Explorer
Newest first — browse tweet threads
Keep on to blur preview images; turn off to show them clearly
To my surprise, Opus 4.5 one-shot my hardest λ-calculus problem (tying with Gemini 3), and it did solve the stack underflow bug that an old checkpoint of Gemini 3 (NOT the deployed version) solved. So, in terms of first hour impression, that couldn't be more promising I guess...
Keep in mind these prompts were all over X last week. Could Opus be accessing the internet? So take this with a grain of salt. I'll design new variations and harder prompts when I find some time, possibly this weekend
RT @joshdbirdwell: Our team got to go to the @aiDotEngineer code conference this past weekend. We learned a lot. Solid talks and conversati…
achieve ambition with intentionality, intensity, & integrity - @dxtipshq - @sveltesociety - @aidotengineer - @latentspacepod - @cognition + @smol_ai
This isn't very good Gemini 3.5 will obliterate it Dario's b2b fortress is strong but not impregnable Inflection, though incomparably wretched, gave up in a similar situation - almost reached the frontier, outmatched on compute Nevertheless, "Ant is dead" postponed by 1 more year
We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1
This isn't very good Gemini 3.5 will obliterate it Dario's b2b fortress is strong but not impregnable Inflection, though incomparably wretched, gave up in a similar situation - almost reached the frontier, outmatched on compute Nevertheless, "Ant is dead" postponed by 1 more year
We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1
wish this had made it in time for AIE CODE but it’s out now! have been testing kevlar internally for a few weeks and people are VERY excited - this thing destroys @cognition’s internal held out benchmarks and is a notable step up in SOTA. Devin only gets an upgrade with step ups so am excited to see this roll out to everyone today.
achieve ambition with intentionality, intensity, & integrity - @dxtipshq - @sveltesociety - @aidotengineer - @latentspacepod - @cognition + @smol_ai
Another cool work from our team! Karpathy’s original quest of controlling computer with pixels in, keyboard+mouse out now can be done with just 7B model!
Mostly research and code. If universe is an optimizer, what is its loss function? All opinions are my own.