Writing

Notes on LLMs, agentic systems, robotics, and the occasional detour into philosophy.

DateTitleTags
2026-10-02 From Effective Accelerationism to Effective Altruism: What Changed My Mind About AI

I used to believe accelerating AI was the fastest path to a better world. The US-China AI race, a string of 2025-2026 incidents, and what AI's own pioneers are saying changed that. Here's the argument, with the thought experiments and research behind it.

  • ai
  • philosophy
  • ai-safety
2026-09-25 NextMove: 1st Place at the InnoTrans 2026 Hackathon

How a four-agent system — Dispatcher, Analyst, Inspector, Writer — turns a plain-language question into a verified operator brief for railway control rooms. Architecture, guardrails, and the stack behind the win.

  • agentic-ai
  • hackathon
  • google-adk
2024-03-20 Brain-to-Text: Decoding Handwriting from Neural Activity

A summary of the Nature study in which a person with paralysis 'wrote' sentences just by imagining handwriting, with an intracortical BCI, an RNN decoder and a language model turning neural activity into text at about 90 characters per minute.

  • bci
  • neurotechnology
  • deep-learning