People & institutions · Article 22 of 60
Researchers and evaluators
The people whose job is to measure, test and disagree.
Evaluators are the group whose claims are most likely to be falsifiable in a mundane way: a benchmark, a replication, a threshold. That makes their records the easiest in the collection to close.
It also makes them the group most likely to be right for reasons that do not generalise, which is why the ledger grades the claim rather than the reputation.
| Person | Organisation | Group | Claims |
|---|---|---|---|
| <div class="person"><span class="face sigil" aria-hidden="true">AN</span><div><strong>Andrew Ng</strong><br /><span class="tiny">Founder, DeepLearning.AI; managing GP, AI Fund · DeepLearning.AI</span></div></div> | DeepLearning.AI | Researchers & evaluation | 2 |
| <div class="person"><span class="face sigil" aria-hidden="true">AN</span><div><strong>Arvind Narayanan</strong><br /><span class="tiny">Professor; director, Center for Information Technology Policy · Princeton University</span></div></div> | Princeton University | Researchers & evaluation | 2 |
| <div class="person"><span class="face sigil" aria-hidden="true">BB</span><div><strong>Beth Barnes</strong><br /><span class="tiny">Founder and CEO · METR</span></div></div> | METR | Researchers & evaluation | 2 |
| <div class="person"><span class="face sigil" aria-hidden="true">GH</span><div><strong>Geoffrey Hinton</strong><br /><span class="tiny">University Professor Emeritus · University of Toronto</span></div></div> | University of Toronto | Researchers & evaluation | 2 |
| <div class="person"><span class="face sigil" aria-hidden="true">YB</span><div><strong>Yoshua Bengio</strong><br /><span class="tiny">Scientific director, LawZero; Université de Montréal professor · LawZero / Mila</span></div></div> | LawZero / Mila | Researchers & evaluation | 2 |
| <div class="person"><span class="face sigil" aria-hidden="true">AD</span><div><strong>Anca Dragan</strong><br /><span class="tiny">Professor; safety lead · UC Berkeley; Google DeepMind</span></div></div> | UC Berkeley; Google DeepMind | Researchers & evaluation | 1 |
| <div class="person"><span class="face sigil" aria-hidden="true">AK</span><div><strong>Anton Korinek</strong><br /><span class="tiny">Professor · University of Virginia</span></div></div> | University of Virginia | Researchers & evaluation | 1 |
| <div class="person"><span class="face sigil" aria-hidden="true">CR</span><div><strong>Christopher Ré</strong><br /><span class="tiny">Professor · Stanford</span></div></div> | Stanford | Researchers & evaluation | 1 |
| <div class="person"><span class="face sigil" aria-hidden="true">DS</span><div><strong>Dawn Song</strong><br /><span class="tiny">Professor · UC Berkeley</span></div></div> | UC Berkeley | Researchers & evaluation | 1 |
| <div class="person"><span class="face sigil" aria-hidden="true">EB</span><div><strong>Erik Brynjolfsson</strong><br /><span class="tiny">Professor; director · Stanford HAI</span></div></div> | Stanford HAI | Researchers & evaluation | 1 |
| <div class="person"><span class="face sigil" aria-hidden="true">JS</span><div><strong>Jacob Steinhardt</strong><br /><span class="tiny">Professor · UC Berkeley</span></div></div> | UC Berkeley | Researchers & evaluation | 1 |
| <div class="person"><span class="face sigil" aria-hidden="true">JL</span><div><strong>Jan Leike</strong><br /><span class="tiny">Leads the Alignment Science team · Anthropic</span></div></div> | Anthropic | Researchers & evaluation | 1 |
| Claim | Horizon | Verdict |
|---|---|---|
| <div><em>“Instead of discussing faulty, unreliable systems, tech leaders are marketing their systems as all-knowing, superpowerful, superintelligent, and lobbying for self-policing and…”</em><br /><span class="tiny">Timnit Gebru — Democracy Now! interview after the Right Livelihood Award, 2026-10-01</span></div> | Current policy cycle | <span class="verdict" data-v="contested">Contested</span> |
| <div><em>“The practical utility of very long-horizon agent tasks, especially relative to their token cost, is overhyped.”</em><br /><span class="tiny">Andrew Ng — DeepLearning.AI essay series on coding agents, 2026-09-04</span></div> | Through 2027 | <span class="verdict" data-v="contested">Contested</span> |
| <div><em>“Going forward, people will have to think about the value-add at the application level.”</em><br /><span class="tiny">Robin Li (Li Yanhong) — TIME interview; Baidu World remarks, 2026-08-04</span></div> | Through the decade | <span class="verdict" data-v="contested">Contested</span> |
| <div><em>“The ChatGPT moment for physical AI is here, when machines begin to understand, reason and act in the real world.”</em><br /><span class="tiny">Jensen Huang — CES 2026 keynote, 2026-01-06</span></div> | From 2026 | <span class="verdict" data-v="contested">Contested</span> |
| <div><em>“I know that it’s going to work in the next six months or 12 months or, maybe worst case, 18 months. It is already superhuman.”</em><br /><span class="tiny">Mustafa Suleyman — Bloomberg Weekend interview, 2025-12-12</span></div> | By mid-2027 | <span class="verdict" data-v="contested">Contested</span> |
Figures computed from docs/predictions/ledger.json at build time. Every claim quoted on this page carries its own sources in the ledger.