I don't remember what rating I gave it, but it wasn't too generous. Reviewing the song lyrics seems like a perfectly good idea, but the thing is - I didn't like the song itself much, didn't think the lyrics were all that amazing, and it was hard for me to really get into the review praising those lyrics and analyzing them in great detail. A review that says: "read this poem, it's genius!" is going to fall a bit flat if I read the poem and think it's kind of meh. This is not a disparagement of your actual writing skill.
I'm old enough to be burned out on people writing Deep Articles analysing song lyrics (way too much of this for Dylan among others in the past) and also reflexively kicking back against song lyric posts in general (way, *way* too many of songfic in fanfiction, though thankfully that seems to have died down again until a fresh crop of kiddiewinkies come along who think their favourite band are, like, sooooo deep and meaningful).
As others have pointed out, the song is not particularly deep'n'meaningful: guy meets gal, they have a fling, they break up, guy broods about it. That's about a good sixty percent of most songs past and present. And this wasn't even a fling but a one night stand.
Overall take; there's a rush through the intro to get to the review, contradictory goals stated, one particular interpretation offered for the lyrics when other interpretations are equally supported, and a general overselling of the song's lyrical tightness.
I have a longer version if you want it. It's a couple of pages, too long to post here.
I don't know if this is the kind of feedback you were looking for, but I just don't feel one song is enough material to justify a thorough review. I absolutely love Dylan and I love this song, but I would have preferred an album review.
I believe I gave it a 7 or 8. It was something I didn't regret the time spent reading, which was my bar for something getting at least a 7. But it didn't change the way I thought about things or give a significant impetus to do new things (aside from listen to a bit more Dylan), so it didn't make it past my bar for 9+.
Thank you so much! I'm glad it resonated with you. That's the best compliment a writer can hope for, that something you've written stays with someone and keeps coming back to them :)
Can we see the ratings for our non-finalist submissions? I'm wondering if my Einstein's' world view review is worth putting up on substack so wondered what score it got.
I feel more comfortable giving rankings in the top half than those in the bottom half, so here are the top 50 runners up: https://slatestarcodex.com/Stuff/50runnersup.xlsx . If you're not in the post above, or on this spreadsheet, yours was in the bottom ~half.
What is less encouraging is that I did Sermon on the Mount Review #2, and it looks like people preferred Review #1 significantly. I preferred my review, but I am extremely biased in this case.
I suspect most raters agreed with something like Billy Hamilton's critique of your review. I actually rated your review more highly than the other one, because I suspect the other one of outright misrepresenting its historical sources and that's a far graver sin, but I do think the other one was more interesting.
Is there a good reason not to show the rest of the ratings? Is it really necessary to shield people from the knowledge of what people thought of their reviews? I thought we were better than that
I tried to see whether I could infer something from this data about the number of votes. With the votes given to 14 decimal places, we can compute the smallest number of votes that could lead to the exact score. For example, the Toki Pona review has a rating of 6.65789473684211 according to the spreadsheet. That's exactly the floating point representation of 253/38, and 38 is the smallest denominator with this property. I reckon that the review likely got 38 votes. Or 76, or 114, although that gets increasingly unlikely.
Some reviews stand out: Orgy review had 101 votes. Dating Apps got 89, Skibidi toilet got 88, and Arbitraging Several Dozen Online Casinos got 77. Also interestingly, the two sermon on the mount reviews got 75 and 45 reviews, respectively... I wonder about the difference, but it could also easily be a bug in my little script :)
I'm quite happy with these numbers, because they mean that scores are reasonably stable. For the Toki Pona review, one extra "10" review would not have affected the ranking at all. One extra "1" review would have moved it down four places though.
I'm surprised Alpha School placed 2nd. I suppose unlike many of the other finalists, it actually delivered on being a *review* rather than a jumping off point for a personal essay, but the flip side is that it didn't possess much of interest if you aren't directly interested in sending your children to Alpha School
I think what happened is that a specific subset of ACX readers *really really* liked the Joan of Arc review and rallied behind it, while people who liked the Alpha School review put it 2nd or 3rd during voting.
That certainly includes me, it was among the better things I'd read all year. Maybe this was a "Shakespeare in Love versus two war movies" voting pattern. The community darlings that were leading on polymarket when I first checked were Alpha School and the thing about dating different categories of young metropolitan stereotypes, and I didn't rank either of those. (To clarify, I didn't look until after voting, and was actually surprised.) My vague sense is that those two appeal to a subset of the ACX readership with specific life circumstances that probably make them more "in the scene" of EA, but it also just isn't relevant to readers outside certain demos or cities.
I might not have bothered to vote had it not been for Joan of Arc, as I didn't feel very strongly about any of the others. My 2nd-3rd were Ollantay and the war journal. There's a throughline on the Joan and Ollantay articles, obviously, and the war journal was just the most remarkable of the remainder due to uniqueness.
I'm sorry to hear that, because the point of Ranked Choice Voting is that it allows participants to vote honestly without worrying about that sort of strategy. Putting Joan of Arc as your first choice, you could've ranked as many others as you'd have liked and still gotten the same result.
> This is not a statement with any connection to reality.
> Wow.
I really really hate drive by comments like this. You might be right - I know nothing about the systems at play - what you aren’t is convincing, in the abscence of an argument.
He's correct about everything except the part where he says I was lying.
I wasn't lying. I was just wrong. IRV is less susceptible to the spoiler effect than FPTP and is still a superior method, but it is not immune, despite what I'd been told by several sources before today.
Thats a strange admission for this blog. Thats against the spirit isnt it? You dont get anything practical out of the one you like best winning. Why shouldnt you be honest about your preferences and let the process end as it will? This is the behavior that ruins rating systems isnt it?
Since all voting systems admit strategizing sometimes, I think everyone should just do that. Although if IRV was used in this case my decision was pointless.
I think that only justified in cases where you have some stakes in the outcome, such as, in theory at least, a political candidate. If you do it here you are not part of team truth seaking
It’s a philosophical question. From one perspective, you’re right I should just trust the process. From another perspective, isn’t it also truth seeking to “help” the process come to what I consider to be the correct answer?
I'm not currently planning on sending children to school but, having been to a school in the past, loved reading about a school system that sounds like it actually tries getting things right.
Why do you not think the question of what actually works in education is of broader intellectual interest? It seems obvious to me that it's one of the most important questions out there.
Absent personal feedback from every reviewer a la Victor, we won't know exactly why readers did or did not like it.
There's also a possibility that this was a satisficing review. I personally did not dislike the review as I said, it just surprised me to see it take top 2. A small core of super fans, and a large mass of readers with nothing against it, would be sufficient to take a high placement
Obviously I can't speak for everyone but I didn't vote for it for being satisficing, I thought it would be a runaway first after reading it (despite not having seen any of the others yet) and stood by it despite several later strong contenders.
I am obsessed with education. I was a teacher for a few years. I currently have 2 kids at school. Some of my friends do various educational projects. Every year my family spends one week of summer vacations at a conference about improving education. And... I feel triggered by this reaction.
School sucks. It's almost the worst possible environment for learning. It teaches children to hate learning. Do you know how 5 years old kids keep asking "why?" about everything all day long? School kills this curiosity in them; trains them to associate learning with stress and boredom. Nerds are a rare minority that remains curious despite school. Many of them are bullied at school.
Most teachers are depressingly conservative. Not politically, but in the sense of: "if we did this for the last 100 years, we will keep doing that for the next 100 years, too; technological progress and educational research be damned". We are lucky that books were invented before general education, otherwise teachers would still consider them suspicious. I think there was a psychological research comparing various professions, and teachers were the third most closed-minded profession, after bureaucrats and professional soldiers.
The "Alpha School" review shows how to do it much better: mostly by doing the obvious. Maybe we could involve computers in education. (Not in the sense of letting kids browse social network, but doing spaced repetition etc. Also, while some kids are training with the computers, the rest can get more individual attention from the teachers.) Maybe we could let children learn also something interesting, in addition to the mandatory. Maybe we could reward them for learning. (Instead of giving them constant stress and fear of punishment.) Maybe we could let them learn at individual speed. Maybe we could just do all of that together, and thereby create something that most current teachers and students wouldn't even consider possible.
My only worry is that this all sounds too good to be true.
The "School" review on the other hand insists that everything I consider a bug is secretly a feature. You know, some kids are lazy morons... but if we invent a system tailored specifically for their personality, it will make school slightly better for them (and much worse for everyone else), and that's apparently a good thing! The children learn little and suffer a lot, but they learn nonzero! Maybe kids ultimately remember almost nothing, but at least the school increased their IQ (that seems to contradict most studies on IQ but who cares) and that's what matters. Different children need different amounts of help, therefore we have to inflict maximum micromanagement on everyone, because apparently the only options are all or nothing. All this suffering and inefficiency is done in the name of motivating the dumbest... without care about how demotivating it is for everyone else.
The author keeps repeating the word "motivation", but in my opinion, school is the opposite of motivation, it is the thing that kills motivation. Some kids love reading books... until you tell them that they will be graded by how many unimportant details from the book they remember; also they need to read the books selected by the teacher, not the ones they like. (My daughter recently got a bad grade for failing to correctly answer who are the best story writers in Slovakia. Who the fuck cares? She has read all seven Harry Potter books and now she is halfway through Narnia, but apparently what matters more is guessing the two names that in teacher's subjective opinion are the most important. If the goal is to make children love reading books, the teacher's strategy tragically goes in the opposite direction.) Maybe if half of the school wasn't optimized for killing motivation, we wouldn't have to optimize the other half to keep it nonzero. Or maybe we could group the morons in a separate classroom and feed them bananas; they might even enjoy it.
tl;dr -- in my opinion, the "School" review represents the worst in the school system, the profoundly anti-fun and anti-learning mentality of the worst teachers
Well said. If "School" accurately portrayed the best we can get, I for one would be prepared to bite the bullet and advocate abolishing public education entirely. The minimal gains we got out of it clearly wouldn't be anywhere near sufficient to offset the suffering it creates.
Mostly in agreement, just doubtful about teachers being "conservative". I wished they were. In Germany, every decade or so one or two new ways to make school even worse are introduced by the ministry of education - usu. under some woke influence. And the schools follow the new guidelines and the teachers defend them, appearing simply unable to say anything critical about those reforms. (Writing before (ie: without) reading / Forcing kids with special needs out of specialised schools: Inclusion. / Keeping kids in class 1 and 2 together - before 28 kids in class 1, and 28 in class 2. Now two classes each with 14 first-year and 14 2nd year pupils. Even more adverse to learning than you can imagine. Huge shock in class 3 when they got to learn all 3rd year material AND the stuff of the first 2 years where very little was learned.
Teachers and ministry of education are not the same kind of people, I think.
Teachers, despite all my criticism, at least do the actual work -- however badly. The ministry is completely disconnected from reality and follow the fads, political and otherwise.
We also have the senseless integration of kids with special needs in Slovakia... the ministry says "no problem, we will give those kids assistants", then in september they say "sorry, no budget for the assistants this year, you need to somehow handle the special needs kids with your existing resources" which is just absurd.
True words. In our primary school they got one assistant, but kids with different(!) special needs in a dozen classes ... At least my son could see: there are kids with more trouble than him. He said, the 4 "special" kids in his class get completely ignored by all the "normal" kids. Heck of an "inclusion". Brutality by Brussels. A friend's daughter is in a special school (speech-issues, logopedical) , they are tears-in-their-eyes-happy they got their girl there instead of a "normal" school.
It would be if it mattered, but it doesn't seem to. It's like working on perpetual motion machines, it's foolish, it can't work, but you occasionally see something that looks like a miracle if you don't know what's been hidden from you. But there is always something hidden from you, and it comes out eventually, every time. Not commenting on the merits, just to say that people who've read about these over the years may see "new educational method" as roughly equivalent to your friend trying to sell you on how his MLM scheme works without drawing a pyramid.
There's a big difference between a hard problem and an impossible problem. I've never seen a convincing argument that getting good effects from education is impossible rather than hard. The right attitude towards a hard problem is to maintain skepticism towards proposed solutions, especially if they claim to work very broadly, but watch with interest when someone gets impressive results, since they might have something that could be a form of incremental progress even if it's not a miracle cure.
It's also worth noting this school got good results in part by making compromises most schools wouldn't be willing to try (e.g. decentering teachers, charging more, being oriented on results over status, only trying to help a narrow segment of the population). It's acting under fewer constraints than most education systems.
Yeah. But that's also the danger in generalizing--the one truth I've ever found in education research is "it works in lab schools only." Everything (within a long stone's throw of vaguely rational) works in lab schools. None of it, to first approximation, works anywhere else.
I'd say that creating differently effective educational systems is provably possible, because we know it's possible to do much worse than our present system. It's been done almost everywhere in all times throughout history. Doing better at scale, economically, is hard, but it's eminently plausible to do better when you know there are plenty of options for better and worse.
Just to be clear, are you saying that unschooling is better? Since that's more or less what was done at most points in history. Having dedicated school buildings, with dedicated teachers, available to everyone in the country, is quite a rarity in the grand scheme of things.
No, I'm saying educational environments for most of history have done less. Mass literacy is a recent development, and in the Middle Ages, being able to read by what we consider middle school age would be notable. If people can pick up more knowledge now without formal education, it's because our educational environment has changed, but there's no particular reason to think we've reached a point where the average person is exposed to the best possible learning environment, and there's no way to do better.
I think there is a good chunk of the readership of this blog who think they are geniuses, but couldn’t flourish due to the structure of school. It’s the easiest genre to pander with. Combine that with the review being very well written and you’ve got a contender.
My teachers were bright enough though it was a private school. You seem to have been dragged through the American state school system. It’s also possible that the school teachers who didn’t think certain kids were smart are actually right.
While I agree with your premise that (well-rounded) "geniuses" accurately observe and analyze social behavior and then use that analysis to productively manipulate the people around them, I'm very dubious that any "actual genius" believes it's possible to teach anyone/everyone critical thinking.
I don't think it's possible to teach everyone to do critical thinking well, but I had a philosophy class in college that covered the basics of formal reasoning and symbolic logic, and while it was covering material I already knew, I was able to see my classmates get noticeably better at critical thinking around me. This was the only class I ever observed that in. I think it's possible to do better than we normally do, because school usually just aims to let students' critical thinking passively improve as a result of learning other things.
I think school generally sorts out the smart from the stupid. The stories of geniuses bored with school are somewhat exaggerated. There’s a subset of people who left school or formal education without much in the way of knowledge or intellectual development who suffer from the Dunning-Kruger effect
Critical intelligence can’t be taught without teaching the facts or accumulated body of knowledge accrued over the centuries. Someone
may feel they can critically think about quantum physics or the causes of WWI, but unless they understand the maths of quantum physics and the historical facts about WWI that opinion is worthless.
This isn’t to say there aren’t autodidacts who contribute to various fields but that’s rare and in fact getting rarer, as knowledge gets more specialised.
As Taymon says, you're conflating intelligence with conformism/pliability here. I'd think genius is, if anything, correlated with contrarianism, not conformism. There's no reason to think that a very intelligent child trapped in a prison built by mediocrities to house cretins won't lash out by maximizing ego damage to the mediocrity placed in a position of direct authority over him.
>(This is the type of person who has literally failed college courses because he did the assignment "the wrong way" -- using energy to solve force problems, say... He did it that way because he believes that the teachers should have to work as hard as the students in order to grade the assignments.).
Having the teacher work as hard on grading assignments as the students are working on the assignment hardly sounds sustainable. The teacher has many more assignments to grade, and only limited time and effort to spend on each assignment.
If a college course is teaching a method of solving some problem, then using a different method to solve the same problem may indeed be a reason for failing. Have your friend considered that the point of the exercise might not have been to find a solution to the problem, but rather to understand a particular path to that solution?
Many people have the ability to come up with creative solutions to a problem on their own, that doesn't take a genius. However, the solution may not be the only or the best solution available. In order to not invent the wheel over and over again, it is necessary to study and understand what solutions have been used in the past. When we understand that, through hard work, that is when we can make progress.
In fairness to these people, most of us aren't eighteen and did successfully flourish after leaving school. It's "I wasted my time there", not "I could've succeeded if only it hadn't stopped me".
I do want to have children, but even if I don't, I have several extended families that I think would do better if I know more about this kind of thing.
Totally in character, Joan of Arc snatches a miracle win from the jaws of defeat, pulling ahead of top contenders Alpha School and Dating Men in the Bay Area.
Thanks for everyone who participated in the Manifold market! It was fun trading with you. Accepting suggestions on how to give back to the community after this.
Congratulations to the winners! Those that won were all easily in my top five. All the finalists were interesting and informative.
I remember reading some comments last year from non-finalists who wished they'd received more feedback. Therefore, here are my notes on the 38 submissions that I reviewed in the first stage of the contest. They were all randomly chosen using the random review button. I also wrote a few paragraphs on my overall thoughts on what makes a good review.
Thank you so much for taking notes and posting them! I did the (F) Sermon on the Mount review, and every year I wonder what I could have done to do better. I imagine most everyone who didn't make the finalist cut wonders the same thing. So thanks for telling us!
I am delighted (and frankly surprised) this won, I voted for it but didn't expect it to do as well as it did (I expected the miraculous element to be the sticking point for this readership).
Stealth Catholic vote ensures victory? Much more broad-minded and generous commentariat than I gave them credit for? You decide! 😁
Eh, a "miracle" is just what interested parties call their favored interpretation of a mystery. And everybody likes a good mystery, even us irreligious irredeemables.
I think that *religious apologism* is, rightly, a sticking point for the ACX readership, and a fair amount of popular historical writing on questions of whether miracles occurred is in fact thinly disguised religious apologism, unless it's instead positioned as debunking. Aevylmar gave a convincing account of why he was interested in this question that did not require a bottom-line position on Christianity to be written first, which allowed us to accept what he was doing as good-faith historical inquiry.
First, I'm not entirely sure that's true in the relevant sense; even if you believe that there's some kind of real supernatural, you'd need an explanation for why it would be in play in any particular case (given that we so rarely encounter it in daily life), or else it's just a god-of-the-gaps argument. You could argue that Joan's miracles were faked with Harry Potter magic or whatever else, but you have to actually make that argument.
Second, in practice, the genre of post I'm complaining about usually ends with something like "so as you can see, we've proven conclusively that something supernatural happened here; therefore, my particular denomination of Christianity is true and you should all convert right now."
I think it's true in the relevant sense, and we're often misled into the opposite assumption, thinking that if something happens which we don't know how to explain naturally, and anyone involved attributes it to God, then God is the best alternative explanation to natural causes. But in the case of a lot of alledged miracles, including Joan's, an all-powerful benevolent God would need to have weird and hard to understand priorities to be responsible for them, when we include our observations of all the places that it doesn't intervene.
From the perpective of a Catholic, I was surprised that while it seemed at the start like another "well miracles can't be real obviously so let's stretch our imagination to think of a way this could have been made up" it ended with the conclusion that there's possibly something to it because the purely skeptical explanation is deeply flawed.
As I'd said in my own comment on the article, the "miraculous" element is in fact what sold me on the article. It really drove home the point for me that Medieval Frenchmen (and perhaps Englishmen) were not like us; they did not engage in realpolitik to the extent that we do; and to them, Joan's powerful conviction and leadership skills (as we'd call them today) would've felt like a miracle. It's obvious that our modern machine technology (cellphones, computers, airplanes, mass-produced corn dogs) would have felt miraculous to them; but I never appreciated that our *social* technology is likewise tremendously advanced !
As a formerly-offensively-atheist now-agnostic - both in the Bayesian sense of "0% and 100% aren't real probabilities" and also in the religious sense - I mostly found the Miracle Question of secondary concern to the mere fact of it simply being a fun and engaging read. Life's weird, lucky things happen more often than mere coincidence, who needs fiction when you have history. Cf. Scott's Fatima post. The object-level question of God's existence is not something that I expect to be argued into/out of by a mere blog post (feels perpetually unknowable with certainty, honestly), yet the stakes are high enough that it makes for a strong topic anyway, especially with a frame that isn't mired in exhausting old New Atheist baggage. Still ended up ranking it 3rd, since I felt it strayed somewhat from the nominal review rules of not being about books, and was more of a quality aggregation of existing well-known sources vs showing new perspectives I wouldn't hear otherwise.
Yeah, the object-level question wasn't nearly as important as the combination of:
- interesting details
- written very well
The part about Joan being "an expert at the top of her game" during the first trial was one that stuck with me, but there were lots of clever turns-of-phrase throughout.
The dialogue was intended to have both "she was crazy" and "it was a genuine miracle" have reasonable points in favor and against, because I think those are both very plausible explanations - or, if you prefer, both implausible.
Joan of Arc was on the top of my list, initially; but upon a subsequent read, I discovered a historical error. The error wasn't a critical one—wrong French king, wrong date for a minor piece of the backstory—but, since I'm not especicially knowledgeable about the Hundred Years War, I had to wonder how many other errors there were that I didn't know enough to spot.
Poor Scott, now he's *never* going to shake off the rumours that he's converting to Catholicism 😁 (something I do not believe is ever going to happen, but now what do I know about what St. Joan and Our Lady of Fatima have in mind?)
Congratulations to all the winners and honorable mentions, and thanks to everyone who entered.
I very much enjoyed the "non-book" aspect of this; I feel like the book reviews for the contests are never really book reviews anyway, so it was nice to have a broader jumping-off point for the submissions to the contest.
