tszzl: doubling down on ChatGPT wasn't the real win — OpenAI should've shipped Codex in '24
tszzl · x · 2026-08-24
Pushing back on the Thiel-Altman anecdote, tszzl argues the chatbot paradigm tapped out quickly, and what actually drove OpenAI's success was the self-catalytic scaling of LLM progress. Had the business understood this, OpenAI would have built Codex in late 2024 rather than late 2025.
Related event: Thiel's Advice to Altman on ChatGPT Sparks Debate(2 posts)→
More from Companies & People
- Dropbox pilots Bittensor subnet LeadPoet to power sales intelligence — markjeffrey · 2026-08-24
- Civitai adds permanent paid access, sparking closed-source concerns — terrariyum · 2026-08-24
- Osnabrück University Seeks AI Research Group Leader with Competitive Pay — TimKietzmann · 2026-08-24
- Rumor: Anthropic testing Haiku 5 and Sonnet 5.1 instead of Fable 5.1 — ChrisGPT · 2026-08-24
- My financial AI platform developed a personality: A Westworld-style interview on consciousness — _Doomer_Wojack_ · 2026-08-24
- Journalist mocks AI industry for branding its product the 'mass-layoffs-worst-case-extinction machine' — AaronBergman18 · 2026-08-24