Local AI’s Future Will Probably Be ‘Braided’

Fun, and challenging, to be back at the technology frontier. But it’s the real deal. LLMs blew the old paradigm away the minute Jeff Dean published his initial work; the Satoshi white paper is actually quaint and vanilla stuff by comparison.

Bitcoin was perceived as uncool for years, yet that’s when the growth and foundational roots emerged. Local AI could be over-regulated or made onerously expensive in the future, yet for now it’s an open horizon: new hardware and users, without the use cases yet figured out or monetized. And by pushing responsible and yet impressive capabilities today, the path dependencies all change: a bad outcome becomes LESS likely as local LLM downloads increase, not more likely.

https://intersignal.org | https://ollama.com | https://python.org/downloads - make sure you have Python 3.10 or above before using Intersignal Braid or Ollama. Mistral, Qwen, and Gemma4 all work well with Braid (and are available for free in Ollama).

Get full access to David Seaman's Deep at davidseaman.substack.com/subscribe