Congrats to the winners! All extremely in-depth and interesting. Joan was my personal favorite so I'm glad to see it get its flowers.
Thank you to everyone who wrote kind comments about my silly Mashed Potatoes essay. I had a lot of fun writing it and it was gratifying to make the finals! I don't have time to go back through all the comments from the original post but I will cop to being, as many pointed out, philosophically inept, spinning out a vast and flawed Baudrillard-recapitulating metaphor before walking it back to a relatively minor point about semantics. I identified the pattern I wanted to write about first and then struggled to tie it all up with a bow in the final section with any kind of consistency or rigor. I do feel it's gesturing at a real dynamic, even if it's imperfectly explicated, and I'm glad some of you found merit in it anyway (except for the guy who thought I was GPT-5? Not cool bro).
I loved your review! It was my favorite. Thank you for identifying this pattern/phenomena and putting it into words. There's definitely a "there" there, even if you "struggled to tie it all up with a bow" at the end.
>except for the guy who thought I was GPT-5? Not cool bro
I legitimately find it less insulting to be called ugly than to be accused of having used generative AI to compose a piece of text I slaved over with my bare hands.
For what it's worth, I have the reverse criticism; I thought your writing was good but remain skeptical that the dynamic you're talking about is actually a real thing.
I passed a happy car journey listening to it! I think it worked on several levels, both concrete and abstract:
- It has straightforwardly interesting facts about the origin of potatoes etc
- It weaves them into a more general and interest narrative about mashed potato, which I'd never really thought of (I grew some potatoes this year so it oddly did interweave with my concrete problem of how to preserve them!)
- It makes a broader abstract point which I think adds something to the mental toolbox (although I think whenever I apply IMPishness I need to think a lot about the context of each individual case)
All the other reviews had their own flaws by the lights of other commenters, too, and I don't think you need to be so apologetic.
I think you're being too modest. I voted for it to win, it was my favourite of the reviews. I learned some new things about potatoes and the simulacra comparison was very apt. Give yourself a pat on the back.
I voted for yours despite having minor criticism of it at the time. I appreciated it being a review of an actual thing and also not being way too long.
I liked your essay and would have voted for it with a more satisfying wrap-up. If you have the time it might be worth trying to resolve the dangling threads and to post a revised version. Thanks for sharing!
I thought your review is the one who most embodied the "everything except book" review. In the way that it's the most similar to John Green's Anthropocene Review. I hope next year would have more reviews like that.
I felt it was the best written review and among the top 3 overall. I am also a fan for the whole unbundling-rebundling of systems and things, or as I am starting to call it , the IMPfication of the world.
I'm still irked that you maligned drywall in a casual drive-by line. Have you ever tried to hang a picture frame on a plaster and lath wall? Or a TV? Or a shelf?
It's such a giant pain. No standard wall anchors work, so you have to drill a big half inch hole, buy expansion anchors from the hardware store, and slide those through. Then you have to manage their alignment to be across laths instead of along laths so they don't break through, then yank to dig them in, then figure out how to adapt your hardware bolt to whatever you're trying to hang...
Putting trim on is a nightmare. Every wall is crooked and bumpy in ways that are invisible to the human eye until you put a piece of straight trim against it and it looks horrible. I've had to put low expansion foam behind trim in places to prevent mice from crawling in and out.
Drywall is a counter-example to instant mashed potatoes, not an example of the same trope. Grinding up soft rocks and gluing them together into perfectly flat 4x8 boards is SO much better than the old alternatives. It's not just price, the quality is better, and those of us who have lived in old houses don't want to go back.
lol very fair! I've never lived in an old house, very few even exist in Florida. I'd say that drywall is an example of something I mention later in the essay, an imitative species that is nevertheless more salubrious than its model and deserves to continue being planted on its own merits.
I liked it a lot and it was in my top 3, but allow me to chastise you for cheating me. You promised me a review of mashed potatoes, gave it a couple of paragraphs, then segwayed into the history of potatoes and their conservative technics. It was a good story. It wasn't the one you promised me
I guess Brandon Hendrickson's anti-precedent (where reviewers tried to limit reviews to <10k words in 2024 after his 2023 review of The Educated Mind got meme'd for being way too long) only lasted one year.
Congrats to all the winners! I wish Edward's kids (along with the rest of the Alpha School students) all the best!
This is just my personal opinion (and so I'm prone to the pundit's fallacy in thinking it's more broadly shared), but this year's winning and second-place reviews were the two longest submissions because they both genuinely had a huge amount of ground to cover, whereas the review of The Educated Mind seemed to be drawing out its argument as long as possible in order to wear down readers' prior inclination to dismiss its bottom-line conclusion.
The HYW stuff was more interesting than the JoA stuff, though. To be honest I think the review benefitted massively from the historical context vs ignoring that to focus more narrowly on JoA.
Before I put them in it was kind of incomprehensible. With longer editing I could probably have gotten it down significantly, but I actually wrote the essay in a desperate rush the week or two before the deadline and so I didn't have time for all the editing I wanted.
I can relate. We've all been there. Editing is hard. Your review was still very well-written and I'm glad you won. I'm looking forward to what you're going to write in the future!
The length was such a sticking point for me personally because it's not good for the overall health of the contest. It's fine if one person submits a super long review; it's not fine if everyone does. Initial voting becomes a slog.
Back in 2023, when Brandon won, I said at the time that I voted for it despite its length, not because of it. Commenters were worried that his review would set a bad example, and subsequent entries would all be super long.
Then 2024 rolled around, and Scott instituted a soft 10k word limit. And... people abided by it! I read (and rated) a significant chunk of the initial submissions that year, and I enjoyed all of them. When the entries are a manageable length, it's really fun to read review after review. All the finalists were under 15k words, with most of them being under the soft 10k word limit.
I don't know why people stopped doing that this year. Scott gave the same (soft) 10k word limit, but a lot more submissions blew right past it. This was a bigger problem than just you, specifically.
I barely read any initial submissions this year, simply because of math. In the time it takes me to read one 30k word post, I could have read and rated three or four <10k word reviews. That adds up when there are over 150 submissions.
I think part of the reason why the ACX commentariat was disappointed in this year's initial submissions was because they had the same constraints I did, and people simply couldn't read as many during initial voting. They didn't have an opportunity to read a ton of them and find their favorites; they probably only read a handful and then stopped. A huge part of the enjoyment I get out of this contest every year is from digging through the initial pile and finding the awesome ones that appeal to me, but not necessarily the rest of the ACX readership (and so never make finalist). I can't do that if all the submissions are long.
All that to say: your review was great! Congrats again!!!
I don't read all the reviews. I pick out ones that seem interesting (based on title and/or subject matter), start to read, and if it bores me, I don't bother with the rest of it but move on to something else.
I can imagine very conscientious people dutifully slogging through the whole of every entry, but I don't think we *have* to do that (barring the voice of conscience compels you).
Why do you think the ACX commentariat was disappointed in this year's initial submissions? Sentiment in the first-round voting thread was largely positive, and where negative was largely about minor-ish process issues (rather than content of submissions), plus some arguing about what's a "review" and some performative complaining about Skibidi Toilet.
In any event, if you're against long reviews, then your complaint is with the voters, not the submitters. There weren't in fact any submissions that came close to 30,000 words. The three longest were Joan of Arc (~24,000 words, first place); Alpha School (~18,000 words, second place); and Dating Men in the Bay Area (~15,000 words, probably fourth place according to Manifold). Besides those three, only four others were over the 10,000 word limit (one of which, The Synaptic Plasticity and Memory Hypothesis, was also a finalist), and only one (Deathbed Ballads, ~14,000 words) was clearly over the limit if it's interpreted as a "soft limit". This is out of 144 submissions in all. (Complete list of word counts here: https://www.astralcodexten.com/p/choose-nonbook-review-finalists-2025/comment/122686238)
So there wasn't in fact a huge wave of super-long reviews. Rather, there were a small number of super-long reviews, and they were the ones that did best in the voting. This makes me think that the pattern of finalist lengths from 2023 to 2024 to 2025 was probably just a fluke.
Maybe your threshold for noping out was actually lower than 10,000 words?
One of my takeaways from the whole review and commentary process is, there's almost no piece of commentary on any of the finalists which isn't balanced be commentary in the opposite direction. At a certain point, you're not making the review better or worse so much as you're adjusting the sort of reader it's intended for.
Could it be that my taste is the exact opposite to ACX readers in general? In my ranking Joan of Arc and Alpha School were 12th and 13th, while School and Mashed Potato were 1st and 2nd. Any chance of publishing the full ordered list?
Alpha school was definitely low for me. Joan of Arc was middling. Although I admire the work put in. The mashed potato was an excellent essay, one of my favourite essays in a long time. Maybe people didn’t think it was a review.
Is there any possible explanation for one submission getting an outlier number of ratings, other than that someone elsewhere on the internet linked it and drew attention to it? I don't think a submission can achieve that by its merits, since we all started off with just the giant Google Docs full of reviews that we hadn't yet read, and so couldn't know in advance which ones were any good.
Speaking as someone who doesn’t ever bother voting in these things, and mostly doesn’t bother even reading honorable mentions because the Google doc is too awkward to deal with on my phone… I had heard of Red Means No, and probably others had too, since Aella is somewhat famous around here. I have not heard of young swingers week. So if I were the going to read stuff, I’d be more likely to read about the thing I’d heard of
Huge congrats to William on the Joan of Arc Review! The other two I voted for (Dating Men in the Bay Area and Instant Mashed Potatoes--both excellent) covered ground that I'm interested in and made arguments I think are important. Joan of Arc had me rivetted to the screen in spite of its length and my lack of prior interest. Very well done!
Great minds! Honestly, I'm surprised that Instant Mashed Potatoes made it to the finals at all--I loved it, but it so cuts against what I stereotype as the ACX view of things. Glad to hear you voted for it!
Not everyone has been following Scott that long. In which case I recommended this old post (don’t worry about the size of the scroll bar, that post had a few hundred comments; the post itself is pretty short)
"No paid consulting. No discounted tuition (or more accurately no discounted tuition that was not provided to every other family in the school)
I did have a couple calls with their marketing team at one point - but it was purely volunteer to help them out. I do the same thing for lots of organizations."
So while the author received no compensation, I think it is fair to call the review "undisclosed infomercial". It was written by a marketer who volunteers to help Alpha School convert customers and apparently employs psychological conversion tactics (analysis here: https://pastebin.com/fJfvQFQ2). The recent posts in the author's new linked blog can be seen in the same light.
ORIGINAL COMMENT: Hi, regarding the "Alpha School" review, the author is a growth marketing expert by profession according to his linkedin: https://www.linkedin.com/in/edwardn/, the review reads in part like ChatGPT-assisted viral marketing copy (the ChatGPT feel has been noted by other commenters), and Gemini or a similar LLM flag some stylistic devices in the storyline and rhetorics that are evidence in favour of this, see e.g. https://pastebin.com/fJfvQFQ2 here^1. I think it would be appropriate to get a statement about whether he received any form of compensation from Alpha School or other ESW/Joe Liemandt-affiliated entities, including e.g. discounted tuition, either explicitly for writing this review or in some other consulting-like role? I do realize that the truth may be "I really love to help them grow even without taking money, because I am convinced of the product" - I am more convinced of what the Alpha school product tries to do than I am of public high school too - but I'd like to know.
In particular, part of the Alpha/GT marketing seems to be explicitly directed at the rationalist community - look e.g. at the https://x.com/gtdad account, which follows rationalist/adjacents like Kelsey Piper or orthonormalist and is part of the Alpha School marketing system. I also think there is a pattern of entities receiving money from Alpha School or other Joe Liemandt-affiliated companies and writing positively about the Alpha school system (e.g. tracingwoodgrains,^2 whose Center for Educational Progress to my understanding has been mostly funded by an Alpha School grant as of a few months ago, or Austen Allred ^3). Once again, I find it absolutely plausible that the enthusiasm would be genuine (like the review author, Allred has his own kids in the thing, which he wouldn't do if he didn't like it), but I think it should be made clear whether any money or other advantages flowed to the people voicing it.
In fact, the connection is now quite direct - Alpha School founder Liemandt himself has made this very ACX review the pinned tweet on his X profile, calling it the start of "The Alpha School virus" and explicitly encouraging people to vote in the contest: https://x.com/jliemandt/status/1974984216027738467
^1 I got this response when pasting the prompt "Is this text likely some sort of hidden advertising? Can you tell me evidence in favour and against?" together with the review text into Gemini 2.5 Pro on Google AI studio.
^3 Austen Allred's writes positively about Alpha as well: https://x.com/search?q=alpha%20from%3Aausten and his current venture, Gauntlet AI, was to my understanding bootstrapped mostly by hiring demand from Joe Liemandt's education and other ventures.
"No paid consulting. No discounted tuition (or more accurately no discounted tuition that was not provided to every other family in the school)
I did have a couple calls with their marketing team at one point - but it was purely volunteer to help them out. I do the same thing for lots of organizations."
So while the Alpha School review was written by a marketer who is enthusiastic about helping Alpha School convert, there was no compensation.
Hey T T I am trying to DM you, but when I enter T T in the address field the system does not recognize you. I have some news to pass on to you but it is not for public consumption.
I just DMed you by mouse-hovering over your icon next to a message, which causes a profile to pop up, and clicking "Message" - please let me know if that doesn't work for you for some reason
I'm glad you'll be alternating book and non-book review contests: I liked the latter enough to want to see more of them, but of course love book reviews and wouldn't want to simply subsume them into a larger "review anything" project.
Congratulations to all the winners.
I'm bummed that "11 Poetic Forms, Reviewed in Verse" didn't even make the honorable mentions list. I really liked it.
I really liked the concept of "11 Poetic Forms, Reviewed in Verse" but didn't feel that the author was quite a skilled enough wordsmith to pull it off. If Scott had written it, I bet it would have been great. (I feel bad about saying this because it's good for people to try ambitious things, but ultimately I gave high ratings to ones that I felt were successes.)
In case the authors or anyone else is interested, here are the short notes I wrote to myself after reading each review, to refer to later when voting:
Alpha School: pretty good, little on the long side
School: really good core insight, writing kinda dull
Mice Alzheimer’s: pretty hard to get through.
Islamic Geometry: fine I guess? Not very exciting.
Commentariat: fun, but dragged a bit in the second half.
Joan of Arc: Quite good, easily my favorite so far.
Instant Potatoes: Fun and interesting, not too long, but point is suspect and also mashed potatoes aren’t actually good.
Dating Men: Pretty good but feels a bit pandery.
Ollantay: Another top contender
Phase 1 Research: Interesting and a good read.
Synapse Memory: Interesting but dubious and kinda long for the value.
Xanadu: Okay I guess.
Russia-Ukraine War: Pretty good but not sure it’s a winner
IIRC I ended up voting for Joan, Ollantay, and Phase 1 Research. Phase 1 might have been confounded in either direction since it was the only finalist that I read and voted on in the preliminary round.
Are we gonna start the Great Potato War of 2025 on here? Because I feel the Irish and Germans would line up on one side, you heretic 🤣
It depends on the type of potato (waxy versus floury), how they're cooked (boiled into mush?), and how they're mashed (drain off the water and allow steam dry a little or mash them into a gruel in the water, then leave them coarsely mashed or put them through a ricer, then lots of liquid such as milk to make them soupy versus no, you want a solid potato not something you can drink with a straw).
One thing I like is to do all the prep up to the point of boiling, but instead of putting the pot onto the stove, letting the potatoes sit in the salted water at room temperature overnight. That's usually enough to get a bit of fermentation going, to break down some of the tough stuff and add a bit of tang.
Of course, slicing them then and then fermenting in Mason jars for a few days before *baking* results into the chips Sea Salt and Vinegar chips want to be when they grow up.
Thanks for running an excellent contest, Scott. I'm always curious to see how alternative voting systems work in practice, is there any way you could share how the RCV method actually determined the results in this case?
why post the contents of footnote 3 about getting past the spam filter, as opposed to send it in an email to finalists? LLMs are good at scraping and finding these sorts of things
I am thrilled "Alpha School" placed, and just want to take a moment to highlight an aspect of the review that I don’t think has gotten enough praise: this man uprooted his entire life to ensure his kids got the highest quality, healthiest education possible. And, unlike some parents who obsess over their kids’ education, he graciously and lovingly admits that both of his kids have unique personalities, and is eager to find individualized motivations that work best for them.
Edward, you're an incredible father. I want to hug you and clone you.
I help with mentorship programs, and my experience with the parents who are obsessed with their kids' education is that they tend to want their children to fit a very cookie-cutter mold of success. They get fixated on their kid getting into Stanford, MIT, etc, and then force their child toward whatever interests/activities/achievements will get them into those schools. They essentially create a personality for their child and then force them to wear it.
I instinctively cringed a bit when I saw a very long post about someone trying to optimize their kids' education. But I was very pleasantly surprised by how in-tune Edward is to his kids' mental well-being, unique interests, strengths, weaknesses, etc.
I think he's a rare balance of someone who is both hyper-fixated on his childrens' success and also deeply respects their children as individuals and cares for their mental health. And I think it's worth applauding. Frankly, I'm baffled by why you'd want to mock it, but so be it.
I wasn't mocking your points, I was critiquing them. But I did not get that you were not comparing him to parents in general, but saying that within the population of parents hyper-fixated on their kids' achieving certain markers of success, he's much better than average. That I do not find absurd, though I'm still inclined to think of it more as being the best of a bad lot than as a precious and remarkable achievement.
Critique is an art form of its own, and I would suggest you learn how to give it in a way that does not sound like mockery. Your point very easily could have been made without sounding snide.
And to each their own. I never got the chance to go to high school, so my mind is absolutely blown when I come across individuals who put so much effort into their children's education. When done in a compassionate and humanizing manner, I absolutely believe it's worth applauding.
You are right that I could have made my point in a pleasanter way. You could lay my snideness to spillover anger from other, related things I object to. First of all, I do not like the sound of Alpha School. The thread where the review was posted for comments has something like 10 long, detailed posts of mine articulating what I think is wrong with its objectives and methods. Second of all, I am very angry about information that has come out about Alpha School and the review's success iin the present thread: The founder of Alpha School had a pinned tweet on X asking his 45,000 followers to vote for the review: https://x.com/jliemandt/status/1974984216027738467. And the request was retweeted 32 times. That's ballot stuffing. Still want to clone him? IMO the world already has enough people who use their smarts exploitively for us to get by ok without cloning this mofo.
On the other hand, calling my anger spillover suggests that it's irrelevant to my present complaints about the writer and dad in question. It's not. Seems to me that both he as a contestant and the school he's promoting in his infomercial are overconcerned with winning competitions and underconcerned with being trustworthy and real.
I guess if everyone he asked to vote did come over here and vote, then yeah, that's ballot stuffing.
Good job St. Joan thwarted his plans just like she did the English, then!
"When done in a compassionate and humanizing manner, I absolutely believe it's worth applauding."
That's the rub, isn't it? Alpha School sounds like rearing turkeys for Christmas. I enjoy my Christmas fowl but I wouldn't raise another animal, let alone a child, like that: plump 'em up fast for the slaughter and the profit of the owner (of the brand which he seems, if Eremolalos is recounting correctly, to be marketing as hard and perhaps as deceitfully* as his little spindly legs will run on the treadmill).
I agree with Eremolalos that I don't like being used for free advertising.
*There seems to be some little blurring about who founded what and who owns what; one source (Wikipedia) says the founders of Alpha School are MacKenzie Price and her husband Andrew Price; another source (Wall Street Journal) says some guy called Bill Ackman is acting as the public face/ambassador for Alpha School, and finally Wikipedia again says the owner of the magic software that the school operates off is this Joe Liemandt, who also owns the companies that Andrew Price works for.
I often feel sort of at home here, which is a rare experience for me, but the present “review” contest has had me wandering deep in familiar alienation. I hope some of that reaction is accounted for by how many outsiders participated as contestants and "judges." One post here today (https://www.astralcodexten.com/p/non-book-review-contest-2025-winners/comment/167323863) says that the founder of Alpha School, Joe Liemandt, had a pinned tweet on his X profile asking readers to vote for Edward Nevraumont's review of Alpha School. Seems to me there's virtually no chance that Nevraumont was not aware of the pinned tweet or that he asked Liemandt to take it down and Liemandt refused. Also seems highly unlikely that any who arrived here via that route actually read any of the other reviews. So they weren't really judges or voters, just unpaid ballot box stuffers.
GPT's assessment of amount of AI-generated content in the essay:
>The piece reads overwhelmingly as human-written with some AI-style scaffolding or heavy-AI paraphrase in certain middle sections. The dominant voice is cohesive, witty, and self-aware—traits that large-language-model prose still rarely sustains across 25 000 words. But there are several passages whose sentence rhythm and connective phrasing match AI composition patterns strongly enough to suggest the author may have used an AI draft or rewrite tool, then edited heavily. . . . Overall, the essay’s conceptual architecture and narrative sensibility are human, but perhaps 15–25 percent of the phrasing bears AI fingerprints—primarily in data-dense or connective prose.
