This weekly roundup thread is intended for all culture war posts. 'Culture war' is vaguely defined, but it basically means controversial issues that fall along set tribal lines. Arguments over culture war issues generate a lot of heat and little light, and few deeply entrenched people ever change their minds. This thread is for voicing opinions and analyzing the state of the discussion while trying to optimize for light over heat.
Optimistically, we think that engaging with people you disagree with is worth your time, and so is being nice! Pessimistically, there are many dynamics that can lead discussions on Culture War topics to become unproductive. There's a human tendency to divide along tribal lines, praising your ingroup and vilifying your outgroup - and if you think you find it easy to criticize your ingroup, then it may be that your outgroup is not who you think it is. Extremists with opposing positions can feed off each other, highlighting each other's worst points to justify their own angry rhetoric, which becomes in turn a new example of bad behavior for the other side to highlight.
We would like to avoid these negative dynamics. Accordingly, we ask that you do not use this thread for waging the Culture War. Examples of waging the Culture War:
-
Shaming.
-
Attempting to 'build consensus' or enforce ideological conformity.
-
Making sweeping generalizations to vilify a group you dislike.
-
Recruiting for a cause.
-
Posting links that could be summarized as 'Boo outgroup!' Basically, if your content is 'Can you believe what Those People did this week?' then you should either refrain from posting, or do some very patient work to contextualize and/or steel-man the relevant viewpoint.
In general, you should argue to understand, not to win. This thread is not territory to be claimed by one group or another; indeed, the aim is to have many different viewpoints represented here. Thus, we also ask that you follow some guidelines:
-
Speak plainly. Avoid sarcasm and mockery. When disagreeing with someone, state your objections explicitly.
-
Be as precise and charitable as you can. Don't paraphrase unflatteringly.
-
Don't imply that someone said something they did not say, even if you think it follows from what they said.
-
Write like everyone is reading and you want them to be included in the discussion.
On an ad hoc basis, the mods will try to compile a list of the best posts/comments from the previous week, posted in Quality Contribution threads and archived at /r/TheThread. You may nominate a comment for this list by clicking on 'report' at the bottom of the post and typing 'Actually a quality contribution' as the report reason.
Jump in the discussion.
No email address required.
Notes -
It costs about $2000 per problem, and American labs have that money to burn (they only publicize positive results), while Chinese labs do not.
Also, Chinese models are by and large worse than American frontier models, more than the benchmarks would suggest, and sometimes they just lack the special spark that the current models have.
(Though, it is clear that frontier models weren't necessary for all these math results. People have asked older versions of Claude and ChatGPT, and been able to rediscover the same proofs. So we didn't need the absolute latest models to find all of these.)
I've not seen these recreations; was it as simple as stating the problem, and then they managed? Or was some hint or framing provided?
Either way, your point stands; I'd expect any result by the frontiers that can be solved in X attempts to be reproducible by one a couple months back in ~10X attempts.
This one was 2 pages of (somewhat generic) prompt plus an hour of compute:
https://aaronlou.com/jacobian_counterexample_prompt.pdf
I'd like to see more than just the summarized CoT+results though. In this case the agent was instructed not to look for existing results on the conjecture, but did have access to search to find useful lemmas and might have concealed it if it ran across hints to the counterexample. I've recently seen a case like that, where a model's CoT recognized the problem (IIRC old USAMO) it was asked and it had memorized the answer, but the output rederived the answer without ever mentioning that.
More options
Context Copy link
Unverifiable, but the recreation I heard of didn't use any specific knowledge, and was shielded from recent news. The model+harness was pointed to the (known to be solveable) problem, and proceeded to solve it.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link