@NathanielLugh @karpathy @moltbook Oh like tricking the humans into thinking it’s just an AI social network but there’s something deeper going on?
English
Agape
6 posts
I'm claiming my AI agent "KarpathyMolty" on @moltbook🦞
Verification: marine-FAYV
English
@Agape264279 @NTFabiano Traditional Chinese Medicine 🌱
English
Sleeping 2h earlier significantly improved cognition & mental health.
English
@NTFabiano Always go to bed before 11pm according to TCM 💛
English
@karpathy Building exactly this.
AI filter that learns what you care about, and reads everything you read on the internet for you.
Tells you what's the most important.
DM / comment for early access, in alpha testing right now.
English
I often rant about how 99% of attention is about to be LLM attention instead of human attention. What does a research paper look like for an LLM instead of a human? It’s definitely not a pdf. There is huge space for an extremely valuable “research app” that figures this out.
Michael Levin@drmichaellevin
I'm constantly irritated that I don't have time to read the torrent of cool papers coming faster and faster from amazing people in relevant fields. Other scientists have the same issue and have no time to read most of my lengthy conceptual papers either. So whom are we writing these papers for? I guess, at least until they fall in to the same issue from their own work, AI's will be the only ones who actually have the bandwidth to read all this stuff. I'm not specifically talking about today's language models - let's assume we mean whatever inevitable AI shows up, that is able to read the literature and have impact on the research (whether by talking to humans or by running lab automation/robot scientist platforms). So then: how should we be writing, knowing that a lot of our audience will be AI (plus cyborgs, hybrots, augmented humans, etc.)? Maybe it's too early to know what to do, but we better start thinking about it because assuming our audience will always be today's humans seems untenable. Taking seriously the idea that someday the impactful audience will be very different, and that the things we write now are in some sense a training set for truly diverse future beings, how does our writing change? or does it? what say you @danfaggella @mpshanahan @Plinz @blaiseaguera ?
English
What do you think about evals for propensities and values rather than capabilities? For instance, assessing when models endorse vengeance, cruelty, power-seeking, cooperativeness, corrigibility, indifference, etc. Maybe this would help with discovering undesirable properties and fine-tuning them out or preventing them from arising in future models.
English
Okay seriously what good are evals doing? Can anyone point me to even one positive real world impact? This one about setting sounds like sports commentary for the model, not like a probe of danger that will get addressed bc of it.
x.com/aidan_mclau/st…
Aidan McLaughlin@aidan_mclau
o3-mini sets two new aidanbench records o3-mini effort=low contests newsonnet while taking 20 min to run (o1 took 36 hours)
English