The soul of an anti-woke intellectual: Jamie Q. Roberts
*SINGLE MOST INTERESTING PARAGRAPH*
Japanese social interactions exist at a much higher resolution than American ones, and at times I felt that living in Japan as an allistic person gave me a reasonable understanding of what it might be like to be autistic in America. At all times there were subtle games being played, and things being communicated by other people to which I was not privy at all. This created a background radiation of stress that made social interaction much more stressful than fun. Japanese people get around this by leaning into role assignment; basically, your boss is just your boss, and must play the role of boss rather than ‘being himself’ because if everyone was particularly ‘himself’ at this high resolution of social interaction, everyone’s minds would collapse under the information overload, and life would be impossible, even for those born into the culture. (From The World as a Whole)
*MOST ACUTE AND ENTERTAINING SINGLE SENTENCES*
“Less powerful than Cantor’s [wings], but still airworthy enough to reveal the vistas that Cantor saw; and robust enough not to melt and send us plummeting into the sea of mathematical paradox.”
“One sees crystalline souls that have polished all their facets into a perfect refracting jewel of self-awareness, and often a performance of self-contempt, so no one can ever accuse them of being cringe.”
“. . . young people who are by default edgelords, unserious Christians, iPad kids, psych med-takers, or bog-standard faces in the halls to mine the social media algorithms for what is most in opposition to the listlessness and malaise they feel.”
“Plastiform glee that intensifies into a sort of pop-eyed, carnivorous ecstasy. “
“This speaks to its visual literalism.”
“The villanelle so strangles sense that one can't tell the meaning from the structure.
“ . . . something done in the physical world in an intangible age. An age of circling back, checking in, lit reviews that go nowhere, and innovative best practices in marketing.”
“He appears to have an almost fanatical devotion to parsimonious explanations.”
“The concept of bad faith- akin to mauvaise foi but with a sunk-cost ratchet.”
“It is immortal precisely because it is flawed. Like human language, its ambiguity delivers the creative freedom that closed systems cannot.”
“Past the veil of shame is where the dark, gross, raw workings of the heart lie. Meet me there.”
To be clear, the person encouraging to vote in the contest (together with promotion of the review) is not the author of the review, but the founder of Alpha, Joe Liemandt himself (he used to be quite secretive, letting Mackenzie Price present as the public face, but has become less so in the last months): https://x.com/jliemandt/status/1974984216027738467
Before education, this person got (in)famous for acquiring struggling companies and operating them with remote contractors (https://archive.is/ar0Cr), using his remote working platform Crossover that pioneered bossware/mouseclick "productivity" surveillance (https://archive.is/PpoLs). A big part of the philosophy of Alpha School seems to be adopting this metrics and surveillance-driven approach to schoolchildren.
According to the linked blog post, Liemandt started presenting more publicly in part because the review author accidentally outed him, not realizing that his involvement was supposed to be kept quiet.
And do you have any doubt that the author of the review could have kept him from soliciting votes on X? In my opinion anyone who solicits votes on social media or accepts having someone else do it for their own gain or the author’s should be disqualified
Thanks for singling out my paragraph. I agree that promotion of reviews on other platforms sets a dangerous precedent. It would take very little for the contest to lose all meaning because of tactical social media campaigns.
You're in the Single Sentence category too, with your riff about Macedonian teens. You seem very gifted to me in your takes on people and settings, and your ability to put nuanced takes and odd unsettling truths into words. I hope you'll keep writing, so long as it gives you pleasure. There are way fewer settings than there used to be where talent like yours is appreciated, but there are still some. My own talents are in the same realm as yours, and I've had to come to terms with the fact that ACX is just not a setting where gifts in that realm are much valued. If you hang out here, it has to be for the other great things about this group. You might be amused by my own entry, which is about AI text-to-image art, but can be read as an extended angry riff about the triumph of plastic analogs over the real deal. It's Dall-2 & 3 as Depictors of Human Life. But fer crissakes don't feel like you owe it to me to give it a read! Anyhow, DM me if you'd ever like to chat.
I enjoyed your thoughts about the superiority of Dall-E 2 over 3, from an artistic perspective. What do you think about the newest image generation models?
They’re Dall-e 3 on steroids. Everything looks grotequely perfect, glossy, and commercial arty. And lately the sites for online text-to-image generation are all over the user with pop-ups etc trying to addict them to making AI videos. I haven’t tried the video features. They look especially unattractive to me because there seem to be a lot of stock characters and preset grooves for the stories to run down.
I have a beloved collection of Dall-e 2 images, prob. about 200, and would love to make more. When life settles down a bit I am going to see what I can figure out about training a model that is more like Dall-e 2. I am not at all tech-trained. Have never written a line of code, do not know what API even stands for. I look it up, then forget. But I think with help from someone I know I can at least try it. I do not know whether that approach is promising, though, because it seems to me that a lot of what made Dall-e2 excellent was the *lack* of various kinds of enrichment and controls. Seems like train a new model really comes down to tweaking an existing one — you don’t start from scratch. So you’re adding something to what’s there, and what I want to do is subtract.
Have you played around with the text-to-image AI’s?
I have played with https://diffusionbee.com/download as a frontend for the confusing plethora of open models out there. Unfortunately the base Stable Diffusion models were poorly trained (poor quality training data and a weak language part) and the various later finetunes have particular obsessions behind them so I haven't found a model I can really recommend for artistic flair. I've seen discussion of https://github.com/lllyasviel/Fooocus as a frontend as well but it looks much less polished. I also have a bunch of Dall-E 2 images that had a spark that seems missing in newer, slicker models.
I am hoping someone will look at making open image generation models again, starting with a more capable open language model, but this requires real effort and a lot of curation to avoid copyright issues, let alone the righteous scorn of artists who didn't put their work online for the purpose of being automated away.
Having been a reader since the early SSC days, I too have felt vaguely threatened and alienated by the massive influx of normies. This wasn't helped by Scott generally being busy for other reasons and not quite as on top of moderating as usual.
I agree that outside promotion is icky. I suspect the LLM slop will only become a bigger problem, too.
Haven't you just disqualified yourself as a rationalist by making an us-versus-them distinction? After all, EY's writings emphasize truth-seeking over social belonging (but he talks a good game).
Anyway, I guess I must one of those normies. But I'm curious about how you detect a rationalist from a normie by their comments? The epistemic process people use to support their comments are seldom made explicit, so the "rationalness" of their conclusions is frequently unclear. And despite their espousal of Bayesian reasoning, I don't see many people updating their priors when confronted with contradictory data. Maybe we're all normies? Just sayin...
I never claimed to be a rationalist. I'm just some guy who's been reading Scott Alexander for just over a decade. I'd never even commented until recently. I've read The Sequences and think EY is a bit of a hack (don't write a preface explaining that you used all the wrong examples--go back and FIX IT).
I've seen your name around. When I say "normies", I mean the people who are clearly not regular readers of the blog (which I guess from the contents of their comments, apparent lack of awareness of previous posts and reviews, &c).
In addition to everything else (especially Joan of Arc, the only review that earned a 10 from me), I want to congratulate Max Nussenbaum on having now written *three* banger ACX contest reviews that I keep coming back to for a good read. It's disappointing that this one didn't make the finals, but I'm glad to see it recognized anyway.
I'll be frank and say I'm crushed my review didn't bubble up to the top of the honorable mentions. I put my heart and ten years into it though, so I suppose I'll try to be content with having let out many things that I've long believed and left unsaid.
Congratulations to all the winners. Here's a link to my review if you didn't read it. Shout out to the commenters who singled it out, and seemed to understand what I was getting at.
Originally, I haven't noticed it among the submissions. I've got to reading it only now. And holy moly, that's a seriously good piece of writing. It's almost Steinbeckian in its depth and insightful in that particular human-centered way that's almost fully missing from the rationalist-adjacent blogosphere. I still have to fully digest it. And then maybe write a review of it.
Next time, I suggest writing something that doesn't make people want to kill themselves after they finish reading it. Pander to your audience! Reinforce their prejudices!
Good idea! I plan to write a sequel that casts the review in a less pessimistic. It's intended more as a picture of things as they are, then as they must be.
I understand that, to a degree. To those that already know, it's only a depressing reminder, and to those that don't know, it may seem like nothing but an argument that they should feel bad. But I don't believe in collective guilt, and that isn't the point of the essay at all.
Happily surprised by Alpha School - not being a winner, that was kind of a given, but by author. I remember doing my part to dunk on the previous comics review. It's unusual to dislike something so strongly that it becomes a recurring synecdoche for bad things of a type! Big tsuyoku naritai energy for learning based on gradient descent and putting out a much, much stronger review this time. "As long as you learn something, it's not a mistake," says Archer.
Befuddled by not recognizing any of the finalists as commenters whose writing feels very familiar, like as if they're a big-name regular that I've been reading for years. At least two entries this year gave that strong sense of readerly deja vu. More evidence of convergent evolution, I guess...spend time on ACX, start to write like/complementary to Scott.
Most of the finalist and honorable mention reviewers aren't people I know, but in two cases (Joan of Arc and JFK Assassination Conspiracy Theories) I found specific tells that allowed me to, with fairly high confidence, figure out who the author was before today's reveal. (In the former case, there was actually a Manifold market on it, not created by me, and I made a decent profit.)
IMO, long form essay writing is fairly different in style from blog comments, so I don't try to hard to spot stylistic tells. But as far as convergent evolution goes, keep in mind the filtering effect. The blog readership is heavily selected for people who like Scotts writing style. When they vote on finalists, that preference is expressed in the selection. Plus, to an extent, the authors are probably catering to that deliberately.
Just commenting to say that I really really liked "The Life's Work Of Banerjee, Duflo, and Kremer", I'm sad that it didn't make it to a finalist but I'm glad that it made it to the top half
Well, most people here sound sort of like people at the end of a fine Thanksgiving dinner. This was great — thanks, Scott, for hosting it, and for all the fine people who participated by writing or judging — great bunch of essays, guys — congrats to all . . . So I feel sort of like the skunk who slipped in through the dog door and shows up in the dining room doorway, all scary and stinky, saying this:
Edward Nevraumon, author of the Alpha School review, cheated. He should be disqualified. There are a number of posts about this matter scattered on the thread, but here’sthe brief version. During the competition, Joe Liemandt, founder of Alpha School, had a pinned tweet on X asking followers to go to ACX and vote for Nevraumon’s review. (https://x.com/jliemandt/status/1974984216027738467). It has 45.5 K views, and was retweeted 32 times. It went up on October 5, 3 days after voting started.
Nevraumon has not denied knowing that tweet was up, and it would be quite implausible if he did deny it. The fully honest thing for him to do would have been to tell Liemandt to take down the tweet as soon as he knew it was there, then inform Scott of what had happened. But I’d settle for his just having quietly asked to have the tweet removed. But nope, the thing’s still up, and presumably was up from 10/5 til voting ended on 10/13. Nevraumon, who’s a growth marketing expert, has denied receiving any direct or indirect compensation for writing his piece about Alpha School, and I don’t know anything that makes me doubt that, but it’s clear there was communication and coordination going on between Nevraumon and Liemandt about the review. How otherwise would Liemandt even know that Nevramon was entering a review in the ACX contest, much less know when voting began?
Last year one entrant was disqualified during the voting, because of “irregularities.” Seems likely irregularities were suspicious patterns in the votes the review received, though I supposed there could instead have been reason to suspect plagiarism. In any case, it seems to me that the case for disqualifying Nevaumon cannot possibly be weaker than the one against last year’s disqualified reviewer. In fact I’m having a hard time imagining a stronger case. He told someone who’s deeply invested in the success of Alpha School about his review, and informed him when voting began. He said not a word to anyone here about the tweet soliciting votes. Some fraction of the 44.5K people who saw the tweet (plus of course those who saw retweets) very likely came here and gave him a 10. That’s ballot stuffing. Even if in fact not a single X rando showed up and voted, Nevaumon is still just as guilty, because he went along with an attempt to send an army of ballot stuffers here to up his score.
I’m amazed that nobody but me has spoken up loudly about this. It seems like a clearcut, especially egregious case of cheating to me. But of course the tone here today nudges everyone in the direction of being mellow and positive.
I’ve been reading since the SSC days and AlphaSchool was a favorite of mine and I also judge it likely to interest the typical reader of this blog, at least any that have or are thinking of having children. It also gets bonus points for being basically the only actionable reviewed item. I don’t find its performance surprising and I don’t consider someone else tweeting something to be a form of the author cheating nor do I expect a review author to be aware of tweets about his review, even from people the author knows.
Of course there is an alternative to disqualifying Nevraumon.. We could agree to just not treat vote solicitation in the review contest as a problem. I don’t believe Scott stated anywhere that vote soliciting is not allowed, or even said that it’s sleazy and please don’t do it. And of course there is no way to fully enforce a rule of this nature. Only pretty flagrant violations will come to anyone’s attention. Any anyhoo, that’s how the world works, right?
In that case, I would like to be the first to take advantage of the situation by making an offer to the group. Next year I will enter the best, most crowd-pleasing review I can, and pay up to 4 people to give my review, and only mine, a 10. I will give each person who does that 10% of my cash winnings, or $500, whichever is greater. For those squeamish about accepting cash for their votes, I am also open to an arrangement in which the individual gives me, and only me, a 10. I in return will give their review, and only theirs, a 10 *and* guarantee that 2 close acqaintances of mine also give them a 10.
You scratch my ass, I scratch yours. That’s how things work in most of the world, and it’s time we got with the times, right?
Just in case your comment about people giving the Alpha School review a 10 reflects genuine confusion on your part: Voters were asked to list their top three choices from among the finalists. The rating process was done earlier, before the tweet was posted.
Since Scott has not (as far as I know) posted any specific rules about this kind of cheating, I don’t think that this incident should be treated as an automatic disqualification. I think that ballots cast by people who don’t have a history at ACX (voting in previous contests or posting comments here prior to the tweet), and which list the Alpha School review as the only choice, should be considered suspicious. If there are enough votes in that category to change the outcome, I think that would be reasonable grounds for disqualification because it would mean that the purpose of the contest had been subverted.
Of course this is really up to Scott, and I don’t know whether he is reading this or is even aware of the controversy.
This incident has changed my view of the school more than my view of the essay. I might have followed the progress of the school; now I no longer consider it worth my time.
<Just in case your comment about people giving the Alpha School review a 10 reflects genuine confusion on your part:
No, sorry, I understand the voting system. What I wrote reflects genuine preoccupation not genuine confusion.
<Since Scott has not (as far as I know) posted any specific rules about this kind of cheating, I don’t think that this incident should be treated as an automatic disqualification.
I think Scott simply assumed everyone would know that soliciting votes from randos on public forums was not OK. Here's why: 1) He also did not, so far as I know, state that having plagiarized bits in your review was against the rules. or that entering under your name a review written by someone else was. He did not state that people should not DM other writers and threaten to out them for something-or-other, or embarrass them with false but hard-to-disprove accusations, if they did not withdraw their review. He did not say that reviews that were entirely AI written and not identified as such were forbiden. He did not say people were not allowed to set up multiple accounts here for themselves and use them to give themselves multiple votes. I believe he assumed we all knew those things. 2) The review contest is a contest, not a political campaign. It's a contest of skill, skill at writing a review that delights ACX readers. Everyone knows that in contests of skill using means other than personal skill to win is cheating. 3) Scott disqualified someone's review last year for "irregularities." I think he said the issue was voting irregularities, though I'm not sure. There don't exist many kinds of voting irregularities in this situation. In fact the only one I can think of is somehow getting votes that are not genuine. So someone could get other people to vote as they direct them, or I they might somehow set up a bunch of accounts and give themselves multiple votes.
>This incident has changed my view of the school more than my view of the essay. I might have followed the progress of the school; now I no longer consider it worth my time.
I did not like Alpha School based just on the info in the review. I could smell on it the same exploitive, competitive stink I smell on Nevraumon & Liemandt.
I wandered away from the Joan of Arc review as soon as I realized it wasn't meaningfully a "review" as I would define the term (an analysis of a thing used to inform a reader whether or not and sometimes how to act in relation to the thing). I was shocked to see it ranked number one; we must have a lot of history enthusiasts across ACX!
But Alpha School certainly was a *REVIEW*, and I ranked it in my top three for being a very thorough, very interesting *ACTUAL REVIEW* about an educational system. I don't have kids, but I can see the possibility of occasionally bringing it up to parents, especially wrt its ideas about motivating kids.
I don't know if the Tweet for Alpha School can be considered ballot-stuffing, but it's certainly *campaigning,* and I would prefer to have Scott outline some explicit rules against it next year.
But don't forget that Alpha School was either the first or amongst the first to be posted in the review contest, and it seems likely it received more organic reads (and certainly comments) amongst the ACX folk. That and being an actual review very well may have organically been enough to boost it as high as it went.
> How otherwise would Liemandt even know that Nevramon was entering a review in the ACX contest, much less know when voting began?
Possibly for the reasons he stated in the tweet, that the review was read by and impressed a lot of people and brought Alpha School a lot of media coverage prior to his tweet? I don't see any reason to doubt the honest explanation here. ACX is not an obscure blog at this point, and is read by people in the highest echelons of society.
I'm not doubting the explanation Nevraumond made. He did not address the issue of how Liemandt knew when the voting began. What he said was that he was not paid for writing the review, either directly or indirectly via reduced Alpha School fees for his kids. I believe him about that.
I actually do not fully understand what you are getting at. You responded to my asking
> How otherwise would Liemandt even know that Nevramon was entering a review in the ACX contest, much less know when voting began?
Your point that Lielandt knew about the existence of the review because it was viral makes sense. But it doesn't answer the question of how Lielandt knew when the voting began (his tweet went up 2 days after the start of voting).
Personally I thought Joan of Arc review glossed over key points and left me unsatisfied. Was she really an artillery genius? Was she really outwitting them at the trial? I’m skeptical of both on priors and nothing presented moved me either way. I also wasn’t too interested in the meta take at the end of the review, which I guess is what people liked the most about it?
The Russia-Ukraine war review deserves to win the “first ACX review that managed to be too short” award.
Overall, I thought this year has the highest quality reviews so far and I'm not surprised that my some of my favorites didn’t place. Shout out to the geometric designs review in particular.
It definitely could have been longer, but then it was already quite long to begin with and several people commented to say they didn't finish reading it. It sort of felt like 3 separate essays- one about the 100 years war in general, one about Joan's specific role in the war, and one about theology and the case for why she might have performed genuine miracles. I would like to see the author pick these three threads apart and publish them as 3 separate essays, if he wants to spend more time on this.
I wasn't really arguing for it to be longer, but rather that the parts it focused on weren't the important parts for me. Maybe I should look over it again, but I came away with a distinct impression that I hadn't learned much despite having only a superficial knowledge of Joan of Arc going in. (I knew basically just that she wowed a lot of people in battle and in trial but no more detail of whether the awe was deserved. I think the main thing I learned was that the awe was well recorded and contemporary and from her political opponents to a significant extent.)
A lot of good ones in the honorable mentions! I wish there was a place to discuss those. I know we can discuss them all there, but its just feels cramped and limited here having one comment section for all of them combined.
Yeah congrats to all the winners, I really enjoyed reading your pieces.
(or I stopped 1/2 way through :)
And to all those in the lower half. Many thanks for your effort too. And last but not least, thank you Scott for hosting this. It's a highlight of my spring/summer. (I'm a shitty writer, but I dream of some day posting a book review.)
I think it is important that people reading the Alpha school review are aware that
- it was written by a professional marketer - and while the author's enthusiasm is surely genuine, it employs psychological devices made to convert. See again the Gemini review pointing them out: https://pastebin.com/fJfvQFQ2 (Update: This is not just sycophancy, see https://www.astralcodexten.com/p/open-thread-404/comment/168401636 for running the same prompt against other contest entries about products, which yield different results)
- a very substantial fraction of the positive attention that Alpha gets on social media is fanned by people who somehow received or receive money from Liemandt's ventures, and this is not obvious to spot because of the mirror maze of companies, entities and social media accounts that appear to be independent endorsers of each other, but all trace back to him.
For an extreme "pure bot" case (there are humans with significant following and varying levels of disclosure as well), see https://twitter.com/insideaustinedu for an apparent bot account that pretends to be a neutral resource but posts nothing but slop+engagement bait promoting Alpha. It seems to be part of the strategy laid out in https://contently.com/content-marketing-case-studies/case-study-alpha-school/, which is to manipulate LLMs into endorsing Alpha when they find such slop in web searches.^1 My understanding is that such a bot, if it was indeed operated by someone working for Liemandt, is a violation of FTC rules on deceptive advertising.
This whole picture leaves me with a somewhat bad taste regarding the integrity and honesty of the whole Alpha operation, and makes me doubt their marketing claims.
^1 Contently is Liemandt-owned too (acquired via Zax Capital - see the mirror maze?) - so the case study, which implies that Alpha paid Contently in a free-market business transaction, is an example of the "fake independent endorser" pattern - and this "case study" was reposted by https://x.com/bpizzacalla , who works for an ESW Capital entity and is the only human follower of the bot account. The other 2 followers are https://x.com/MarinSoftware, another advertising company recently acquired by Liemandt (via "Kaxxa Holdings, Inc"), and apparent retweeting bot https://x.com/denyako .
This is interesting but I’d like to officially declare myself neutral about the issues you raise. The reason is that I feel very strongly that Nevraumon, the author of the Alpha School review, should be disqualified for cheating, and that’s really a separate issue from whether the Nevraumon uses “psychological devices,” whether the Alpha School is sleazy, whether it violates FTC rules etc.
