if not working in ai safety, doesnt have educated opinion on dangers of ai? would imagine ai safety >= post train > pretrain but all are still significantly well positioned
Sorry, but asking Jacob about AI extinction risk is like asking your AC guy about climate change.
Not saying it's necessarily uninteresting or wrong per se but let’s keep things in perspective and hear from the full range of expertise across the ecosystem!
This young lady got her undergrad degree last spring. She was an intern at oAI until last September.
At this point we're going to have posts from Amodei's janitor on the need to be "concerned" about the pace of development get a million views.
Previously: an unlabeled graph — no x-axis, no y-axis, just vibes.
Today? Both axes are labeled. Clearly. Intentionally. Unambiguously.
And that? That’s growth. ✨
Figure is scaling up data in a big way. Last week we hit 69,900 weekly active users
Users are uploading 35 minutes of data every second across the world
@chris_j_paxton i can imagine astra is significantly better for computer use tasks but from a robotics and real world understanding pov, imo gemini has always been the best
@difficultyang i also feel that with open source, most non-coding tasks (and even coding ones in a lot of cases) dont actually need the shiniest, newest, can-do-all model. openrouter/auto works v well for the non technical ppl ive set stuff up for