I built GTA Toronto + Uber driving with Codex for my new Modretro Chromatic that i got from OpenAI devday two weeks ago.
the most fun ive had in year. pure GTA II vibes. game is open source here:
enterprise maturity arc with AI = Dunning–Kruger
Stage 1: building with agent’s internally
Stage 2: a lot f*kin’ harder than we thought
Stage 3: take our $$$, here’s what we want; let’s go
Theres this local tech company that built a good enough agent harness a while back but never took off/had momentum. I realized only the founder/curated team that’s a little crazy that can move mountains to create crazy momentum. They didnt have that kind of ppl.
IMO the only tests you should have in the age of coding agents are, in order of priority:
1. Full E2E tests. Nothing mocked out at all. Can even be something that runs in prod on test accounts via Playwright or equivalent.
2. Integration tests. Agents make more mistakes as things bleed between data/API boundaries where schemas can drift.
3. Golden tests. This helps ground the code with real examples of data, and can be used as regressions against edge cases.
Unit tests are bloat 99% of the time if you’re using a frontier model since they’re smart enough now to get the implementation (and subsequent iterations) right on the first shot, so they only add bloat at this point.
updates on our homebrew bookeeping software:
- ollama support and can run local 100% (free AI powered quickbooks)
- works with openjev
- open-sourced next week