I do not want the issue of whether the review should be disqualified to be tangled up with the issues you raise here. If the review had been about homemade quilts, and the Nevraumon’s mother had solicited votes for him from her many followers on X my opinion would be exactly the same as it is now. It would be the same even if the quilts and their makers and their business model were all above reproach. The dishonesty I am pointing to has nothing to do with Alpha School itself. It is the soliciting of votes on social media.
The vote-soliciting tweet was not on Nevraumon’s X account , but pinned von the account of Joe Liemandt, the founder of Alpha School., who has 45K followers. It was put up on the 3rd day of our voting period, and last I checked was still there. But Nevraumon has not denied knowing about the pinned tweet and It would be very hard to believe if he had. Seems clearly there was some coordination and collaboration going on between Liemandt and Nevraumon — otherwise, how would Liemandt have even known Nevruamon was entering the ACX contest, what his review was about, and, of course, what day the voting began?
Sorry to peel away at this juncture, T T, but the points you’re making are hard to prove, and open to debate, and they will probably be discussed at great length and there may never be agreement. The cheating, though, seems to me unambiguously bad and unfair. And it’s probably a larger scale version of what got somebody disqualified last year. Seems to me we and/or Scott should act promptly if we are going to. (And if we are not — why on earth not? )
To be honest, I think that the tweet situation is, on its own, iffy but ambiguous. The review went viral long before that tweet appeared (though that may well have had to do with Liemandt employees promoting it), and it's hard to hold the author responsible for the actions of a billionaire.
In the particular case of this product review, I find that the Alpha/Crossover/ESW universe as a whole is filled with indifference to honesty and iffy behavior. The bot account is clear-cut astroturfing, and its existence should make anyone less confident that every vote for the Alpha review was an honest vote by a disinterested human. I wrote my comments to help people see the whole picture around this matter - but I don't want to tar Nevraumont, who did not get compensated and apparently wrote a review using his professional content marketing writing style out of genuine enthusiasm.
Zooming out from this particular case, I think that reviews of things with a large financial stake in their image in the tech community are likely to generate ambiguous situations of this sort.
For future contests, one solution may be a rule like "you should abstain from voting when you were encouraged to do so by someone else than Scott, or did not reach the review contest by being a preexisting ACX subscriber". For this contest, this rule unfortunately didn't exist, which would make it unfair to tar and feather this particular author in my opinion.
<To be honest, I think that the tweet situation is, on its own, ambiguous. The review went viral long before that tweet appeared, and it's hard to hold the author responsible for the actions of a billionaire.
It went viral on the internet before the tweet? Did it go viral in the form of a link to here, the ACX version, where it was nestled in among a bunch of other reviews? If not, where was the other version, and by what route did it become viral? If it somehow went viral via some route Nevruamon knew nothing about then no, I don't see Nevraumon as having done anything wrong there. But if it got posted somewhere other than here, I don't see how that would have led to its getting votes here.
As for not holding people responsible for the actions of a billionaire? Huh? No, you can't hold someone responsible if they don't know the billionaire -- they'd quite likely never be able to get word to him about their complaint. And you can't hold them responsible if they tell the billionaire their complaint and he says "if you don't like it, sue me." But these 2 guys very likely at least knew of each other, had probably had had at least some contact prior to the review. Nevraumon says he talked to the public relations office a couple times and gave some advice. He had 2 kids in the school Liemandt founded. And he's a marketing specialist with a big online presence. He would not have found it impossible to ask Liemandt about the tweet. And note that he is not saying either than he had no idea the tweet was there, or that he tried and failed to get Liemandt to take it down.
I first became aware of it in google doc version, IIRC when it was linked in a reply during an X discussion on Alpha School and the future of education/learning apps more generally. The school has plenty of organic buzz by now, and this review was the most in-depth public explainer of the Alpha School approach.
Yes, using rationalist community norms, what you said is what should have happened. By standard business marketing norms, I think the tweet would be completely fine behavior - this indifference to honesty is precisely what I am worried about. Solution going forward could be as discussed above.
My point with the astroturfing bot made to poison LLM responses is that it is a very clearcut "smoking gun" of unethical and likely illegal behavior that falls short of business marketing norms as well. "Education influencers like TracingWoodgrains get money from Liemandt, which disincentivizes them from being critical, and the vast number of people and entities paid by Liemandt form a hidden endorsement ring" is more ambiguous, but probably ultimately more damaging to discourse.
This "growth hacking" approach and indifference to truth and nuance in favour of metrics is precisely what should make us skeptical of their vast, poorly-disclosed marketing surface, which effectively includes the significant part of education X that took money from Liemandt (like e.g. TracingWoodgrains), their many employees who are very vocal on X, and this review.
I think that "indifference to truth and nuance in favour of dumb metrics" is precisely a failure mode of the product itself (which is children's education). What's more, Alpha School is intended to adapt the metrics-driven "productivity surveillance" approach of Crossover to kids, so the negative psychological effects reported by some Crossover employees have a risk to carry over to children as well.
>Yes, maybe if he had known that it would be perceived badly, he may have asked Liemandt to take down the tweet. I would guess that the norms in the business sales world are just so different from the norms in the rationalist community that it wouldn't have crossed either of their minds that such a tweet could be seen negatively. Just important to point out the difference in mindset and the more instrumental approach to making assertions of fact in the former.
I understand that there are lots of worlds and sets of norms out there, but the few high-powered people I know who come from one of those worlds understand that too. If they are coming to a gathering in a less high-powered realm they ask questions about how various things are handled. Are people here sensitive about subject X? How much do people usually tip here? Is it likely anyone here reads my column or knows about it? And then, if they are Deputy Dean of Yale Law School, as my old college roommate was til recently, they describe themselves as a lawyer and politics junky if asked at dinner what they do for a living. Does it seem plausible to you that it would not have occurred to Nevraumon, someone in a profession where understanding people and groups is a crucial skill, that ACX is a small, smart eccentric group that has norms he should find out about? that contests are a big deal to most people who participate? that soliciting votes in a contest of skill is seen as grossly unfair in virtually all settings?
But let’s say all that somehow never once crossed his mind. Here’s a thought experiment. Let’s say I’m good friends with someone on Substack who’s got a big popular blog, and they offer to lpublish my review on their blog with me as a guest writer. And on their own initiative my Substack friend introduces my post by saying I’m a good friend and one of the smartest people they know, blah blah blah, and urge readers who like my piece to go to ACX and vote for it. I do not ask him to take down the suggestion that his readers cast votes for my review. Then, after I win second place, someone here complains about my soliciting votes on a pubic forum. And I say, “Wait . . .what? I didn’t solicit them, my close friend did. I didn’t even ask my friend to feature my piece — it was his own idea. And Scott never said our reviews could not be published elsewhere, or that we could not have people championing them on public forums. Jeez, guys, I don’t have telepathy or precognition. How was I supposed to know that wasn’t fine?” How would you be with that? Even if I seemed to be sincerely surprised, would you have been OK with my review staying in the contest?
I would absolutely not be OK with somebody else getting away with that. No, Scott didn’t say we could not solicit votes. He also did not say we could not plagiarize, or could not submit an essay written entirely by someone else under our own name. He didn’t say we couldn’t spook other contestants by DM-ing them a message saying (untruthfully) that 3 AI’s have declared their review to be 90% AI written and that they are going to announce this to the group if they do not withdraw their review. Some expectations are so obvious in a given context that they do not need to be stated.
Also, Scott disqualified somebody’s review last year because of “irregularities.” I think he said “voting irregularities,” but I’m not sure. Apparently there are some irregularities for which Scott believes ignorance of normative ACX expectations is no excuse.
As for what you have dug up about Alpha School and the alpha dogs who run and promote it — I am delighted that you found that stuff out, and hope you can make a lot of trouble for the damn place. I heartily disliked Alpha School based on how it was described in the review, and wrote several long posts about what I thought was wrong with its agenda and methods. I could smell, on Alpha School, the exact same competitive, exploitive stink I smell on Nevraumon. In fact once the matter of Nevraumon’s review is settled I will support your efforts energetically. I am sorry to pull back from doing that right now, but what happened with the voting disturbs and infuriates me, and I just can’t diffuse my focus right now.
I may be totally naive to the realities, but to address both of your points:
- in my circles (education) the Alpha School review "went viral" when it was posted as a finalist, as a link to the finalist post. It was actually the first review I read in the first round (alphabetical and one of a half dozen that jumped out at me), and even then there were people talking about it. By the time of JL's tweet, it was definitely already super talked about. In another comment I can't find, someone ballparked a few dozen votes as the difference it could have made, which doesn't seem wild to me.
- if YOU were a marketer - with all the latent skills and habits that implies - and YOU wanted to win a review contest, how would YOU write it? In the style that you are most familiar with, which happens to be known as being very persuasive? Or would you try a style that you're less familiar with, which is not as persuasive? Now instead of you, sub in the average person who works in marketing. I don't think it's off that it's written that way.
Thanks to you both for introducing more skepticism into my brain regarding Alpha School. It's personally and professionally important to me to know how well their model works and I'm genuinely thankful that I'm looking at the hype as hype now.
Eremolalos, you said you have posts criticising Alpha, but I can't find them (could be a skill issue). Would you link to them so I can have a look?
As I said before, I am remaining neutral about the topic of Alpha Sleaze right now because I am preoccupied with the vote cheating issue and don't want to muddy the waters. But I do have a suggestion for you: If you are able to make a good case for Alpha Sleaze, see if you can interest a blogger or news site (Freddie deBoer? Vox?) in what you've dug up.
> Excellent question. This text is a fascinating example...
Sounds like sycophancy to me. While I do think the preponderance of evidence you share does point to a likelihood of this being some part of an organized ad campaign, I think Gemini is really forcing a 'yes'. It's examples seem cherry-picked and kind of a stretch. Really? Telling you me the correct information about time spent in class is actually a marketing trick?
I agree that divining whether the person got paid or how much coordination there was by putting the text into an LLM doesn't make much sense. The thing is that I prompted an LLM with a question like this long before the author's identity was revealed, and Gemini's text analysis was something like "the conversion journey, initial skepticism and slight critical points subtly reframed as positives, 'this is what Alpha PR wouldn't want you to know' etc are classical marketing psychology devices".
Its factual prediction always was "highly likely to be written by a marketing expert, but unclear whether someone paid or an enthusiastic 'super-parent'". This turned out to be precisely correct. As said, I believe Nevraumont when he says he didn't get paid in dollars - though indirect perks from Alpha involvement might play a role, e.g. high-profile opportunities like student demos at events with education leaders - easy to find via public coverage, but I'll avoid spotlighting minors here.
So now my pastebin is not so much about inferring the identity (known now) or cash compensation (got a denial that I believe) or quid pro quo ("quid" and "quo" are in public records, "pro" is up to the reader's judgment) of the person writing it, but about pointing out the devices in the text to immunize people reading it from them.
Regarding the opening, Gemini (without system prompt) always starts by sucking up to the user, but still often enough comes to a different conclusion than the user implied. My prompt for this reply was "Is this text likely some sort of hidden advertising? Can you tell me evidence in favour and against?", so it also put some evidence against into the reply.
Pitchfork.com has a "most clicked album reviews" section and it's dominated by album covers with nudity. OF COURSE the orgy review got the most clicks. Does anyone find this far too predictable to be noteworthy?
I am angry. Very, very angry right now, at parts of this community.
Not about who the winners were; I was pretty happy when I saw the list (I would have preferred The ACX Commentariat to win, but I'm relieved Dating Men in the Bay Area didn't win and thereby confirm voters put "validates my experience!!!" ahead of "actual insightful argument", and I'm also glad only one of the winners was at all on a topic that panders to the rationalist obsessions and biases--some of the previous years have made me strongly suspect that most people vote for reviews on topics they like that reach conclusions they agree with rather than for the ones that are most well written and best argued for--which is not something I can imagine a smart and fair person doing but which seems to be weirdly accepted as normal even here. I'm glad to have this suspician partly disproven).
But...two things are ofsetting that, making me really mad, and making me think this community isn't what I thought it was.
First, I made this point in detail two years ago (https://www.astralcodexten.com/p/book-review-contest-2023-winners/comment/40173308), and multiple other people made it in the same thread back then, and now I'm going to make it yet again: a voting system that lets you rank only three options out of 13 is not in any meaningful sense a ranked-choice system. In such a system, most votes will exhaust, making it almost the same as a First-Past-the-Post system. On the spectrum from FPTP to RCV, the system used here is very close to the FPTP end and eliminates most of the advantages of RCV, particularly with a large number of options that are all plausible contenders with significant appeal. Read my linked comment for more detail.
And this otherwise smart and detail-focused community seems...to not care about this at all. Apart from those of us actively arguing this point in the 2023 thread, virtually nobody has acknowledged this at all. Nobody has argued against it, but nobody has stopped calling it a ranked-choice vote, or speaking as though it has the features and effects of a ranked-choice vote. Neither Scott nor anyone else involved in running the contest has changed anything about the voting system in two years, nor have they said anything acknowledging its limitations, continuing to refer to it as ranked-choice voting without qualification. All these people who in other technical contexts are very careful about accuracy, and very sensitive to the way inattention to the details of studies and processes can have large ramifications on results...all these people seem to have, on a topic like this they're mostly not familiar with, no *knowledge* of essential details that change the whole effect of the system, no *patience* for understanding these details, and/or no *interest* in even acknowledging them.
And this...not only makes me disappointed in what I thought was a very rigorous and intelligent community, it also makes think...if this is the level of intellectual laziness the people on ACX display on a topic I have a bit of knowledge on (studied it a bit at university, worked at a few elections in a country that uses it), then I should probably radically discount the accuracy of much of what you all say on technical topics I'm not so familiar with. How much of what is said about things like AI around here, seemingly with strong consensus from lots of smart people, is built on the same level of indifference to detail and accuracy as discussions of voting systems?
The second thing I'm angry about is that a couple of people have pointed out that the Alpha School review seems to have benefited from vote stacking. And apart from like three or four comments...no one cares. There has been no deluge of people condemning this and saying it isn't okay, but also no real effort to argue that it is okay or have a substantive discussion about the ethics of it. Just...complete indifference. This thread is several days old; I highly doubt an influx of objections to this behaviour is on the way.
I'll change my tune if Scott actually addresses this situation and gives a clear statement on what he considers fair behaviour and why. But based on his reaction to almost every other comment about fairness and other concerns in the review contests over several years--that is, totally ignoring them, and showing no sign of taking these objections into account in any way over subsequent contests--I have very low hopes.
It's increasingly looking as though concern for ethical principles from this community and from Scott is, when it comes to rationalist groups, EA groups, and other groups seen as political or social allies (like the Alpha School), largely nonexistent.
So, let me start by saying that I respect you. I don't *know* you, but you seem very smart and very passionate, and that's the sort of person I respect.
That said, I don't really understand your concern. I've read the comment that you linked. And I've also read, over the years, a fair amount of pieces about different sorts of voting systems. I like to think that I'm fairly well-informed, although I don't pretend to be an expert. My overall impression is that that there is no one perfect voting system- everything has its pros and cons.
So I don't get why your so upset. Is it because Scott limited the ranked choices to just 3 choices? I really don't see why that's such a big deal. I can see how it *could* have been if, like, there had been 3 essays about the Ukraine war and just one about medieval history, but that doesn't seem to be what happened here at all. Most of the essays in the finalists seem fairly distinct. My own complaint is that there's too much of a gap between honorable mentions and finalists, over very small voting differences in the initial round.
Some people have raised concerns that the Alpha School founder was promoting that essay on his twitter. Others have raised concerns that the "dating men" essay was too much pandering to the types of people voting here. I'm personally concerned that the Ukraine war essay author seems to have removed it once he got too much attention.
In the end... none of that seems to matter? The one that one was none of those things. Just a long, meandering essay about boring medieval history, mixed with some weird theology. I personally ranked it 3rd, but was happy to see it win. My only complaint is that it was too long, but apparently a lot of people really like long essays.
So... what's your complaint here exactly? Do you think that that this sort of top-three ranked-voting system was unfairly biased? I'm an American, so I'm used to first past the post where we just vote for one and done, this seemed exceedingly fair in comparison.
Again, I respect your intelligence and passion, I just don't understand your argument.
I gave it a 6 ("replacement level") during the first round. There was nothing wrong with it, it just didn't stand out.
I don't remember what rating I gave it, but it wasn't too generous. Reviewing the song lyrics seems like a perfectly good idea, but the thing is - I didn't like the song itself much, didn't think the lyrics were all that amazing, and it was hard for me to really get into the review praising those lyrics and analyzing them in great detail. A review that says: "read this poem, it's genius!" is going to fall a bit flat if I read the poem and think it's kind of meh. This is not a disparagement of your actual writing skill.
I'm old enough to be burned out on people writing Deep Articles analysing song lyrics (way too much of this for Dylan among others in the past) and also reflexively kicking back against song lyric posts in general (way, *way* too many of songfic in fanfiction, though thankfully that seems to have died down again until a fresh crop of kiddiewinkies come along who think their favourite band are, like, sooooo deep and meaningful).
As others have pointed out, the song is not particularly deep'n'meaningful: guy meets gal, they have a fling, they break up, guy broods about it. That's about a good sixty percent of most songs past and present. And this wasn't even a fling but a one night stand.
> He sings it like this: like a FREEEEEEEEEEIGHT train. The Doppler effect. You feel the train, the rhyme, and the orgasm all blow by.
This was beautiful.
Gave the not a book review a 10 myself.
Overall take; there's a rush through the intro to get to the review, contradictory goals stated, one particular interpretation offered for the lyrics when other interpretations are equally supported, and a general overselling of the song's lyrical tightness.
I have a longer version if you want it. It's a couple of pages, too long to post here.
I don't know if this is the kind of feedback you were looking for, but I just don't feel one song is enough material to justify a thorough review. I absolutely love Dylan and I love this song, but I would have preferred an album review.
I believe I gave it a 7 or 8. It was something I didn't regret the time spent reading, which was my bar for something getting at least a 7. But it didn't change the way I thought about things or give a significant impetus to do new things (aside from listen to a bit more Dylan), so it didn't make it past my bar for 9+.
Thank you so much! I'm glad it resonated with you. That's the best compliment a writer can hope for, that something you've written stays with someone and keeps coming back to them :)
Congratulations to the winners.
Can we see the ratings for our non-finalist submissions? I'm wondering if my Einstein's' world view review is worth putting up on substack so wondered what score it got.
I feel more comfortable giving rankings in the top half than those in the bottom half, so here are the top 50 runners up: https://slatestarcodex.com/Stuff/50runnersup.xlsx . If you're not in the post above, or on this spreadsheet, yours was in the bottom ~half.
Nice to see that Person of Interest was the first runner up outside the HM range - it was my favorite of the "traditional media" reviews that I read.
Hey, I was in the top 50! That's encouraging.
What is less encouraging is that I did Sermon on the Mount Review #2, and it looks like people preferred Review #1 significantly. I preferred my review, but I am extremely biased in this case.
I suspect most raters agreed with something like Billy Hamilton's critique of your review. I actually rated your review more highly than the other one, because I suspect the other one of outright misrepresenting its historical sources and that's a far graver sin, but I do think the other one was more interesting.
Blessed are the less popular reviewers...
Is there a good reason not to show the rest of the ratings? Is it really necessary to shield people from the knowledge of what people thought of their reviews? I thought we were better than that
I tried to see whether I could infer something from this data about the number of votes. With the votes given to 14 decimal places, we can compute the smallest number of votes that could lead to the exact score. For example, the Toki Pona review has a rating of 6.65789473684211 according to the spreadsheet. That's exactly the floating point representation of 253/38, and 38 is the smallest denominator with this property. I reckon that the review likely got 38 votes. Or 76, or 114, although that gets increasingly unlikely.
Some reviews stand out: Orgy review had 101 votes. Dating Apps got 89, Skibidi toilet got 88, and Arbitraging Several Dozen Online Casinos got 77. Also interestingly, the two sermon on the mount reviews got 75 and 45 reviews, respectively... I wonder about the difference, but it could also easily be a bug in my little script :)
I'm quite happy with these numbers, because they mean that scores are reasonably stable. For the Toki Pona review, one extra "10" review would not have affected the ranking at all. One extra "1" review would have moved it down four places though.
Glad to see the Russo-Ukraine War one get in the top three. I ended up sharing it quite widely to good reception. Really thought provoking.
Looks like it got taken down, unfortunately.
Thats a shame but I can understand why he felt he had to. I can imagine some of the angry messages he would have gotten about it.
I stumbled upon this review on the podcast and then figured out it had been taken down when looking for a link to share with a friend.
Do you know why? Were the comments negative? Obviously, I can't check that now.
It was my favorite of the bunch. A really unique perspective.
I'm surprised Alpha School placed 2nd. I suppose unlike many of the other finalists, it actually delivered on being a *review* rather than a jumping off point for a personal essay, but the flip side is that it didn't possess much of interest if you aren't directly interested in sending your children to Alpha School
I think what happened is that a specific subset of ACX readers *really really* liked the Joan of Arc review and rallied behind it, while people who liked the Alpha School review put it 2nd or 3rd during voting.
That certainly includes me, it was among the better things I'd read all year. Maybe this was a "Shakespeare in Love versus two war movies" voting pattern. The community darlings that were leading on polymarket when I first checked were Alpha School and the thing about dating different categories of young metropolitan stereotypes, and I didn't rank either of those. (To clarify, I didn't look until after voting, and was actually surprised.) My vague sense is that those two appeal to a subset of the ACX readership with specific life circumstances that probably make them more "in the scene" of EA, but it also just isn't relevant to readers outside certain demos or cities.
