Weekly Top Picks #49

AI-generated video propaganda / AI industry gets an F on transparency / The BAH method / GPT-4V + DALL-E 3 loop / Will GPT-5 meet expectations? / Techno-optimism / Analogy of raw models vs. APIs

This week at a glance

• AI-generated political propaganda: Jill Biden calls for Gaza ceasefire

• How transparent are foundation model developers? Hint: Not much

• The Best Available Human: A pragmatic approach to AI

• The GPT-4V - DALL-E 3 recursive loop

• OpenAI stopped working on Arrakis — what does that mean for GPT-5?

• The Techno-Optimist Manifesto and responses

• A good analogy to think of raw models (e.g. Llama) vs. APIs (e.g. GPT-4)