All-In, under review / Episode archive
Big claims.
A closer look.
Six episodes. Four hosts. Claims checked against the evidence.
The latest review · #291
Where the hosts stand
Scores reflect the claims reviewed. See rubric.
The longer view
Host scores over time
Toggle hosts. Select a point to open its review. Spacing follows episode order.
View scores and sample sizes
| Episode / date | Jason | Sacks | Friedberg | Chamath |
|---|---|---|---|---|
| #286 2026-08-21 | 55 (3 claims) | 55 (4 claims) | 70 (4 claims) | 50 (3 claims) |
| #287 2026-08-28 | 60 (3 claims) | 75 (4 claims) | 50 (4 claims) | 65 (3 claims) |
| #288 2026-09-04 | 70 (3 claims) | 45 (4 claims) | 50 (4 claims) | 45 (3 claims) |
| #289 2026-09-11 | 60 (3 claims) | 60 (4 claims) | 45 (4 claims) | 50 (3 claims) |
| #290 2026-09-25 | 55 (3 claims) | 50 (4 claims) | 60 (4 claims) | 50 (3 claims) |
| #291 2026-10-02 | 60 (3 claims) | 55 (4 claims) | 50 (4 claims) | 50 (3 claims) |
The evidence trail
Six episodes. Open to scrutiny.
Regular panel episodes.
Trump's Super Intelligence Summit, AI Safety Accord, GDP Beats, Midterm Predictions
Anthropic IPO at Risk, Meta's Muse Pop, Token Prices Fall, Open Source Gains Share, Alignment Fails
AI Kills Everybody or Doomer Psyop? OpenAI's Math Breakthrough, Nike's $200B Collapse
GPT-6 Hits AGI? Tech Euphoria 2.0, SF Mansion Shortage, NYC Bans AI in Schools & Venezuela Oil Deal
Nvidia's Historic Quarter, SaaS Comeback, Bessent vs Druck, America's Debt Crisis, Cancer Vaccine
Dario Defends Himself, Datacenter Panic, AI Doomer Trap, Senate Toss-Up
How the scores work
One rubric.
Visible judgments.
Five dimensions scored 0–10. Weighted total converted to 100 and rounded to the nearest five. Evidence balance assesses cherry-picking, omitted counterevidence, inconsistent standards and misrepresented alternatives.
Letter grades use the rounded score: A ≥80, B ≥70, C ≥60, D ≥50, F <50.
Download all claims, dimensions and sources ↗Review scope & sources
Editorial scores cover selected claims, not overall accuracy. Each episode reviews 14 claims, with three or four per host. All six episodes use rubric v2, including evidence balance at 15%. Small changes are not statistically meaningful. Summit interviews are excluded.
Attribution uses automated transcripts and has not been fully audio-audited. Primary sources support each assessment; current reference pages provide retrospective context. Forecasts are graded on their support. Method v2.0.