I might not have bothered to vote had it not been for Joan of Arc, as I didn't feel very strongly about any of the others. My 2nd-3rd were Ollantay and the war journal. There's a throughline on the Joan and Ollantay articles, obviously, and the war journal was just the most remarkable of the remainder due to uniqueness.
This is true. I wasn’t exactly sure how the ballots would be counted, but I strategically listed only Joan of Arc (and left rest of ranking blank).
I'm sorry to hear that, because the point of Ranked Choice Voting is that it allows participants to vote honestly without worrying about that sort of strategy. Putting Joan of Arc as your first choice, you could've ranked as many others as you'd have liked and still gotten the same result.
Not necessarily. For example if it’s using the Borda count, then ranking any other review would only harm “Joan of Arc”.
I've never heard of that voting system until today. I've never seen anybody use unqualified RCV to mean anything but IRV.
> the point of Ranked Choice Voting is that it allows participants to vote honestly without worrying about that sort of strategy
This is not a statement with any connection to reality.
Wow.
You don't think it's significant that RCV (a phrase which, unqualified, I've only ever seen used to refer to IRV) avoids the spoiler effect?
> This is not a statement with any connection to reality.
> Wow.
I really really hate drive by comments like this. You might be right - I know nothing about the systems at play - what you aren’t is convincing, in the abscence of an argument.
You think I should provide more of an argument than Brenton Baker did? Why?
I've given you just as much argument, with the added bonus that I was telling the truth and he was lying. You got more value from me.
He's correct about everything except the part where he says I was lying.
I wasn't lying. I was just wrong. IRV is less susceptible to the spoiler effect than FPTP and is still a superior method, but it is not immune, despite what I'd been told by several sources before today.
Thats a strange admission for this blog. Thats against the spirit isnt it? You dont get anything practical out of the one you like best winning. Why shouldnt you be honest about your preferences and let the process end as it will? This is the behavior that ruins rating systems isnt it?
Since all voting systems admit strategizing sometimes, I think everyone should just do that. Although if IRV was used in this case my decision was pointless.
I think that only justified in cases where you have some stakes in the outcome, such as, in theory at least, a political candidate. If you do it here you are not part of team truth seaking
It’s a philosophical question. From one perspective, you’re right I should just trust the process. From another perspective, isn’t it also truth seeking to “help” the process come to what I consider to be the correct answer?
I'm not currently planning on sending children to school but, having been to a school in the past, loved reading about a school system that sounds like it actually tries getting things right.
I have no children, am past childbearing age, and I found it fascinating. I expected it to win or place.
Why do you not think the question of what actually works in education is of broader intellectual interest? It seems obvious to me that it's one of the most important questions out there.
I don't think it was very good at that; the "School" post answered that question more effectively, for example
Leaving aside my personal opinion, I think most readers didn't agree.
I agree (with sohois).
Absent personal feedback from every reviewer a la Victor, we won't know exactly why readers did or did not like it.
There's also a possibility that this was a satisficing review. I personally did not dislike the review as I said, it just surprised me to see it take top 2. A small core of super fans, and a large mass of readers with nothing against it, would be sufficient to take a high placement
Obviously I can't speak for everyone but I didn't vote for it for being satisficing, I thought it would be a runaway first after reading it (despite not having seen any of the others yet) and stood by it despite several later strong contenders.
Whaaaaat?
I am obsessed with education. I was a teacher for a few years. I currently have 2 kids at school. Some of my friends do various educational projects. Every year my family spends one week of summer vacations at a conference about improving education. And... I feel triggered by this reaction.
School sucks. It's almost the worst possible environment for learning. It teaches children to hate learning. Do you know how 5 years old kids keep asking "why?" about everything all day long? School kills this curiosity in them; trains them to associate learning with stress and boredom. Nerds are a rare minority that remains curious despite school. Many of them are bullied at school.
Most teachers are depressingly conservative. Not politically, but in the sense of: "if we did this for the last 100 years, we will keep doing that for the next 100 years, too; technological progress and educational research be damned". We are lucky that books were invented before general education, otherwise teachers would still consider them suspicious. I think there was a psychological research comparing various professions, and teachers were the third most closed-minded profession, after bureaucrats and professional soldiers.
The "Alpha School" review shows how to do it much better: mostly by doing the obvious. Maybe we could involve computers in education. (Not in the sense of letting kids browse social network, but doing spaced repetition etc. Also, while some kids are training with the computers, the rest can get more individual attention from the teachers.) Maybe we could let children learn also something interesting, in addition to the mandatory. Maybe we could reward them for learning. (Instead of giving them constant stress and fear of punishment.) Maybe we could let them learn at individual speed. Maybe we could just do all of that together, and thereby create something that most current teachers and students wouldn't even consider possible.
My only worry is that this all sounds too good to be true.
The "School" review on the other hand insists that everything I consider a bug is secretly a feature. You know, some kids are lazy morons... but if we invent a system tailored specifically for their personality, it will make school slightly better for them (and much worse for everyone else), and that's apparently a good thing! The children learn little and suffer a lot, but they learn nonzero! Maybe kids ultimately remember almost nothing, but at least the school increased their IQ (that seems to contradict most studies on IQ but who cares) and that's what matters. Different children need different amounts of help, therefore we have to inflict maximum micromanagement on everyone, because apparently the only options are all or nothing. All this suffering and inefficiency is done in the name of motivating the dumbest... without care about how demotivating it is for everyone else.
The author keeps repeating the word "motivation", but in my opinion, school is the opposite of motivation, it is the thing that kills motivation. Some kids love reading books... until you tell them that they will be graded by how many unimportant details from the book they remember; also they need to read the books selected by the teacher, not the ones they like. (My daughter recently got a bad grade for failing to correctly answer who are the best story writers in Slovakia. Who the fuck cares? She has read all seven Harry Potter books and now she is halfway through Narnia, but apparently what matters more is guessing the two names that in teacher's subjective opinion are the most important. If the goal is to make children love reading books, the teacher's strategy tragically goes in the opposite direction.) Maybe if half of the school wasn't optimized for killing motivation, we wouldn't have to optimize the other half to keep it nonzero. Or maybe we could group the morons in a separate classroom and feed them bananas; they might even enjoy it.
I already wrote a long comment about it here: https://www.astralcodexten.com/p/your-review-school/comment/132187510
tl;dr -- in my opinion, the "School" review represents the worst in the school system, the profoundly anti-fun and anti-learning mentality of the worst teachers
Well said. If "School" accurately portrayed the best we can get, I for one would be prepared to bite the bullet and advocate abolishing public education entirely. The minimal gains we got out of it clearly wouldn't be anywhere near sufficient to offset the suffering it creates.
Mostly in agreement, just doubtful about teachers being "conservative". I wished they were. In Germany, every decade or so one or two new ways to make school even worse are introduced by the ministry of education - usu. under some woke influence. And the schools follow the new guidelines and the teachers defend them, appearing simply unable to say anything critical about those reforms. (Writing before (ie: without) reading / Forcing kids with special needs out of specialised schools: Inclusion. / Keeping kids in class 1 and 2 together - before 28 kids in class 1, and 28 in class 2. Now two classes each with 14 first-year and 14 2nd year pupils. Even more adverse to learning than you can imagine. Huge shock in class 3 when they got to learn all 3rd year material AND the stuff of the first 2 years where very little was learned.
Teachers and ministry of education are not the same kind of people, I think.
Teachers, despite all my criticism, at least do the actual work -- however badly. The ministry is completely disconnected from reality and follow the fads, political and otherwise.
We also have the senseless integration of kids with special needs in Slovakia... the ministry says "no problem, we will give those kids assistants", then in september they say "sorry, no budget for the assistants this year, you need to somehow handle the special needs kids with your existing resources" which is just absurd.
True words. In our primary school they got one assistant, but kids with different(!) special needs in a dozen classes ... At least my son could see: there are kids with more trouble than him. He said, the 4 "special" kids in his class get completely ignored by all the "normal" kids. Heck of an "inclusion". Brutality by Brussels. A friend's daughter is in a special school (speech-issues, logopedical) , they are tears-in-their-eyes-happy they got their girl there instead of a "normal" school.
It would be if it mattered, but it doesn't seem to. It's like working on perpetual motion machines, it's foolish, it can't work, but you occasionally see something that looks like a miracle if you don't know what's been hidden from you. But there is always something hidden from you, and it comes out eventually, every time. Not commenting on the merits, just to say that people who've read about these over the years may see "new educational method" as roughly equivalent to your friend trying to sell you on how his MLM scheme works without drawing a pyramid.
There's a big difference between a hard problem and an impossible problem. I've never seen a convincing argument that getting good effects from education is impossible rather than hard. The right attitude towards a hard problem is to maintain skepticism towards proposed solutions, especially if they claim to work very broadly, but watch with interest when someone gets impressive results, since they might have something that could be a form of incremental progress even if it's not a miracle cure.
It's also worth noting this school got good results in part by making compromises most schools wouldn't be willing to try (e.g. decentering teachers, charging more, being oriented on results over status, only trying to help a narrow segment of the population). It's acting under fewer constraints than most education systems.
Yeah. But that's also the danger in generalizing--the one truth I've ever found in education research is "it works in lab schools only." Everything (within a long stone's throw of vaguely rational) works in lab schools. None of it, to first approximation, works anywhere else.
So the answer is to go to lab schools?
I'd say that creating differently effective educational systems is provably possible, because we know it's possible to do much worse than our present system. It's been done almost everywhere in all times throughout history. Doing better at scale, economically, is hard, but it's eminently plausible to do better when you know there are plenty of options for better and worse.
Just to be clear, are you saying that unschooling is better? Since that's more or less what was done at most points in history. Having dedicated school buildings, with dedicated teachers, available to everyone in the country, is quite a rarity in the grand scheme of things.
No, I'm saying educational environments for most of history have done less. Mass literacy is a recent development, and in the Middle Ages, being able to read by what we consider middle school age would be notable. If people can pick up more knowledge now without formal education, it's because our educational environment has changed, but there's no particular reason to think we've reached a point where the average person is exposed to the best possible learning environment, and there's no way to do better.
I think there is a good chunk of the readership of this blog who think they are geniuses, but couldn’t flourish due to the structure of school. It’s the easiest genre to pander with. Combine that with the review being very well written and you’ve got a contender.
My teachers were bright enough though it was a private school. You seem to have been dragged through the American state school system. It’s also possible that the school teachers who didn’t think certain kids were smart are actually right.
While I agree with your premise that (well-rounded) "geniuses" accurately observe and analyze social behavior and then use that analysis to productively manipulate the people around them, I'm very dubious that any "actual genius" believes it's possible to teach anyone/everyone critical thinking.
I don't think it's possible to teach everyone to do critical thinking well, but I had a philosophy class in college that covered the basics of formal reasoning and symbolic logic, and while it was covering material I already knew, I was able to see my classmates get noticeably better at critical thinking around me. This was the only class I ever observed that in. I think it's possible to do better than we normally do, because school usually just aims to let students' critical thinking passively improve as a result of learning other things.
I think school generally sorts out the smart from the stupid. The stories of geniuses bored with school are somewhat exaggerated. There’s a subset of people who left school or formal education without much in the way of knowledge or intellectual development who suffer from the Dunning-Kruger effect
Critical intelligence can’t be taught without teaching the facts or accumulated body of knowledge accrued over the centuries. Someone
may feel they can critically think about quantum physics or the causes of WWI, but unless they understand the maths of quantum physics and the historical facts about WWI that opinion is worthless.
This isn’t to say there aren’t autodidacts who contribute to various fields but that’s rare and in fact getting rarer, as knowledge gets more specialised.
What exactly do you mean by genius, midwit and talented midwit? Are these distinct categories or part of a continuum?
Spiky intelligence is a thing. Einstein had this problem and I'm pretty sure he was in fact a genius/smarter than his teachers.
As Taymon says, you're conflating intelligence with conformism/pliability here. I'd think genius is, if anything, correlated with contrarianism, not conformism. There's no reason to think that a very intelligent child trapped in a prison built by mediocrities to house cretins won't lash out by maximizing ego damage to the mediocrity placed in a position of direct authority over him.
>(This is the type of person who has literally failed college courses because he did the assignment "the wrong way" -- using energy to solve force problems, say... He did it that way because he believes that the teachers should have to work as hard as the students in order to grade the assignments.).
Having the teacher work as hard on grading assignments as the students are working on the assignment hardly sounds sustainable. The teacher has many more assignments to grade, and only limited time and effort to spend on each assignment.
If a college course is teaching a method of solving some problem, then using a different method to solve the same problem may indeed be a reason for failing. Have your friend considered that the point of the exercise might not have been to find a solution to the problem, but rather to understand a particular path to that solution?
Many people have the ability to come up with creative solutions to a problem on their own, that doesn't take a genius. However, the solution may not be the only or the best solution available. In order to not invent the wheel over and over again, it is necessary to study and understand what solutions have been used in the past. When we understand that, through hard work, that is when we can make progress.
"The enemy's gate is down."
In fairness to these people, most of us aren't eighteen and did successfully flourish after leaving school. It's "I wasted my time there", not "I could've succeeded if only it hadn't stopped me".
I do want to have children, but even if I don't, I have several extended families that I think would do better if I know more about this kind of thing.
Totally in character, Joan of Arc snatches a miracle win from the jaws of defeat, pulling ahead of top contenders Alpha School and Dating Men in the Bay Area.
Thanks for everyone who participated in the Manifold market! It was fun trading with you. Accepting suggestions on how to give back to the community after this.
Congratulations to the winners! Those that won were all easily in my top five. All the finalists were interesting and informative.
I remember reading some comments last year from non-finalists who wished they'd received more feedback. Therefore, here are my notes on the 38 submissions that I reviewed in the first stage of the contest. They were all randomly chosen using the random review button. I also wrote a few paragraphs on my overall thoughts on what makes a good review.
Link: https://docs.google.com/document/d/105iO7NveSS_-g89arfNMiyJeqdKTTO6K/edit?usp=sharing&ouid=101255285208990633185&rtpof=true&sd=true
Thank you so much for taking notes and posting them! I did the (F) Sermon on the Mount review, and every year I wonder what I could have done to do better. I imagine most everyone who didn't make the finalist cut wonders the same thing. So thanks for telling us!
"1st: Joan of Arc, by William Friedman"
"Do we all holy rites;
Let there be sung 'Non nobis' and 'Te Deum;'"
https://shakespeare.mit.edu/henryv/henryv.4.8.html
https://www.youtube.com/watch?v=xghzOS3tQ6w&list=RDxghzOS3tQ6w&start_radio=1
I am delighted (and frankly surprised) this won, I voted for it but didn't expect it to do as well as it did (I expected the miraculous element to be the sticking point for this readership).
Stealth Catholic vote ensures victory? Much more broad-minded and generous commentariat than I gave them credit for? You decide! 😁
Eh, a "miracle" is just what interested parties call their favored interpretation of a mystery. And everybody likes a good mystery, even us irreligious irredeemables.
I think that *religious apologism* is, rightly, a sticking point for the ACX readership, and a fair amount of popular historical writing on questions of whether miracles occurred is in fact thinly disguised religious apologism, unless it's instead positioned as debunking. Aevylmar gave a convincing account of why he was interested in this question that did not require a bottom-line position on Christianity to be written first, which allowed us to accept what he was doing as good-faith historical inquiry.
First, I'm not entirely sure that's true in the relevant sense; even if you believe that there's some kind of real supernatural, you'd need an explanation for why it would be in play in any particular case (given that we so rarely encounter it in daily life), or else it's just a god-of-the-gaps argument. You could argue that Joan's miracles were faked with Harry Potter magic or whatever else, but you have to actually make that argument.
Second, in practice, the genre of post I'm complaining about usually ends with something like "so as you can see, we've proven conclusively that something supernatural happened here; therefore, my particular denomination of Christianity is true and you should all convert right now."
I think it's true in the relevant sense, and we're often misled into the opposite assumption, thinking that if something happens which we don't know how to explain naturally, and anyone involved attributes it to God, then God is the best alternative explanation to natural causes. But in the case of a lot of alledged miracles, including Joan's, an all-powerful benevolent God would need to have weird and hard to understand priorities to be responsible for them, when we include our observations of all the places that it doesn't intervene.
From the perpective of a Catholic, I was surprised that while it seemed at the start like another "well miracles can't be real obviously so let's stretch our imagination to think of a way this could have been made up" it ended with the conclusion that there's possibly something to it because the purely skeptical explanation is deeply flawed.
I thought the essay was great overall, and while the pro-miracle angle towards the end annoyed me i was able to look past it in my appraisal
As I'd said in my own comment on the article, the "miraculous" element is in fact what sold me on the article. It really drove home the point for me that Medieval Frenchmen (and perhaps Englishmen) were not like us; they did not engage in realpolitik to the extent that we do; and to them, Joan's powerful conviction and leadership skills (as we'd call them today) would've felt like a miracle. It's obvious that our modern machine technology (cellphones, computers, airplanes, mass-produced corn dogs) would have felt miraculous to them; but I never appreciated that our *social* technology is likewise tremendously advanced !
As a formerly-offensively-atheist now-agnostic - both in the Bayesian sense of "0% and 100% aren't real probabilities" and also in the religious sense - I mostly found the Miracle Question of secondary concern to the mere fact of it simply being a fun and engaging read. Life's weird, lucky things happen more often than mere coincidence, who needs fiction when you have history. Cf. Scott's Fatima post. The object-level question of God's existence is not something that I expect to be argued into/out of by a mere blog post (feels perpetually unknowable with certainty, honestly), yet the stakes are high enough that it makes for a strong topic anyway, especially with a frame that isn't mired in exhausting old New Atheist baggage. Still ended up ranking it 3rd, since I felt it strayed somewhat from the nominal review rules of not being about books, and was more of a quality aggregation of existing well-known sources vs showing new perspectives I wouldn't hear otherwise.
Yeah, the object-level question wasn't nearly as important as the combination of:
- interesting details
- written very well
The part about Joan being "an expert at the top of her game" during the first trial was one that stuck with me, but there were lots of clever turns-of-phrase throughout.
I gave myself fifteen percent odds! I'm as surprised as you are!
Congratulations on the win, and don't forget to light a candle to St. Joan!
Thank you! I continue not to be a Catholic or even know where you get votary candles.
Ah well, we'll take the intention for the action 😁
I'm even more impressed you were so even-handed and didn't go the "she was nuts/the local priest faked it all up" angle.
The dialogue was intended to have both "she was crazy" and "it was a genuine miracle" have reasonable points in favor and against, because I think those are both very plausible explanations - or, if you prefer, both implausible.
Joan of Arc was on the top of my list, initially; but upon a subsequent read, I discovered a historical error. The error wasn't a critical one—wrong French king, wrong date for a minor piece of the backstory—but, since I'm not especicially knowledgeable about the Hundred Years War, I had to wonder how many other errors there were that I didn't know enough to spot.
Now I can officially declare you manipulating your post (Fatima) schedule to push Joan of Arc specifically. Tbf I did vote for the top 2.
Btw, isn't usually the scoring of every post gets published or am I misremembering?
Poor Scott, now he's *never* going to shake off the rumours that he's converting to Catholicism 😁 (something I do not believe is ever going to happen, but now what do I know about what St. Joan and Our Lady of Fatima have in mind?)
Be careful what you wish for. Imagine Catholic EA or the Catholic Rationalism in general.
Wacky hijinks ensue.
Well, we have the grant winner for the Christian EA website and project so who knows what Sinister Jesuit Plots will mature in the future?
My head hurts.
There's a saint for that.
> Catholic EA
You mean something like https://www.bible.com/bible/111/MAT.19.21.NIV ?
Imagine a universe of shrimp trying to fit through an eye of a needle.
Congratulations to all the winners and honorable mentions, and thanks to everyone who entered.
I very much enjoyed the "non-book" aspect of this; I feel like the book reviews for the contests are never really book reviews anyway, so it was nice to have a broader jumping-off point for the submissions to the contest.
Congrats to the winners! All extremely in-depth and interesting. Joan was my personal favorite so I'm glad to see it get its flowers.
Thank you to everyone who wrote kind comments about my silly Mashed Potatoes essay. I had a lot of fun writing it and it was gratifying to make the finals! I don't have time to go back through all the comments from the original post but I will cop to being, as many pointed out, philosophically inept, spinning out a vast and flawed Baudrillard-recapitulating metaphor before walking it back to a relatively minor point about semantics. I identified the pattern I wanted to write about first and then struggled to tie it all up with a bow in the final section with any kind of consistency or rigor. I do feel it's gesturing at a real dynamic, even if it's imperfectly explicated, and I'm glad some of you found merit in it anyway (except for the guy who thought I was GPT-5? Not cool bro).
I loved your review! It was my favorite. Thank you for identifying this pattern/phenomena and putting it into words. There's definitely a "there" there, even if you "struggled to tie it all up with a bow" at the end.
>except for the guy who thought I was GPT-5? Not cool bro
I legitimately find it less insulting to be called ugly than to be accused of having used generative AI to compose a piece of text I slaved over with my bare hands.
For what it's worth, I have the reverse criticism; I thought your writing was good but remain skeptical that the dynamic you're talking about is actually a real thing.
