En 30 segundos
Theo says an AI-assisted audit of recently updated T3 Code pull requests produced useful insight for 12 cents. He proposes running a bot every three hours to flag PRs it considers ready to merge. The spoken token totals differ from the figures visible in the on-screen snapshot, so they should not be assumed to describe the same measurement.
- 1A low-cost auditTheo describes getting useful insight for 12 cents from a recent PR audit.
- 2Scope and usageThe request covered recently updated T3 Code PRs. The spoken token counts and the displayed usage snapshot do not match.
- 3A proposed recurring botHe plans a bot that checks periodically and alerts him to PRs it considers ready to merge.
Conclusiones
- Theo asked an agent to audit T3 Code pull requests updated in the previous five days.
- He describes the audit as costing 12 cents and proposes repeating it every three hours.
- The spoken token figures differ from the visible snapshot, and the clip does not explain the discrepancy.
Lee el texto editado por capítulos
Auditing PRs and planning a recurring check
Theo describes a low-cost audit of recently updated pull requests. He proposes automating a readiness check every three hours.
I got a useful insight for 12 cents. I did not just ask the agent to audit my own pull requests: I asked it to examine the open T3 Code PRs updated in the last five days.
I say the audit used six billion input tokens, mostly cached, and 300,000 output tokens. The usage snapshot shown in the clip instead displays about 6.23 million input tokens and 30,642 output tokens, with a pricing calculation being revised; the clip does not explain whether these are the same measurement.
I plan to set up a bot to repeat this every three hours and notify me when it thinks a pull request is ready to merge. I expect that to help us ship faster.
Más de este canal
38:47EN
Theo - t3․gg· 28 oct 2026Theo’s Long-Running Agent Experiments: Verification, Compute and CostTheo Browne · Ping LabsA skateboard game and two compiler projects show how repeated checks and flexible compute can sustain agent work—and how expensive those loops can become.
36:07ENTheo - t3․gg· 27 oct 2026Theo on Token Costs, Coding Agents, and the Price of AI CapabilityTheo Browne · Ping LabsA viral estimate of Lauren's AI spending prompts a closer look at cache pricing, benchmarks, and the changing work of coding agents.
30:37ENTheo - t3․gg· 26 oct 2026Theo’s Three Skills for Reviewing and Shipping with Coding AgentsTheo Browne · Ping LabsHow sol-loop, full-send and full-stop encode review, verification and different levels of merge authority.
0:43ENTheo - t3․gg· 23 oct 2026Theo Distinguishes Security Vulnerabilities from Active AttacksTheo Browne · Ping LabsA short explanation of why disclosed software flaws and reported attacks are different kinds of security news.
29:22ENTheo - t3․gg· 22 oct 2026Theo on Security Vulnerabilities, Stolen API Keys and AITheo Browne · Ping LabsA look at the gap between vulnerability disclosure and patching, and a speculative account of why stolen inference keys may be valuable.
31:13ENTheo - t3․gg· 20 oct 2026Theo on Why Smarter Models Help More Than DevelopersTheo Browne · Ping LabsFrom game modifications to compiler experiments, Theo revisits his assumptions about who benefits from more capable AI.