Time & horizons · Article 29 of 60
The short fuses
The claims that can be settled soonest, and why they matter most.
Short-dated claims are the ledger's real test. They cannot be rescued by reinterpretation, and they generate the evidence that later claims are judged against.
If the graded column ever grows quickly, it will be because of this set, not because the grand claims about 2040 finally resolved.
| Person | Said on | Due | Days |
|---|---|---|---|
| Elon Musk | 2026-01-22 | 2026-12-31 | 343 |
| Mark Zuckerberg | 2025-01-10 | 2025-12-31 | 355 |
| Sam Altman | 2025-01-05 | 2025-12-31 | 360 |
| Sundar Pichai | 2025-12-07 | 2026-12-07 | 365 |
| Dario Amodei | 2025-03-10 | 2026-03-10 | 365 |
| Narendra Modi | 2026-08-15 | 2027-08-15 | 365 |
| Clément Delangue | 2026-08-03 | 2027-12-31 | 515 |
| Mustafa Suleyman | 2025-12-12 | 2027-06-12 | 547 |
| Person | Claim | Due | Days |
|---|---|---|---|
| Mark Zuckerberg | <em>“By roughly October 2026, most code written for Meta’s AI efforts is authored by AI agents.”</em> | 2026-10-29 | 4 |
| Sundar Pichai | <em>“Within roughly a year, AI agents will routinely act on a user’s behalf for high-stakes decisions.”</em> | 2026-12-07 | 43 |
| Elon Musk | <em>“AGI — AI smarter than any human — arrives by the end of 2026, “no later than next year”.”</em> | 2026-12-31 | 67 |
| Mustafa Suleyman | <em>“AI systems are already superhuman in the tasks that matter, with agentic reliability arriving within 6–18 months.”</em> | 2027-06-12 | 230 |
| Jan Leike | <em>“The core technical challenges of aligning superintelligence would be solved by mid-2027, with 20% of OpenAI’s compute.”</em> | 2027-07-05 | 253 |
| Narendra Modi | <em>“Ten million young Indians receive AI skills training within one year — by August 2027.”</em> | 2027-08-15 | 294 |
Figures computed from docs/predictions/ledger.json at build time. Every claim quoted on this page carries its own sources in the ledger.