I passed a happy car journey listening to it! I think it worked on several levels, both concrete and abstract:
- It has straightforwardly interesting facts about the origin of potatoes etc
- It weaves them into a more general and interest narrative about mashed potato, which I'd never really thought of (I grew some potatoes this year so it oddly did interweave with my concrete problem of how to preserve them!)
- It makes a broader abstract point which I think adds something to the mental toolbox (although I think whenever I apply IMPishness I need to think a lot about the context of each individual case)
All the other reviews had their own flaws by the lights of other commenters, too, and I don't think you need to be so apologetic.
I think you're being too modest. I voted for it to win, it was my favourite of the reviews. I learned some new things about potatoes and the simulacra comparison was very apt. Give yourself a pat on the back.
Fantastic Chris. Loved it.
I voted for yours despite having minor criticism of it at the time. I appreciated it being a review of an actual thing and also not being way too long.
I liked your essay and would have voted for it with a more satisfying wrap-up. If you have the time it might be worth trying to resolve the dangling threads and to post a revised version. Thanks for sharing!
I thought your review is the one who most embodied the "everything except book" review. In the way that it's the most similar to John Green's Anthropocene Review. I hope next year would have more reviews like that.
I felt it was the best written review and among the top 3 overall. I am also a fan for the whole unbundling-rebundling of systems and things, or as I am starting to call it , the IMPfication of the world.
I'm still irked that you maligned drywall in a casual drive-by line. Have you ever tried to hang a picture frame on a plaster and lath wall? Or a TV? Or a shelf?
It's such a giant pain. No standard wall anchors work, so you have to drill a big half inch hole, buy expansion anchors from the hardware store, and slide those through. Then you have to manage their alignment to be across laths instead of along laths so they don't break through, then yank to dig them in, then figure out how to adapt your hardware bolt to whatever you're trying to hang...
Putting trim on is a nightmare. Every wall is crooked and bumpy in ways that are invisible to the human eye until you put a piece of straight trim against it and it looks horrible. I've had to put low expansion foam behind trim in places to prevent mice from crawling in and out.
Drywall is a counter-example to instant mashed potatoes, not an example of the same trope. Grinding up soft rocks and gluing them together into perfectly flat 4x8 boards is SO much better than the old alternatives. It's not just price, the quality is better, and those of us who have lived in old houses don't want to go back.
lol very fair! I've never lived in an old house, very few even exist in Florida. I'd say that drywall is an example of something I mention later in the essay, an imitative species that is nevertheless more salubrious than its model and deserves to continue being planted on its own merits.
I liked it a lot and it was in my top 3, but allow me to chastise you for cheating me. You promised me a review of mashed potatoes, gave it a couple of paragraphs, then segwayed into the history of potatoes and their conservative technics. It was a good story. It wasn't the one you promised me
God rewards the faithful.
I guess Brandon Hendrickson's anti-precedent (where reviewers tried to limit reviews to <10k words in 2024 after his 2023 review of The Educated Mind got meme'd for being way too long) only lasted one year.
Congrats to all the winners! I wish Edward's kids (along with the rest of the Alpha School students) all the best!
This is just my personal opinion (and so I'm prone to the pundit's fallacy in thinking it's more broadly shared), but this year's winning and second-place reviews were the two longest submissions because they both genuinely had a huge amount of ground to cover, whereas the review of The Educated Mind seemed to be drawing out its argument as long as possible in order to wear down readers' prior inclination to dismiss its bottom-line conclusion.
No, the Joan of Arc review did not need an entire socratic dialogue in order to cover the ground it had to.
It definitely did not need several thousand words just to go over all the minutia of the Hundred Years' War.
What can I say, some of us like those minutiae!
Yes I never got bored reading it. I didn’t even notice it was particularly long.
The HYW stuff was more interesting than the JoA stuff, though. To be honest I think the review benefitted massively from the historical context vs ignoring that to focus more narrowly on JoA.
Before I put them in it was kind of incomprehensible. With longer editing I could probably have gotten it down significantly, but I actually wrote the essay in a desperate rush the week or two before the deadline and so I didn't have time for all the editing I wanted.
I can relate. We've all been there. Editing is hard. Your review was still very well-written and I'm glad you won. I'm looking forward to what you're going to write in the future!
The length was such a sticking point for me personally because it's not good for the overall health of the contest. It's fine if one person submits a super long review; it's not fine if everyone does. Initial voting becomes a slog.
Back in 2023, when Brandon won, I said at the time that I voted for it despite its length, not because of it. Commenters were worried that his review would set a bad example, and subsequent entries would all be super long.
Then 2024 rolled around, and Scott instituted a soft 10k word limit. And... people abided by it! I read (and rated) a significant chunk of the initial submissions that year, and I enjoyed all of them. When the entries are a manageable length, it's really fun to read review after review. All the finalists were under 15k words, with most of them being under the soft 10k word limit.
I don't know why people stopped doing that this year. Scott gave the same (soft) 10k word limit, but a lot more submissions blew right past it. This was a bigger problem than just you, specifically.
I barely read any initial submissions this year, simply because of math. In the time it takes me to read one 30k word post, I could have read and rated three or four <10k word reviews. That adds up when there are over 150 submissions.
I think part of the reason why the ACX commentariat was disappointed in this year's initial submissions was because they had the same constraints I did, and people simply couldn't read as many during initial voting. They didn't have an opportunity to read a ton of them and find their favorites; they probably only read a handful and then stopped. A huge part of the enjoyment I get out of this contest every year is from digging through the initial pile and finding the awesome ones that appeal to me, but not necessarily the rest of the ACX readership (and so never make finalist). I can't do that if all the submissions are long.
All that to say: your review was great! Congrats again!!!
I don't read all the reviews. I pick out ones that seem interesting (based on title and/or subject matter), start to read, and if it bores me, I don't bother with the rest of it but move on to something else.
I can imagine very conscientious people dutifully slogging through the whole of every entry, but I don't think we *have* to do that (barring the voice of conscience compels you).
Why do you think the ACX commentariat was disappointed in this year's initial submissions? Sentiment in the first-round voting thread was largely positive, and where negative was largely about minor-ish process issues (rather than content of submissions), plus some arguing about what's a "review" and some performative complaining about Skibidi Toilet.
In any event, if you're against long reviews, then your complaint is with the voters, not the submitters. There weren't in fact any submissions that came close to 30,000 words. The three longest were Joan of Arc (~24,000 words, first place); Alpha School (~18,000 words, second place); and Dating Men in the Bay Area (~15,000 words, probably fourth place according to Manifold). Besides those three, only four others were over the 10,000 word limit (one of which, The Synaptic Plasticity and Memory Hypothesis, was also a finalist), and only one (Deathbed Ballads, ~14,000 words) was clearly over the limit if it's interpreted as a "soft limit". This is out of 144 submissions in all. (Complete list of word counts here: https://www.astralcodexten.com/p/choose-nonbook-review-finalists-2025/comment/122686238)
So there wasn't in fact a huge wave of super-long reviews. Rather, there were a small number of super-long reviews, and they were the ones that did best in the voting. This makes me think that the pattern of finalist lengths from 2023 to 2024 to 2025 was probably just a fluke.
Maybe your threshold for noping out was actually lower than 10,000 words?
"If I had more time, I would have written you a shorter letter."
Several thousand words for the Hundred Years War is, what, twenty words a year? Sounds way too fast, actually.
One of my takeaways from the whole review and commentary process is, there's almost no piece of commentary on any of the finalists which isn't balanced be commentary in the opposite direction. At a certain point, you're not making the review better or worse so much as you're adjusting the sort of reader it's intended for.
Could it be that my taste is the exact opposite to ACX readers in general? In my ranking Joan of Arc and Alpha School were 12th and 13th, while School and Mashed Potato were 1st and 2nd. Any chance of publishing the full ordered list?
Scott linked (https://www.astralcodexten.com/p/non-book-review-contest-2025-winners/comment/167278351) to the top 50 entries outside of the winners, finalists and honorable mentions.
Alpha school was definitely low for me. Joan of Arc was middling. Although I admire the work put in. The mashed potato was an excellent essay, one of my favourite essays in a long time. Maybe people didn’t think it was a review.
Being an Irish Catholic, of course I voted for Joan and mashed spuds 😁 Some things just go perfectly together.
Is there any possible explanation for one submission getting an outlier number of ratings, other than that someone elsewhere on the internet linked it and drew attention to it? I don't think a submission can achieve that by its merits, since we all started off with just the giant Google Docs full of reviews that we hadn't yet read, and so couldn't know in advance which ones were any good.
I think it was because it was titled "Orgies", and people were naturally interested in that topic.
Would this not also apply to Young Swingers' Week?
Speaking as someone who doesn’t ever bother voting in these things, and mostly doesn’t bother even reading honorable mentions because the Google doc is too awkward to deal with on my phone… I had heard of Red Means No, and probably others had too, since Aella is somewhat famous around here. I have not heard of young swingers week. So if I were the going to read stuff, I’d be more likely to read about the thing I’d heard of
Huge congrats to William on the Joan of Arc Review! The other two I voted for (Dating Men in the Bay Area and Instant Mashed Potatoes--both excellent) covered ground that I'm interested in and made arguments I think are important. Joan of Arc had me rivetted to the screen in spite of its length and my lack of prior interest. Very well done!
Thank you!
We voted for the same three!
Great minds! Honestly, I'm surprised that Instant Mashed Potatoes made it to the finals at all--I loved it, but it so cuts against what I stereotype as the ACX view of things. Glad to hear you voted for it!
Disgusting to date someone while married.
Seems pretty weird to bring that up all of a sudden. Scott has been open about being polyamorous for well over a decade.
Not everyone has been following Scott that long. In which case I recommended this old post (don’t worry about the size of the scroll bar, that post had a few hundred comments; the post itself is pretty short)
https://slatestarcodex.com/2013/04/06/polyamory-is-boring/
That post was amazing. I've missed quite a lot of ancient SSC content, especially from 2013. Thanks for sharing!
Welcome :)
Aw, your disgust is so quaint.
UPDATE: Edward Nevraumont reassured me that he didn't receive any direct compensation for the Alpha School review, and while he did consult them on marketing, this was pro-bono. I quote from https://everestera.substack.com/p/welcome-acx-readers/comment/167347729:
"No paid consulting. No discounted tuition (or more accurately no discounted tuition that was not provided to every other family in the school)
I did have a couple calls with their marketing team at one point - but it was purely volunteer to help them out. I do the same thing for lots of organizations."
So while the author received no compensation, I think it is fair to call the review "undisclosed infomercial". It was written by a marketer who volunteers to help Alpha School convert customers and apparently employs psychological conversion tactics (analysis here: https://pastebin.com/fJfvQFQ2). The recent posts in the author's new linked blog can be seen in the same light.
ORIGINAL COMMENT: Hi, regarding the "Alpha School" review, the author is a growth marketing expert by profession according to his linkedin: https://www.linkedin.com/in/edwardn/, the review reads in part like ChatGPT-assisted viral marketing copy (the ChatGPT feel has been noted by other commenters), and Gemini or a similar LLM flag some stylistic devices in the storyline and rhetorics that are evidence in favour of this, see e.g. https://pastebin.com/fJfvQFQ2 here^1. I think it would be appropriate to get a statement about whether he received any form of compensation from Alpha School or other ESW/Joe Liemandt-affiliated entities, including e.g. discounted tuition, either explicitly for writing this review or in some other consulting-like role? I do realize that the truth may be "I really love to help them grow even without taking money, because I am convinced of the product" - I am more convinced of what the Alpha school product tries to do than I am of public high school too - but I'd like to know.
In particular, part of the Alpha/GT marketing seems to be explicitly directed at the rationalist community - look e.g. at the https://x.com/gtdad account, which follows rationalist/adjacents like Kelsey Piper or orthonormalist and is part of the Alpha School marketing system. I also think there is a pattern of entities receiving money from Alpha School or other Joe Liemandt-affiliated companies and writing positively about the Alpha school system (e.g. tracingwoodgrains,^2 whose Center for Educational Progress to my understanding has been mostly funded by an Alpha School grant as of a few months ago, or Austen Allred ^3). Once again, I find it absolutely plausible that the enthusiasm would be genuine (like the review author, Allred has his own kids in the thing, which he wouldn't do if he didn't like it), but I think it should be made clear whether any money or other advantages flowed to the people voicing it.
In fact, the connection is now quite direct - Alpha School founder Liemandt himself has made this very ACX review the pinned tweet on his X profile, calling it the start of "The Alpha School virus" and explicitly encouraging people to vote in the contest: https://x.com/jliemandt/status/1974984216027738467
^1 I got this response when pasting the prompt "Is this text likely some sort of hidden advertising? Can you tell me evidence in favour and against?" together with the review text into Gemini 2.5 Pro on Google AI studio.
^2 https://x.com/search?q=alpha%20from%3Atracewoodgrains&src=typed_query&f=top grant announcement: https://www.educationprogress.org/p/june-updates-gt-school-grant-and
^3 Austen Allred's writes positively about Alpha as well: https://x.com/search?q=alpha%20from%3Aausten and his current venture, Gauntlet AI, was to my understanding bootstrapped mostly by hiring demand from Joe Liemandt's education and other ventures.
^4 https://everestera.substack.com/p/welcome-acx-readers/comment/167332671
You could ask him about this on his own blog? There's a link in the post.
Thanks for the suggestion, I asked him on his blog - I will post a reply here
UPDATE: Edward Nevraumont reassured me that he didn't receive any direct compensation for the review, and while he did consult them on marketing, this was pro-bono. I quote from https://everestera.substack.com/p/welcome-acx-readers/comment/167347729:
"No paid consulting. No discounted tuition (or more accurately no discounted tuition that was not provided to every other family in the school)
I did have a couple calls with their marketing team at one point - but it was purely volunteer to help them out. I do the same thing for lots of organizations."
So while the Alpha School review was written by a marketer who is enthusiastic about helping Alpha School convert, there was no compensation.
Hey T T I am trying to DM you, but when I enter T T in the address field the system does not recognize you. I have some news to pass on to you but it is not for public consumption.
I just DMed you by mouse-hovering over your icon next to a message, which causes a profile to pop up, and clicking "Message" - please let me know if that doesn't work for you for some reason
Ewww. Hey, ACX, send your kids to Alpha Dog School and with luck they, too, will grow up to be rich and successful growth marketing experts.
Hey Scott, I'm the author of Bishop's Castle. I sent you an email with my bio. See you at Inkhaven!
Woot woot! See you at Inkhaven, too! Did you fill out your entry in the bio doc?
Yup! You can find it by Ctrl+F searching "collisteru".
See you at Inkhaven!
I'm glad you'll be alternating book and non-book review contests: I liked the latter enough to want to see more of them, but of course love book reviews and wouldn't want to simply subsume them into a larger "review anything" project.
Congratulations to all the winners.
I'm bummed that "11 Poetic Forms, Reviewed in Verse" didn't even make the honorable mentions list. I really liked it.
Found it in an open thread: https://www.astralcodexten.com/p/open-thread-387/comment/128562595
Ok, that's funny.
I really liked the concept of "11 Poetic Forms, Reviewed in Verse" but didn't feel that the author was quite a skilled enough wordsmith to pull it off. If Scott had written it, I bet it would have been great. (I feel bad about saying this because it's good for people to try ambitious things, but ultimately I gave high ratings to ones that I felt were successes.)
So. Much. To. Read.
Brain. Dying.
Feels. Good. Must. Keep. Going...
In case the authors or anyone else is interested, here are the short notes I wrote to myself after reading each review, to refer to later when voting:
Alpha School: pretty good, little on the long side
School: really good core insight, writing kinda dull
Mice Alzheimer’s: pretty hard to get through.
Islamic Geometry: fine I guess? Not very exciting.
Commentariat: fun, but dragged a bit in the second half.
Joan of Arc: Quite good, easily my favorite so far.
Instant Potatoes: Fun and interesting, not too long, but point is suspect and also mashed potatoes aren’t actually good.
Dating Men: Pretty good but feels a bit pandery.
Ollantay: Another top contender
Phase 1 Research: Interesting and a good read.
Synapse Memory: Interesting but dubious and kinda long for the value.
Xanadu: Okay I guess.
Russia-Ukraine War: Pretty good but not sure it’s a winner
IIRC I ended up voting for Joan, Ollantay, and Phase 1 Research. Phase 1 might have been confounded in either direction since it was the only finalist that I read and voted on in the preliminary round.
"also mashed potatoes aren’t actually good."
Are we gonna start the Great Potato War of 2025 on here? Because I feel the Irish and Germans would line up on one side, you heretic 🤣
It depends on the type of potato (waxy versus floury), how they're cooked (boiled into mush?), and how they're mashed (drain off the water and allow steam dry a little or mash them into a gruel in the water, then leave them coarsely mashed or put them through a ricer, then lots of liquid such as milk to make them soupy versus no, you want a solid potato not something you can drink with a straw).
One thing I like is to do all the prep up to the point of boiling, but instead of putting the pot onto the stove, letting the potatoes sit in the salted water at room temperature overnight. That's usually enough to get a bit of fermentation going, to break down some of the tough stuff and add a bit of tang.
Of course, slicing them then and then fermenting in Mason jars for a few days before *baking* results into the chips Sea Salt and Vinegar chips want to be when they grow up.
You should have saved this to enter in the 2027 review contest
Thanks for running an excellent contest, Scott. I'm always curious to see how alternative voting systems work in practice, is there any way you could share how the RCV method actually determined the results in this case?
why post the contents of footnote 3 about getting past the spam filter, as opposed to send it in an email to finalists? LLMs are good at scraping and finding these sorts of things
I am thrilled "Alpha School" placed, and just want to take a moment to highlight an aspect of the review that I don’t think has gotten enough praise: this man uprooted his entire life to ensure his kids got the highest quality, healthiest education possible. And, unlike some parents who obsess over their kids’ education, he graciously and lovingly admits that both of his kids have unique personalities, and is eager to find individualized motivations that work best for them.
Edward, you're an incredible father. I want to hug you and clone you.
>he graciously and lovingly admits that both of his kids have unique personalities . . .
Jeez, that's not exactly a high bar to clear. In fact I'd say any parent who fails to recognize this fact is 15th percentile or below.
> . . . and is eager to find individualized motivations that work best for them.
Now he's up to 40th percentile.
I help with mentorship programs, and my experience with the parents who are obsessed with their kids' education is that they tend to want their children to fit a very cookie-cutter mold of success. They get fixated on their kid getting into Stanford, MIT, etc, and then force their child toward whatever interests/activities/achievements will get them into those schools. They essentially create a personality for their child and then force them to wear it.
I instinctively cringed a bit when I saw a very long post about someone trying to optimize their kids' education. But I was very pleasantly surprised by how in-tune Edward is to his kids' mental well-being, unique interests, strengths, weaknesses, etc.
I think he's a rare balance of someone who is both hyper-fixated on his childrens' success and also deeply respects their children as individuals and cares for their mental health. And I think it's worth applauding. Frankly, I'm baffled by why you'd want to mock it, but so be it.
I wasn't mocking your points, I was critiquing them. But I did not get that you were not comparing him to parents in general, but saying that within the population of parents hyper-fixated on their kids' achieving certain markers of success, he's much better than average. That I do not find absurd, though I'm still inclined to think of it more as being the best of a bad lot than as a precious and remarkable achievement.
Critique is an art form of its own, and I would suggest you learn how to give it in a way that does not sound like mockery. Your point very easily could have been made without sounding snide.
And to each their own. I never got the chance to go to high school, so my mind is absolutely blown when I come across individuals who put so much effort into their children's education. When done in a compassionate and humanizing manner, I absolutely believe it's worth applauding.
You are right that I could have made my point in a pleasanter way. You could lay my snideness to spillover anger from other, related things I object to. First of all, I do not like the sound of Alpha School. The thread where the review was posted for comments has something like 10 long, detailed posts of mine articulating what I think is wrong with its objectives and methods. Second of all, I am very angry about information that has come out about Alpha School and the review's success iin the present thread: The founder of Alpha School had a pinned tweet on X asking his 45,000 followers to vote for the review: https://x.com/jliemandt/status/1974984216027738467. And the request was retweeted 32 times. That's ballot stuffing. Still want to clone him? IMO the world already has enough people who use their smarts exploitively for us to get by ok without cloning this mofo.
On the other hand, calling my anger spillover suggests that it's irrelevant to my present complaints about the writer and dad in question. It's not. Seems to me that both he as a contestant and the school he's promoting in his infomercial are overconcerned with winning competitions and underconcerned with being trustworthy and real.
I guess if everyone he asked to vote did come over here and vote, then yeah, that's ballot stuffing.
Good job St. Joan thwarted his plans just like she did the English, then!
"When done in a compassionate and humanizing manner, I absolutely believe it's worth applauding."
That's the rub, isn't it? Alpha School sounds like rearing turkeys for Christmas. I enjoy my Christmas fowl but I wouldn't raise another animal, let alone a child, like that: plump 'em up fast for the slaughter and the profit of the owner (of the brand which he seems, if Eremolalos is recounting correctly, to be marketing as hard and perhaps as deceitfully* as his little spindly legs will run on the treadmill).
I agree with Eremolalos that I don't like being used for free advertising.
