Author: Jose S
Ask, edit, test, review loop
Across IDE assist and terminal agents, keep the same loop: clear ask, small diffs, tests that matter, human review before merge. The surface changes. The loop does not.
System One Models: The Fast, Non-Autoregressive Decision Layer for Software
System One models process program state in a single parallel pass to output typed, probabilistic decisions rather than chat prose. Here is how they eliminate hallucinations and bring…
Chat APIs for open models: Groq, Together, Fireworks
Hosted APIs for open weights bill by tokens and still send prompts off your machine. Separate Groq from Grok, set a spend cap, and treat retention pages as…
Terminal agents: CLI and Antigravity mental model
Terminal agents like Gemini CLI and Antigravity CLI can edit a real repo. Use a feature branch, a small goal, plan-first habits, and a human merge. Treat mid-2026…
Sharing a local setup with family or a small team
A kitchen laptop running a local AI model is a shared brain: same login means same chat history. Separate users, keep the app on localhost, and quit it…
