LessWrong 2.0 Reader

1. Age is extremely compressed/skewed because it's OKCupid. So I can think of a couple issues there: there might be a problem of distribution mismatch where a GPT is trained on a much more even distribution of text (I would assume tons of text is written by age 50-100 IRL rather than a young techie dating website) and so is simply taking into account a very different base rate; another issue is that maybe the GPT is accurate but restriction of range creates misleading statistical artifacts. Binarization wouldn't help, and might worsen matters - how many people tweak their age to avoid the dreaded leading '3' and turning into Christmas cake?

   As far as sexuality goes, this is something the LLMs may be trained very heavily on, with unpredictable effects. But it's also a much weirder category here too:

   Dating sites in general have more males than females, reflecting the mating behavior seen offline (more males being on the lookout). OKCupid features a very broad selection of possible genders. One must choose at least one category and up to 5 categories of which the possible options are: Man, Woman, Agender, Androgynous, Bigender, Cis Man, Cis Woman, Genderfluid, Genderqueer, Gender Nonconforming, Hijra, Intersex, Non-binary, Other, Pangender, Transfeminine, Transgender, Transmasculine, Transsexual, Trans Man, Trans Women and Two Spirit. Nevertheless, almost everybody chooses one of the first two (39.1 % Women, 60.6 % Men, binary total = 99.7 %)^5. The full count by type can be found in the supplementary materials sheet "Genders").

   I'm not sure how OP handled that. So the predictive power here should be considered as a loose lower bound, given all the potential sources of measurement error/noise.

I have indeed been publicly advocating against the inside game strategy at labs for many years (going all the way back to 2018), predicting it would fail due to incentive issues and have large negative externalities due to conflict of interest issues. I could dig up my comments, but I am confident almost anyone who I've interfaced with at the labs, or who I've talked to about any adjacent topic in leadership would be happy to confirm.

I didn't know it meant either.

If the strategy failed in predictable ways, shouldn't we expect to be "pre-registered" predictions that it would fail?

Oh, weird. I always thought "ETA" means "Edited To Add". 

ETA = edit time addition

I should probably not use this term, I think I picked up this habit from some other people on LW.

I agree with ~all of your subpoints but it seems like we disagree in terms of the overall appraisal.

Thanks for explaining your overall reasoning though. Also big +1 that the internal deployment stuff is scary. I don’t think either lab has told me what protections they’re going to use for internally deploying dangerous (~ASL-4) systems, but the fact that Anthropic treats internal deployment like external deployment is a good sign. OpenAI at least acknowledges that internal deployment can be dangerous through its distinction between high risk (can be internally deployed) and critical risk (cannot be), but I agree that the thresholds are too high, particularly for model autonomy.

"ETA" commonly is short for "estimated time of arrival". I understand you are using it to mean "edited" but I don't quite know what it is short for, and also it seems like using this is just confusing for people in general.

Who is updating? I haven't seen anyone change their mind yet.

Wasn't edited, based on my memory.