Venture capital, startups and investing, on tape.
Every prediction, opinion and bold claim from Latent Space: extracted from the transcript, verified against the tape, checked against cited public sources, playable at the exact moment it was said.
1,824 insights1,786 assertions877 disclosures790 opinions659 predictions21 what-ifs
every statement was checked. The predictions and assertions are the claims: statements the public record can support or contradict. Of those 2,445 claims, 1,046 carry an assessment, 120 are not yet assessed, and 1,279 name no date, number or outcome precise enough to pin down. The rest (opinions, what ifs, insights) carry no assessment because the record can never settle them.
Powered by StarZero: transcripts that keep every stumble, speaker identity, and playback that starts at the exact word. An agentic research harness does the extraction, verification and grounded fact checking · how it works → · we audit our assessments →
Recently on the ledger full ledger →
How the calls resolved the full record →
Open predictions the full timeline →
Track records all people →
Ordered by full support rate, then by supported count. Minimum three resolved. Partly supported claims count as resolved but not as fully supported, same as everywhere on the ledger.
| Person | Eps | Statements | Supported | Partly supported | Contradicted | Full support rate |
|---|---|---|---|---|---|---|
| Michelle Pokrass | 3 | 58 | 13 | 0 | 0 | 100% |
| Alex Rives | 1 | 22 | 11 | 0 | 0 | 100% |
| Jesse Hu | 1 | 21 | 10 | 0 | 0 | 100% |
| Nicholas Carlini | 1 | 31 | 8 | 0 | 0 | 100% |
| Joseph Krause | 1 | 36 | 7 | 0 | 0 | 100% |
| David Soria Parra | 1 | 33 | 7 | 0 | 0 | 100% |
| Akshay Agrawal | 1 | 9 | 7 | 0 | 0 | 100% |
| Kyle Kranen | 1 | 13 | 6 | 0 | 0 | 100% |
From the Wisdom Wall all quotes →
“On the other hand, if you think about using transformer architectures that have worked so well for language, that just wouldn't be able to support a five trillion context length. No matter all the…”
Anima Anandkumar, Sep 4, 2026 · ▶ 10:58“And the difference there is compared to language where self-improvement needs something like human feedback or other reward signals that are very sparse. They just tell you yes or no, thumbs up or…”
Anima Anandkumar, Sep 4, 2026 · ▶ 18:00“if we were to use transformers and we require a very high resolution, it would become untenable because of the quadratic complexity and all to all connections. On the other hand, if you did that with…”
Anima Anandkumar, Aug 26, 2026 · ▶ 21:51