The price tag isn't the cost of AI
My recovered AI history contains 76.85 billion tokens. Most are cached input, which changes what the models cost and how much capability I get for the money.
How I work with coding agents: review their changes in layers, fix small friction with small tools, and check what the agent work really costs.
Connected topics, ranked by shared posts.
My recovered AI history contains 76.85 billion tokens. Most are cached input, which changes what the models cost and how much capability I get for the money.
Every check misses stuff. A bug only reaches production if all of them miss it, which is why I think AI code review needs different models with different prompts.
A small helper to switch between multiple Codex CLI accounts without constantly logging in and out.
'move fast and break things' is basically the default now, and AI-coding will crank that up even more. the real fix is boring: code review + managing agents well, so you don't ship the next outage or security mess.