En 30 segundos
Theo calls OpenAI’s GPT-6 Astra a next-generation model, citing its apparent range across coding, computer use, 3D modeling and office work. He also says it can over-engineer and struggle to escape loops, although he finds that problem less severe than in Sol.
- 1A next-generation modelTheo says OpenAI has delivered an impressive model, while acknowledging that it still has rough edges.
- 2A broad range of workHe highlights coding, computer use, 3D modeling and office work as areas where the model stands out.
- 3Over-engineering and loopsIn his experience, the model still tends to over-engineer and can struggle to escape loops, although he considers it much less prone to this than Sol.
Conclusiones
- Theo considers GPT-6 a next-generation model with impressive capabilities.
- He praises its performance across coding, computer use, 3D modeling and office work.
- He says it still over-engineers and can get trapped in loops, though less severely than Sol.
Lee el texto editado por capítulos
Strong capabilities, lingering rough edges
Theo praises GPT-6 as a next-generation model across several kinds of work. He also describes over-engineering and getting stuck in loops as remaining weaknesses.
GPT-6 is the next generation. OpenAI has done it: this model is incredible, although it still has rough edges. From code and computer use to 3D modeling and office work, it truly feels built different.
It still has a problem with over-engineering. It is nowhere near as bad as Sol, but it does tend to get trapped in loops and struggles to get out of them.
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.