Open Thread 386

...

This is the weekly visible open thread. Post about anything you want, ask random questions, whatever. ACX has an unofficial subreddit, Discord, and bulletin board, and in-person meetups around the world. Most content is free, some is subscriber only; you can subscribe here. Also:

1: Thanks to everyone who commented on P-Zombies Would Report Qualia. Although as usual most of your attempts to link it to other philosophical writings were vague and not very relevant, a correspondent was eventually able to find a very similar argument starting on page 289 of David Chalmers’ The Conscious Mind (which I had not previously read). Thanks to Professor Chalmers for not getting upset about this unintentional duplication of his work. And if you think you have a more original insight about consciousness, Erik Hoel highlights a $50,000 consciousness essay contest.

2: Thanks to everyone who commented on The Claude Bliss Attractor. Two especially good categories of comment - one arguing that the tendency of iterative AI art to produce black people was caused not by a bias towards diversity, but by AI art adding a dark sepia tone to everything - as the figures in the art get darker and darker, the AI interprets them as ethnically African and adds black features to “match” (and see this comment by Michael) - and another arguing that even though Claude had a slight bias towards being a hippie, it has slight biases towards lots of things (eg coding) and it still requires more explanation for why the hippie spirituality is what eventually shines through.

3: New rule: if you post a link in an open thread (as a standalone comment advertising the thing being linked to, not as a source for something you’re saying), please also include at least two paragraphs of original commentary on the link and discussion of why you think it’s interesting, as “proof of work” that you’re willing to put effort into promoting this and aren’t just spamming us with links. Even with this proof-of-work, please try to avoid having more than one Open Thread link per few months.

4: Another AI alignment summer workshop in Europe, this one Prague, July 22 - 25, apply here.