*There seems to be some little blurring about who founded what and who owns what; one source (Wikipedia) says the founders of Alpha School are MacKenzie Price and her husband Andrew Price; another source (Wall Street Journal) says some guy called Bill Ackman is acting as the public face/ambassador for Alpha School, and finally Wikipedia again says the owner of the magic software that the school operates off is this Joe Liemandt, who also owns the companies that Andrew Price works for.
I often feel sort of at home here, which is a rare experience for me, but the present “review” contest has had me wandering deep in familiar alienation. I hope some of that reaction is accounted for by how many outsiders participated as contestants and "judges." One post here today (https://www.astralcodexten.com/p/non-book-review-contest-2025-winners/comment/167323863) says that the founder of Alpha School, Joe Liemandt, had a pinned tweet on his X profile asking readers to vote for Edward Nevraumont's review of Alpha School. Seems to me there's virtually no chance that Nevraumont was not aware of the pinned tweet or that he asked Liemandt to take it down and Liemandt refused. Also seems highly unlikely that any who arrived here via that route actually read any of the other reviews. So they weren't really judges or voters, just unpaid ballot box stuffers.
GPT's assessment of amount of AI-generated content in the essay:
>The piece reads overwhelmingly as human-written with some AI-style scaffolding or heavy-AI paraphrase in certain middle sections. The dominant voice is cohesive, witty, and self-aware—traits that large-language-model prose still rarely sustains across 25 000 words. But there are several passages whose sentence rhythm and connective phrasing match AI composition patterns strongly enough to suggest the author may have used an AI draft or rewrite tool, then edited heavily. . . . Overall, the essay’s conceptual architecture and narrative sensibility are human, but perhaps 15–25 percent of the phrasing bears AI fingerprints—primarily in data-dense or connective prose.
Full GPT response is here: https://chatgpt.com/s/t_68f2bab449a081919ac5ba3e85cab785
Ewww to the whole thing.
**********************
So I am posting here my personal awards:
*BIGGEST HAUL OF NEW AND INTERESTING INFORMATION*
The ACX Commentariat
*BEST NUANCE PARSING REGARDING SUBJECTIVE MATTERS*
The soul of an anti-woke intellectual: Jamie Q. Roberts
*SINGLE MOST INTERESTING PARAGRAPH*
Japanese social interactions exist at a much higher resolution than American ones, and at times I felt that living in Japan as an allistic person gave me a reasonable understanding of what it might be like to be autistic in America. At all times there were subtle games being played, and things being communicated by other people to which I was not privy at all. This created a background radiation of stress that made social interaction much more stressful than fun. Japanese people get around this by leaning into role assignment; basically, your boss is just your boss, and must play the role of boss rather than ‘being himself’ because if everyone was particularly ‘himself’ at this high resolution of social interaction, everyone’s minds would collapse under the information overload, and life would be impossible, even for those born into the culture. (From The World as a Whole)
*MOST ACUTE AND ENTERTAINING SINGLE SENTENCES*
“Less powerful than Cantor’s [wings], but still airworthy enough to reveal the vistas that Cantor saw; and robust enough not to melt and send us plummeting into the sea of mathematical paradox.”
“One sees crystalline souls that have polished all their facets into a perfect refracting jewel of self-awareness, and often a performance of self-contempt, so no one can ever accuse them of being cringe.”
“. . . young people who are by default edgelords, unserious Christians, iPad kids, psych med-takers, or bog-standard faces in the halls to mine the social media algorithms for what is most in opposition to the listlessness and malaise they feel.”
“Plastiform glee that intensifies into a sort of pop-eyed, carnivorous ecstasy. “
“This speaks to its visual literalism.”
“The villanelle so strangles sense that one can't tell the meaning from the structure.
“ . . . something done in the physical world in an intangible age. An age of circling back, checking in, lit reviews that go nowhere, and innovative best practices in marketing.”
“He appears to have an almost fanatical devotion to parsimonious explanations.”
“The concept of bad faith- akin to mauvaise foi but with a sunk-cost ratchet.”
“It is immortal precisely because it is flawed. Like human language, its ambiguity delivers the creative freedom that closed systems cannot.”
“Past the veil of shame is where the dark, gross, raw workings of the heart lie. Meet me there.”
To be clear, the person encouraging to vote in the contest (together with promotion of the review) is not the author of the review, but the founder of Alpha, Joe Liemandt himself (he used to be quite secretive, letting Mackenzie Price present as the public face, but has become less so in the last months): https://x.com/jliemandt/status/1974984216027738467
Before education, this person got (in)famous for acquiring struggling companies and operating them with remote contractors (https://archive.is/ar0Cr), using his remote working platform Crossover that pioneered bossware/mouseclick "productivity" surveillance (https://archive.is/PpoLs). A big part of the philosophy of Alpha School seems to be adopting this metrics and surveillance-driven approach to schoolchildren.
According to the linked blog post, Liemandt started presenting more publicly in part because the review author accidentally outed him, not realizing that his involvement was supposed to be kept quiet.
And do you have any doubt that the author of the review could have kept him from soliciting votes on X? In my opinion anyone who solicits votes on social media or accepts having someone else do it for their own gain or the author’s should be disqualified
Fuck this shit, every stinking aspect of it.
OK T T. I have changed my post to reflect all your corrections and clarifications.
Thanks for singling out my paragraph. I agree that promotion of reviews on other platforms sets a dangerous precedent. It would take very little for the contest to lose all meaning because of tactical social media campaigns.
You're in the Single Sentence category too, with your riff about Macedonian teens. You seem very gifted to me in your takes on people and settings, and your ability to put nuanced takes and odd unsettling truths into words. I hope you'll keep writing, so long as it gives you pleasure. There are way fewer settings than there used to be where talent like yours is appreciated, but there are still some. My own talents are in the same realm as yours, and I've had to come to terms with the fact that ACX is just not a setting where gifts in that realm are much valued. If you hang out here, it has to be for the other great things about this group. You might be amused by my own entry, which is about AI text-to-image art, but can be read as an extended angry riff about the triumph of plastic analogs over the real deal. It's Dall-2 & 3 as Depictors of Human Life. But fer crissakes don't feel like you owe it to me to give it a read! Anyhow, DM me if you'd ever like to chat.
I enjoyed your thoughts about the superiority of Dall-E 2 over 3, from an artistic perspective. What do you think about the newest image generation models?
They’re Dall-e 3 on steroids. Everything looks grotequely perfect, glossy, and commercial arty. And lately the sites for online text-to-image generation are all over the user with pop-ups etc trying to addict them to making AI videos. I haven’t tried the video features. They look especially unattractive to me because there seem to be a lot of stock characters and preset grooves for the stories to run down.
I have a beloved collection of Dall-e 2 images, prob. about 200, and would love to make more. When life settles down a bit I am going to see what I can figure out about training a model that is more like Dall-e 2. I am not at all tech-trained. Have never written a line of code, do not know what API even stands for. I look it up, then forget. But I think with help from someone I know I can at least try it. I do not know whether that approach is promising, though, because it seems to me that a lot of what made Dall-e2 excellent was the *lack* of various kinds of enrichment and controls. Seems like train a new model really comes down to tweaking an existing one — you don’t start from scratch. So you’re adding something to what’s there, and what I want to do is subtract.
Have you played around with the text-to-image AI’s?
I have played with https://diffusionbee.com/download as a frontend for the confusing plethora of open models out there. Unfortunately the base Stable Diffusion models were poorly trained (poor quality training data and a weak language part) and the various later finetunes have particular obsessions behind them so I haven't found a model I can really recommend for artistic flair. I've seen discussion of https://github.com/lllyasviel/Fooocus as a frontend as well but it looks much less polished. I also have a bunch of Dall-E 2 images that had a spark that seems missing in newer, slicker models.
I am hoping someone will look at making open image generation models again, starting with a more capable open language model, but this requires real effort and a lot of curation to avoid copyright issues, let alone the righteous scorn of artists who didn't put their work online for the purpose of being automated away.
Having been a reader since the early SSC days, I too have felt vaguely threatened and alienated by the massive influx of normies. This wasn't helped by Scott generally being busy for other reasons and not quite as on top of moderating as usual.
I agree that outside promotion is icky. I suspect the LLM slop will only become a bigger problem, too.
Haven't you just disqualified yourself as a rationalist by making an us-versus-them distinction? After all, EY's writings emphasize truth-seeking over social belonging (but he talks a good game).
Anyway, I guess I must one of those normies. But I'm curious about how you detect a rationalist from a normie by their comments? The epistemic process people use to support their comments are seldom made explicit, so the "rationalness" of their conclusions is frequently unclear. And despite their espousal of Bayesian reasoning, I don't see many people updating their priors when confronted with contradictory data. Maybe we're all normies? Just sayin...
I never claimed to be a rationalist. I'm just some guy who's been reading Scott Alexander for just over a decade. I'd never even commented until recently. I've read The Sequences and think EY is a bit of a hack (don't write a preface explaining that you used all the wrong examples--go back and FIX IT).
I've seen your name around. When I say "normies", I mean the people who are clearly not regular readers of the blog (which I guess from the contents of their comments, apparent lack of awareness of previous posts and reviews, &c).
Sorry, the broken link to Liemandt's pinned tweet was my fault, corrected now.
Thanks for these gems. I've now read Whimsi's piece, definitely agree with your recommendation.
yeah, outside manipulation feels gross to me, too.
In addition to everything else (especially Joan of Arc, the only review that earned a 10 from me), I want to congratulate Max Nussenbaum on having now written *three* banger ACX contest reviews that I keep coming back to for a good read. It's disappointing that this one didn't make the finals, but I'm glad to see it recognized anyway.
I'll be frank and say I'm crushed my review didn't bubble up to the top of the honorable mentions. I put my heart and ten years into it though, so I suppose I'll try to be content with having let out many things that I've long believed and left unsaid.
Congratulations to all the winners. Here's a link to my review if you didn't read it. Shout out to the commenters who singled it out, and seemed to understand what I was getting at.
https://whimsi.substack.com/p/my-review-of-the-world-as-a-whole
Originally, I haven't noticed it among the submissions. I've got to reading it only now. And holy moly, that's a seriously good piece of writing. It's almost Steinbeckian in its depth and insightful in that particular human-centered way that's almost fully missing from the rationalist-adjacent blogosphere. I still have to fully digest it. And then maybe write a review of it.
Thank you!
Next time, I suggest writing something that doesn't make people want to kill themselves after they finish reading it. Pander to your audience! Reinforce their prejudices!
Good idea! I plan to write a sequel that casts the review in a less pessimistic. It's intended more as a picture of things as they are, then as they must be.
Sequel idea: FOREIGN PANTIES: GETTING LAID EVERYWHERE IN THE WHOLE WORLD
I should write about this aspect of it, honestly.
Very well written. A bit too far in the you are so privileged direction for my taste but well done for sure.
I understand that, to a degree. To those that already know, it's only a depressing reminder, and to those that don't know, it may seem like nothing but an argument that they should feel bad. But I don't believe in collective guilt, and that isn't the point of the essay at all.
Happily surprised by Alpha School - not being a winner, that was kind of a given, but by author. I remember doing my part to dunk on the previous comics review. It's unusual to dislike something so strongly that it becomes a recurring synecdoche for bad things of a type! Big tsuyoku naritai energy for learning based on gradient descent and putting out a much, much stronger review this time. "As long as you learn something, it's not a mistake," says Archer.
Befuddled by not recognizing any of the finalists as commenters whose writing feels very familiar, like as if they're a big-name regular that I've been reading for years. At least two entries this year gave that strong sense of readerly deja vu. More evidence of convergent evolution, I guess...spend time on ACX, start to write like/complementary to Scott.
Most of the finalist and honorable mention reviewers aren't people I know, but in two cases (Joan of Arc and JFK Assassination Conspiracy Theories) I found specific tells that allowed me to, with fairly high confidence, figure out who the author was before today's reveal. (In the former case, there was actually a Manifold market on it, not created by me, and I made a decent profit.)
JFK author here. SO curious what the tell was — can you reveal?
The review contained the same anecdote about LBJ as the review of The Outlier, with similar wording.
damn, impressively close read! I didn’t even realize I’d done that
IMO, long form essay writing is fairly different in style from blog comments, so I don't try to hard to spot stylistic tells. But as far as convergent evolution goes, keep in mind the filtering effect. The blog readership is heavily selected for people who like Scotts writing style. When they vote on finalists, that preference is expressed in the selection. Plus, to an extent, the authors are probably catering to that deliberately.
Glad Gallow did so well, the military theme was really unusual and refreshing to me.
Elons musk engineering algorithm and Bukele are too interesting reviews, thanks for bubbling them up.
Funny story: Sabaton released a bunch of new music videos today... including 1 about Joan of Arc! https://www.youtube.com/watch?v=fTqB-kWVnBA
Thank you, I needed some music recommendations and this will go nicely with their Swiss Guard and Winged Hussars songs! 😁
Someone did a great video for the Winged Hussars:
https://www.youtube.com/watch?v=CxlRJsQ7p2k&list=RDCxlRJsQ7p2k&start_radio=1
Dude. So nice to see this mentioned here. That songs gives me SHIVERS.
Congrats to the winners!
Just commenting to say that I really really liked "The Life's Work Of Banerjee, Duflo, and Kremer", I'm sad that it didn't make it to a finalist but I'm glad that it made it to the top half
Well, most people here sound sort of like people at the end of a fine Thanksgiving dinner. This was great — thanks, Scott, for hosting it, and for all the fine people who participated by writing or judging — great bunch of essays, guys — congrats to all . . . So I feel sort of like the skunk who slipped in through the dog door and shows up in the dining room doorway, all scary and stinky, saying this:
Edward Nevraumon, author of the Alpha School review, cheated. He should be disqualified. There are a number of posts about this matter scattered on the thread, but here’sthe brief version. During the competition, Joe Liemandt, founder of Alpha School, had a pinned tweet on X asking followers to go to ACX and vote for Nevraumon’s review. (https://x.com/jliemandt/status/1974984216027738467). It has 45.5 K views, and was retweeted 32 times. It went up on October 5, 3 days after voting started.
Nevraumon has not denied knowing that tweet was up, and it would be quite implausible if he did deny it. The fully honest thing for him to do would have been to tell Liemandt to take down the tweet as soon as he knew it was there, then inform Scott of what had happened. But I’d settle for his just having quietly asked to have the tweet removed. But nope, the thing’s still up, and presumably was up from 10/5 til voting ended on 10/13. Nevraumon, who’s a growth marketing expert, has denied receiving any direct or indirect compensation for writing his piece about Alpha School, and I don’t know anything that makes me doubt that, but it’s clear there was communication and coordination going on between Nevraumon and Liemandt about the review. How otherwise would Liemandt even know that Nevramon was entering a review in the ACX contest, much less know when voting began?
Last year one entrant was disqualified during the voting, because of “irregularities.” Seems likely irregularities were suspicious patterns in the votes the review received, though I supposed there could instead have been reason to suspect plagiarism. In any case, it seems to me that the case for disqualifying Nevaumon cannot possibly be weaker than the one against last year’s disqualified reviewer. In fact I’m having a hard time imagining a stronger case. He told someone who’s deeply invested in the success of Alpha School about his review, and informed him when voting began. He said not a word to anyone here about the tweet soliciting votes. Some fraction of the 44.5K people who saw the tweet (plus of course those who saw retweets) very likely came here and gave him a 10. That’s ballot stuffing. Even if in fact not a single X rando showed up and voted, Nevaumon is still just as guilty, because he went along with an attempt to send an army of ballot stuffers here to up his score.
I’m amazed that nobody but me has spoken up loudly about this. It seems like a clearcut, especially egregious case of cheating to me. But of course the tone here today nudges everyone in the direction of being mellow and positive.
I’ve been reading since the SSC days and AlphaSchool was a favorite of mine and I also judge it likely to interest the typical reader of this blog, at least any that have or are thinking of having children. It also gets bonus points for being basically the only actionable reviewed item. I don’t find its performance surprising and I don’t consider someone else tweeting something to be a form of the author cheating nor do I expect a review author to be aware of tweets about his review, even from people the author knows.
Of course there is an alternative to disqualifying Nevraumon.. We could agree to just not treat vote solicitation in the review contest as a problem. I don’t believe Scott stated anywhere that vote soliciting is not allowed, or even said that it’s sleazy and please don’t do it. And of course there is no way to fully enforce a rule of this nature. Only pretty flagrant violations will come to anyone’s attention. Any anyhoo, that’s how the world works, right?
In that case, I would like to be the first to take advantage of the situation by making an offer to the group. Next year I will enter the best, most crowd-pleasing review I can, and pay up to 4 people to give my review, and only mine, a 10. I will give each person who does that 10% of my cash winnings, or $500, whichever is greater. For those squeamish about accepting cash for their votes, I am also open to an arrangement in which the individual gives me, and only me, a 10. I in return will give their review, and only theirs, a 10 *and* guarantee that 2 close acqaintances of mine also give them a 10.
You scratch my ass, I scratch yours. That’s how things work in most of the world, and it’s time we got with the times, right?
Whadda say, guys? ?
Just in case your comment about people giving the Alpha School review a 10 reflects genuine confusion on your part: Voters were asked to list their top three choices from among the finalists. The rating process was done earlier, before the tweet was posted.
Since Scott has not (as far as I know) posted any specific rules about this kind of cheating, I don’t think that this incident should be treated as an automatic disqualification. I think that ballots cast by people who don’t have a history at ACX (voting in previous contests or posting comments here prior to the tweet), and which list the Alpha School review as the only choice, should be considered suspicious. If there are enough votes in that category to change the outcome, I think that would be reasonable grounds for disqualification because it would mean that the purpose of the contest had been subverted.
Of course this is really up to Scott, and I don’t know whether he is reading this or is even aware of the controversy.
This incident has changed my view of the school more than my view of the essay. I might have followed the progress of the school; now I no longer consider it worth my time.
<Just in case your comment about people giving the Alpha School review a 10 reflects genuine confusion on your part:
No, sorry, I understand the voting system. What I wrote reflects genuine preoccupation not genuine confusion.
<Since Scott has not (as far as I know) posted any specific rules about this kind of cheating, I don’t think that this incident should be treated as an automatic disqualification.
I think Scott simply assumed everyone would know that soliciting votes from randos on public forums was not OK. Here's why: 1) He also did not, so far as I know, state that having plagiarized bits in your review was against the rules. or that entering under your name a review written by someone else was. He did not state that people should not DM other writers and threaten to out them for something-or-other, or embarrass them with false but hard-to-disprove accusations, if they did not withdraw their review. He did not say that reviews that were entirely AI written and not identified as such were forbiden. He did not say people were not allowed to set up multiple accounts here for themselves and use them to give themselves multiple votes. I believe he assumed we all knew those things. 2) The review contest is a contest, not a political campaign. It's a contest of skill, skill at writing a review that delights ACX readers. Everyone knows that in contests of skill using means other than personal skill to win is cheating. 3) Scott disqualified someone's review last year for "irregularities." I think he said the issue was voting irregularities, though I'm not sure. There don't exist many kinds of voting irregularities in this situation. In fact the only one I can think of is somehow getting votes that are not genuine. So someone could get other people to vote as they direct them, or I they might somehow set up a bunch of accounts and give themselves multiple votes.
>This incident has changed my view of the school more than my view of the essay. I might have followed the progress of the school; now I no longer consider it worth my time.
I did not like Alpha School based just on the info in the review. I could smell on it the same exploitive, competitive stink I smell on Nevraumon & Liemandt.
I wandered away from the Joan of Arc review as soon as I realized it wasn't meaningfully a "review" as I would define the term (an analysis of a thing used to inform a reader whether or not and sometimes how to act in relation to the thing). I was shocked to see it ranked number one; we must have a lot of history enthusiasts across ACX!
But Alpha School certainly was a *REVIEW*, and I ranked it in my top three for being a very thorough, very interesting *ACTUAL REVIEW* about an educational system. I don't have kids, but I can see the possibility of occasionally bringing it up to parents, especially wrt its ideas about motivating kids.
I don't know if the Tweet for Alpha School can be considered ballot-stuffing, but it's certainly *campaigning,* and I would prefer to have Scott outline some explicit rules against it next year.
But don't forget that Alpha School was either the first or amongst the first to be posted in the review contest, and it seems likely it received more organic reads (and certainly comments) amongst the ACX folk. That and being an actual review very well may have organically been enough to boost it as high as it went.
And for what it's worth, my voting was:
1. Phase 1 Clinical Trials
2. The Review Which Shall Not Be Named
3. Alpha School
> How otherwise would Liemandt even know that Nevramon was entering a review in the ACX contest, much less know when voting began?
Possibly for the reasons he stated in the tweet, that the review was read by and impressed a lot of people and brought Alpha School a lot of media coverage prior to his tweet? I don't see any reason to doubt the honest explanation here. ACX is not an obscure blog at this point, and is read by people in the highest echelons of society.
I'm not doubting the explanation Nevraumond made. He did not address the issue of how Liemandt knew when the voting began. What he said was that he was not paid for writing the review, either directly or indirectly via reduced Alpha School fees for his kids. I believe him about that.
I actually do not fully understand what you are getting at. You responded to my asking
> How otherwise would Liemandt even know that Nevramon was entering a review in the ACX contest, much less know when voting began?
Your point that Lielandt knew about the existence of the review because it was viral makes sense. But it doesn't answer the question of how Lielandt knew when the voting began (his tweet went up 2 days after the start of voting).
