Dictated 203,000 words in 44 days. Finished almost nothing. The interesting result wasn't output — it was a different posture on the human side of @karpathy's boundary. Voice + LLM as a thinking instrument, not voice-to-text. Wrote it up:
- The hard part of agentic AI isn’t the agent. It’s the boundary. @karpathy autoresearch hints at what most people are missing: one agent, one spec, one metric. If a human can’t define success clearly, the AI can’t improve it. That’s the real bottleneck.
- I recently spoke with Jeffrey Emanuel @doodlestein who is widely attributed with catalyzing the recent $2 trillion dollar market correction in the wake of DeeSpeek. Lots of additional context on AI valuations.
- @nateliason book on the technology and psychology behind the crypto boom is a cautionary tale and manual for best practices. Here is my take on it.
- OpenAI CEO @sama wants to spend $7 trillion on an AI chip maximizer that promises to scale AI faster. How about we #acceleratetrust instead? Here are some reasons why accelerating trust frameworks may scale AI even quicker.
