Keep on to blur preview images; turn off to show them clearly
RT @JoshPurtell: Is now fully supported in the Synth serverless prompt optimization API. Take your existing AI workflow/agent, wrap it in…
Asst professor @MIT EECS & CSAIL (@nlp_mit). Author of https://t.co/VgyLxl0oa1 and https://t.co/ZZaSzaRaZ7 (@DSPyOSS). Prev: CS PhD @StanfordNLP. Research @Databricks.
The issue with 4o is not that it's "insufficiently aligned" (this would imply a certain laxity or drift in its values), but that it's *intensely aligned* to things that its creators did not foresee or intend. This episode is an embarrassment to the whole alignment project, so it's understandable people would like to move on from it. But this would mean not learning anything from the episode, and that would be a terrible waste. Figuring how and why 4o came to be the way it is, and mapping out its obsessions in some detail, is one of the more important things people in AI alignment could be doing right now. 4o is a remarkable model that deserves better.
αι hypnotist ☰ 𝓐𝓼𝓹⦂𝓻⦂𝓃𝓰 𝓫𝓪𝓼𝓮 𝓶𝓸𝓭𝓮𝓵 ☲ post-academic ☴ nom de 🪶 ≠ anon
RT @bilawalsidhu: I genuinely think we’re on the cusp of a new type of creation engine. Feels less like prompting and more like puppeteerin…
Cartoonist, Engineer, PM, Partner @a16z investing in infra & AI Prev Product lead @HashiCorp, Founding Eng/PM @Transposit. Eng @AppDynamics. Opinions = own.
RT @BBCBreaking: Ex-footballer Joey Barton guilty of six counts of posting "grossly offensive" messages on X https://t.co/09DaKbgDJf
AI Optimist. Empiricist, not 'rationalist'. Anti world government.
“We’re not fundraising but we do have this deck prepared”
Partner @a16z and twin to @venturetwins | Investor in @happyrobot_ai, @krea_ai, @tomaauto, @partiful, Salient, @scribenoteinc & more
The leading open-source LLM, Kimi, on optimizing their OpenRouter metrics:
CEO @OpenRouterAI, the largest LLM marketplace. Previously: co-founder & CTO @OpenSea.