Verbatim Index: LLM judges blind-rank 50 tech leaders on the thinking in their public speech
tonygwu · x · 2026-10-01
Tony Wu open-sourced Verbatim Index, which ranks 50 well-known tech leaders purely on the thinking their public speech demonstrates — graded blind by three LLM judges from verbatim transcripts of interviews, podcasts and keynotes, with company performance and reputation excluded. Full rankings, methodology, transcripts, code, rubric and harness are on GitHub. A sibling project, Verbatim Predictions, converts the same transcripts into an auditable index of falsifiable predictions, each grounded in a timestamped quote.
Related event: LLM blind review of 8.8M interview words ranks LeCun top for insight(3 posts)→
More from Fun
- 'We Use Agents to Build' — Or Is It 'Asians'? Dylan Patel's Global Economy Joke — dylan522p · 2026-10-01
- Sentdex baffled Hugging Face still hasn't built a coding agent — Sentdex · 2026-10-01
- "Babe calm down, Juliet is my OpenAI personal agent" meme goes around — tech__unicorn · 2026-10-01
- Autonomous company platform Polsia stumbles: onboarding failure leaves test stuck for over a week — Fun-Telephone-681 · 2026-10-01
- "I Know Kung Fu" Is Really Just Reading, Says Author Betting Against Neuralink — StewartalsopIII · 2026-10-01
- Roboticist questions Figure's lava robot tests as unrealistic for home safety — MarwaEldiwiny · 2026-10-01