Personally I thought Joan of Arc review glossed over key points and left me unsatisfied. Was she really an artillery genius? Was she really outwitting them at the trial? I’m skeptical of both on priors and nothing presented moved me either way. I also wasn’t too interested in the meta take at the end of the review, which I guess is what people liked the most about it?
The Russia-Ukraine war review deserves to win the “first ACX review that managed to be too short” award.
Overall, I thought this year has the highest quality reviews so far and I'm not surprised that my some of my favorites didn’t place. Shout out to the geometric designs review in particular.
It definitely could have been longer, but then it was already quite long to begin with and several people commented to say they didn't finish reading it. It sort of felt like 3 separate essays- one about the 100 years war in general, one about Joan's specific role in the war, and one about theology and the case for why she might have performed genuine miracles. I would like to see the author pick these three threads apart and publish them as 3 separate essays, if he wants to spend more time on this.
I wasn't really arguing for it to be longer, but rather that the parts it focused on weren't the important parts for me. Maybe I should look over it again, but I came away with a distinct impression that I hadn't learned much despite having only a superficial knowledge of Joan of Arc going in. (I knew basically just that she wowed a lot of people in battle and in trial but no more detail of whether the awe was deserved. I think the main thing I learned was that the awe was well recorded and contemporary and from her political opponents to a significant extent.)
A lot of good ones in the honorable mentions! I wish there was a place to discuss those. I know we can discuss them all there, but its just feels cramped and limited here having one comment section for all of them combined.
Wow I only just saw that someone reviewed the Miniatur Wunderland. I really liked the review.
Thank you! Glad you enjoyed it.
> can be found at his nascent Substack : https://substack.com/@gallowglassglen
This link does not work
Sooo can we now ask for Alex's dating doc? Her description of the whole man made me smile
Yeah congrats to all the winners, I really enjoyed reading your pieces.
(or I stopped 1/2 way through :)
And to all those in the lower half. Many thanks for your effort too. And last but not least, thank you Scott for hosting this. It's a highlight of my spring/summer. (I'm a shitty writer, but I dream of some day posting a book review.)
I think it is important that people reading the Alpha school review are aware that
- it was written by a professional marketer - and while the author's enthusiasm is surely genuine, it employs psychological devices made to convert. See again the Gemini review pointing them out: https://pastebin.com/fJfvQFQ2 (Update: This is not just sycophancy, see https://www.astralcodexten.com/p/open-thread-404/comment/168401636 for running the same prompt against other contest entries about products, which yield different results)
- a very substantial fraction of the positive attention that Alpha gets on social media is fanned by people who somehow received or receive money from Liemandt's ventures, and this is not obvious to spot because of the mirror maze of companies, entities and social media accounts that appear to be independent endorsers of each other, but all trace back to him.
For an extreme "pure bot" case (there are humans with significant following and varying levels of disclosure as well), see https://twitter.com/insideaustinedu for an apparent bot account that pretends to be a neutral resource but posts nothing but slop+engagement bait promoting Alpha. It seems to be part of the strategy laid out in https://contently.com/content-marketing-case-studies/case-study-alpha-school/, which is to manipulate LLMs into endorsing Alpha when they find such slop in web searches.^1 My understanding is that such a bot, if it was indeed operated by someone working for Liemandt, is a violation of FTC rules on deceptive advertising.
This whole picture leaves me with a somewhat bad taste regarding the integrity and honesty of the whole Alpha operation, and makes me doubt their marketing claims.
^1 Contently is Liemandt-owned too (acquired via Zax Capital - see the mirror maze?) - so the case study, which implies that Alpha paid Contently in a free-market business transaction, is an example of the "fake independent endorser" pattern - and this "case study" was reposted by https://x.com/bpizzacalla , who works for an ESW Capital entity and is the only human follower of the bot account. The other 2 followers are https://x.com/MarinSoftware, another advertising company recently acquired by Liemandt (via "Kaxxa Holdings, Inc"), and apparent retweeting bot https://x.com/denyako .
This is interesting but I’d like to officially declare myself neutral about the issues you raise. The reason is that I feel very strongly that Nevraumon, the author of the Alpha School review, should be disqualified for cheating, and that’s really a separate issue from whether the Nevraumon uses “psychological devices,” whether the Alpha School is sleazy, whether it violates FTC rules etc.
I do not want the issue of whether the review should be disqualified to be tangled up with the issues you raise here. If the review had been about homemade quilts, and the Nevraumon’s mother had solicited votes for him from her many followers on X my opinion would be exactly the same as it is now. It would be the same even if the quilts and their makers and their business model were all above reproach. The dishonesty I am pointing to has nothing to do with Alpha School itself. It is the soliciting of votes on social media.
The vote-soliciting tweet was not on Nevraumon’s X account , but pinned von the account of Joe Liemandt, the founder of Alpha School., who has 45K followers. It was put up on the 3rd day of our voting period, and last I checked was still there. But Nevraumon has not denied knowing about the pinned tweet and It would be very hard to believe if he had. Seems clearly there was some coordination and collaboration going on between Liemandt and Nevraumon — otherwise, how would Liemandt have even known Nevruamon was entering the ACX contest, what his review was about, and, of course, what day the voting began?
Sorry to peel away at this juncture, T T, but the points you’re making are hard to prove, and open to debate, and they will probably be discussed at great length and there may never be agreement. The cheating, though, seems to me unambiguously bad and unfair. And it’s probably a larger scale version of what got somebody disqualified last year. Seems to me we and/or Scott should act promptly if we are going to. (And if we are not — why on earth not? )
To be honest, I think that the tweet situation is, on its own, iffy but ambiguous. The review went viral long before that tweet appeared (though that may well have had to do with Liemandt employees promoting it), and it's hard to hold the author responsible for the actions of a billionaire.
In the particular case of this product review, I find that the Alpha/Crossover/ESW universe as a whole is filled with indifference to honesty and iffy behavior. The bot account is clear-cut astroturfing, and its existence should make anyone less confident that every vote for the Alpha review was an honest vote by a disinterested human. I wrote my comments to help people see the whole picture around this matter - but I don't want to tar Nevraumont, who did not get compensated and apparently wrote a review using his professional content marketing writing style out of genuine enthusiasm.
Zooming out from this particular case, I think that reviews of things with a large financial stake in their image in the tech community are likely to generate ambiguous situations of this sort.
For future contests, one solution may be a rule like "you should abstain from voting when you were encouraged to do so by someone else than Scott, or did not reach the review contest by being a preexisting ACX subscriber". For this contest, this rule unfortunately didn't exist, which would make it unfair to tar and feather this particular author in my opinion.
<To be honest, I think that the tweet situation is, on its own, ambiguous. The review went viral long before that tweet appeared, and it's hard to hold the author responsible for the actions of a billionaire.
It went viral on the internet before the tweet? Did it go viral in the form of a link to here, the ACX version, where it was nestled in among a bunch of other reviews? If not, where was the other version, and by what route did it become viral? If it somehow went viral via some route Nevruamon knew nothing about then no, I don't see Nevraumon as having done anything wrong there. But if it got posted somewhere other than here, I don't see how that would have led to its getting votes here.
As for not holding people responsible for the actions of a billionaire? Huh? No, you can't hold someone responsible if they don't know the billionaire -- they'd quite likely never be able to get word to him about their complaint. And you can't hold them responsible if they tell the billionaire their complaint and he says "if you don't like it, sue me." But these 2 guys very likely at least knew of each other, had probably had had at least some contact prior to the review. Nevraumon says he talked to the public relations office a couple times and gave some advice. He had 2 kids in the school Liemandt founded. And he's a marketing specialist with a big online presence. He would not have found it impossible to ask Liemandt about the tweet. And note that he is not saying either than he had no idea the tweet was there, or that he tried and failed to get Liemandt to take it down.
I first became aware of it in google doc version, IIRC when it was linked in a reply during an X discussion on Alpha School and the future of education/learning apps more generally. The school has plenty of organic buzz by now, and this review was the most in-depth public explainer of the Alpha School approach.
Yes, using rationalist community norms, what you said is what should have happened. By standard business marketing norms, I think the tweet would be completely fine behavior - this indifference to honesty is precisely what I am worried about. Solution going forward could be as discussed above.
My point with the astroturfing bot made to poison LLM responses is that it is a very clearcut "smoking gun" of unethical and likely illegal behavior that falls short of business marketing norms as well. "Education influencers like TracingWoodgrains get money from Liemandt, which disincentivizes them from being critical, and the vast number of people and entities paid by Liemandt form a hidden endorsement ring" is more ambiguous, but probably ultimately more damaging to discourse.
This "growth hacking" approach and indifference to truth and nuance in favour of metrics is precisely what should make us skeptical of their vast, poorly-disclosed marketing surface, which effectively includes the significant part of education X that took money from Liemandt (like e.g. TracingWoodgrains), their many employees who are very vocal on X, and this review.
I think that "indifference to truth and nuance in favour of dumb metrics" is precisely a failure mode of the product itself (which is children's education). What's more, Alpha School is intended to adapt the metrics-driven "productivity surveillance" approach of Crossover to kids, so the negative psychological effects reported by some Crossover employees have a risk to carry over to children as well.
>Yes, maybe if he had known that it would be perceived badly, he may have asked Liemandt to take down the tweet. I would guess that the norms in the business sales world are just so different from the norms in the rationalist community that it wouldn't have crossed either of their minds that such a tweet could be seen negatively. Just important to point out the difference in mindset and the more instrumental approach to making assertions of fact in the former.
I understand that there are lots of worlds and sets of norms out there, but the few high-powered people I know who come from one of those worlds understand that too. If they are coming to a gathering in a less high-powered realm they ask questions about how various things are handled. Are people here sensitive about subject X? How much do people usually tip here? Is it likely anyone here reads my column or knows about it? And then, if they are Deputy Dean of Yale Law School, as my old college roommate was til recently, they describe themselves as a lawyer and politics junky if asked at dinner what they do for a living. Does it seem plausible to you that it would not have occurred to Nevraumon, someone in a profession where understanding people and groups is a crucial skill, that ACX is a small, smart eccentric group that has norms he should find out about? that contests are a big deal to most people who participate? that soliciting votes in a contest of skill is seen as grossly unfair in virtually all settings?
But let’s say all that somehow never once crossed his mind. Here’s a thought experiment. Let’s say I’m good friends with someone on Substack who’s got a big popular blog, and they offer to lpublish my review on their blog with me as a guest writer. And on their own initiative my Substack friend introduces my post by saying I’m a good friend and one of the smartest people they know, blah blah blah, and urge readers who like my piece to go to ACX and vote for it. I do not ask him to take down the suggestion that his readers cast votes for my review. Then, after I win second place, someone here complains about my soliciting votes on a pubic forum. And I say, “Wait . . .what? I didn’t solicit them, my close friend did. I didn’t even ask my friend to feature my piece — it was his own idea. And Scott never said our reviews could not be published elsewhere, or that we could not have people championing them on public forums. Jeez, guys, I don’t have telepathy or precognition. How was I supposed to know that wasn’t fine?” How would you be with that? Even if I seemed to be sincerely surprised, would you have been OK with my review staying in the contest?
I would absolutely not be OK with somebody else getting away with that. No, Scott didn’t say we could not solicit votes. He also did not say we could not plagiarize, or could not submit an essay written entirely by someone else under our own name. He didn’t say we couldn’t spook other contestants by DM-ing them a message saying (untruthfully) that 3 AI’s have declared their review to be 90% AI written and that they are going to announce this to the group if they do not withdraw their review. Some expectations are so obvious in a given context that they do not need to be stated.
Also, Scott disqualified somebody’s review last year because of “irregularities.” I think he said “voting irregularities,” but I’m not sure. Apparently there are some irregularities for which Scott believes ignorance of normative ACX expectations is no excuse.
As for what you have dug up about Alpha School and the alpha dogs who run and promote it — I am delighted that you found that stuff out, and hope you can make a lot of trouble for the damn place. I heartily disliked Alpha School based on how it was described in the review, and wrote several long posts about what I thought was wrong with its agenda and methods. I could smell, on Alpha School, the exact same competitive, exploitive stink I smell on Nevraumon. In fact once the matter of Nevraumon’s review is settled I will support your efforts energetically. I am sorry to pull back from doing that right now, but what happened with the voting disturbs and infuriates me, and I just can’t diffuse my focus right now.
I may be totally naive to the realities, but to address both of your points:
- in my circles (education) the Alpha School review "went viral" when it was posted as a finalist, as a link to the finalist post. It was actually the first review I read in the first round (alphabetical and one of a half dozen that jumped out at me), and even then there were people talking about it. By the time of JL's tweet, it was definitely already super talked about. In another comment I can't find, someone ballparked a few dozen votes as the difference it could have made, which doesn't seem wild to me.
- if YOU were a marketer - with all the latent skills and habits that implies - and YOU wanted to win a review contest, how would YOU write it? In the style that you are most familiar with, which happens to be known as being very persuasive? Or would you try a style that you're less familiar with, which is not as persuasive? Now instead of you, sub in the average person who works in marketing. I don't think it's off that it's written that way.
- GEO (generative engine optimisation) scares the shit out of me. I mean, it's a totally natural evolution of SEO, but it's like we're poisoning everything.. as soon as something becomes popular, we'll suck the value out of it and make it bad. (https://www.npr.org/sections/planet-money/2025/10/21/g-s1-94264/a-theory-why-the-internet-is-going-down-the-toilet)
Also:
Thanks to you both for introducing more skepticism into my brain regarding Alpha School. It's personally and professionally important to me to know how well their model works and I'm genuinely thankful that I'm looking at the hype as hype now.
Eremolalos, you said you have posts criticising Alpha, but I can't find them (could be a skill issue). Would you link to them so I can have a look?
As I said before, I am remaining neutral about the topic of Alpha Sleaze right now because I am preoccupied with the vote cheating issue and don't want to muddy the waters. But I do have a suggestion for you: If you are able to make a good case for Alpha Sleaze, see if you can interest a blogger or news site (Freddie deBoer? Vox?) in what you've dug up.
(from the Gemini review)
> Excellent question. This text is a fascinating example...
Sounds like sycophancy to me. While I do think the preponderance of evidence you share does point to a likelihood of this being some part of an organized ad campaign, I think Gemini is really forcing a 'yes'. It's examples seem cherry-picked and kind of a stretch. Really? Telling you me the correct information about time spent in class is actually a marketing trick?
But thank you for sharing this context.
I agree that divining whether the person got paid or how much coordination there was by putting the text into an LLM doesn't make much sense. The thing is that I prompted an LLM with a question like this long before the author's identity was revealed, and Gemini's text analysis was something like "the conversion journey, initial skepticism and slight critical points subtly reframed as positives, 'this is what Alpha PR wouldn't want you to know' etc are classical marketing psychology devices".
Its factual prediction always was "highly likely to be written by a marketing expert, but unclear whether someone paid or an enthusiastic 'super-parent'". This turned out to be precisely correct. As said, I believe Nevraumont when he says he didn't get paid in dollars - though indirect perks from Alpha involvement might play a role, e.g. high-profile opportunities like student demos at events with education leaders - easy to find via public coverage, but I'll avoid spotlighting minors here.
So now my pastebin is not so much about inferring the identity (known now) or cash compensation (got a denial that I believe) or quid pro quo ("quid" and "quo" are in public records, "pro" is up to the reader's judgment) of the person writing it, but about pointing out the devices in the text to immunize people reading it from them.
Regarding the opening, Gemini (without system prompt) always starts by sucking up to the user, but still often enough comes to a different conclusion than the user implied. My prompt for this reply was "Is this text likely some sort of hidden advertising? Can you tell me evidence in favour and against?", so it also put some evidence against into the reply.
Are you / have you published the names / substacks of the other review writers?
Can you please post literally any evidence that Alex King is not Aella beyond a pinky promise?
Aella would have included in her review observations about what men were like in bed.
I don't think this is a context where Aella would be reserved about admitting to being Aella.
Pitchfork.com has a "most clicked album reviews" section and it's dominated by album covers with nudity. OF COURSE the orgy review got the most clicks. Does anyone find this far too predictable to be noteworthy?
I am angry. Very, very angry right now, at parts of this community.
Not about who the winners were; I was pretty happy when I saw the list (I would have preferred The ACX Commentariat to win, but I'm relieved Dating Men in the Bay Area didn't win and thereby confirm voters put "validates my experience!!!" ahead of "actual insightful argument", and I'm also glad only one of the winners was at all on a topic that panders to the rationalist obsessions and biases--some of the previous years have made me strongly suspect that most people vote for reviews on topics they like that reach conclusions they agree with rather than for the ones that are most well written and best argued for--which is not something I can imagine a smart and fair person doing but which seems to be weirdly accepted as normal even here. I'm glad to have this suspician partly disproven).
But...two things are ofsetting that, making me really mad, and making me think this community isn't what I thought it was.
First, I made this point in detail two years ago (https://www.astralcodexten.com/p/book-review-contest-2023-winners/comment/40173308), and multiple other people made it in the same thread back then, and now I'm going to make it yet again: a voting system that lets you rank only three options out of 13 is not in any meaningful sense a ranked-choice system. In such a system, most votes will exhaust, making it almost the same as a First-Past-the-Post system. On the spectrum from FPTP to RCV, the system used here is very close to the FPTP end and eliminates most of the advantages of RCV, particularly with a large number of options that are all plausible contenders with significant appeal. Read my linked comment for more detail.
And this otherwise smart and detail-focused community seems...to not care about this at all. Apart from those of us actively arguing this point in the 2023 thread, virtually nobody has acknowledged this at all. Nobody has argued against it, but nobody has stopped calling it a ranked-choice vote, or speaking as though it has the features and effects of a ranked-choice vote. Neither Scott nor anyone else involved in running the contest has changed anything about the voting system in two years, nor have they said anything acknowledging its limitations, continuing to refer to it as ranked-choice voting without qualification. All these people who in other technical contexts are very careful about accuracy, and very sensitive to the way inattention to the details of studies and processes can have large ramifications on results...all these people seem to have, on a topic like this they're mostly not familiar with, no *knowledge* of essential details that change the whole effect of the system, no *patience* for understanding these details, and/or no *interest* in even acknowledging them.
And this...not only makes me disappointed in what I thought was a very rigorous and intelligent community, it also makes think...if this is the level of intellectual laziness the people on ACX display on a topic I have a bit of knowledge on (studied it a bit at university, worked at a few elections in a country that uses it), then I should probably radically discount the accuracy of much of what you all say on technical topics I'm not so familiar with. How much of what is said about things like AI around here, seemingly with strong consensus from lots of smart people, is built on the same level of indifference to detail and accuracy as discussions of voting systems?
The second thing I'm angry about is that a couple of people have pointed out that the Alpha School review seems to have benefited from vote stacking. And apart from like three or four comments...no one cares. There has been no deluge of people condemning this and saying it isn't okay, but also no real effort to argue that it is okay or have a substantive discussion about the ethics of it. Just...complete indifference. This thread is several days old; I highly doubt an influx of objections to this behaviour is on the way.
I'll change my tune if Scott actually addresses this situation and gives a clear statement on what he considers fair behaviour and why. But based on his reaction to almost every other comment about fairness and other concerns in the review contests over several years--that is, totally ignoring them, and showing no sign of taking these objections into account in any way over subsequent contests--I have very low hopes.
It's increasingly looking as though concern for ethical principles from this community and from Scott is, when it comes to rationalist groups, EA groups, and other groups seen as political or social allies (like the Alpha School), largely nonexistent.
I confess myself disappointed.
Hi!
So, let me start by saying that I respect you. I don't *know* you, but you seem very smart and very passionate, and that's the sort of person I respect.
That said, I don't really understand your concern. I've read the comment that you linked. And I've also read, over the years, a fair amount of pieces about different sorts of voting systems. I like to think that I'm fairly well-informed, although I don't pretend to be an expert. My overall impression is that that there is no one perfect voting system- everything has its pros and cons.
So I don't get why your so upset. Is it because Scott limited the ranked choices to just 3 choices? I really don't see why that's such a big deal. I can see how it *could* have been if, like, there had been 3 essays about the Ukraine war and just one about medieval history, but that doesn't seem to be what happened here at all. Most of the essays in the finalists seem fairly distinct. My own complaint is that there's too much of a gap between honorable mentions and finalists, over very small voting differences in the initial round.
Some people have raised concerns that the Alpha School founder was promoting that essay on his twitter. Others have raised concerns that the "dating men" essay was too much pandering to the types of people voting here. I'm personally concerned that the Ukraine war essay author seems to have removed it once he got too much attention.
In the end... none of that seems to matter? The one that one was none of those things. Just a long, meandering essay about boring medieval history, mixed with some weird theology. I personally ranked it 3rd, but was happy to see it win. My only complaint is that it was too long, but apparently a lot of people really like long essays.
So... what's your complaint here exactly? Do you think that that this sort of top-three ranked-voting system was unfairly biased? I'm an American, so I'm used to first past the post where we just vote for one and done, this seemed exceedingly fair in comparison.
Again, I respect your intelligence and passion, I just don't understand your argument.
Which criteria to determine a good non-book review?
For me it's (1) Informative (i.e: introduce something new and relevant to me); (2) Diversity of perspectives; (3) Highlight the constraints/limitation https://slatestarcodex.com/2013/03/07/we-wrestle-not-with-flesh-and-blood-but-against-powers-and-principalities/
With such criteria, "Red Means No" and Russo war are the best for me.