New paper: Towards a science of AI agent reliability Arvind Narayanan & Sayash Kapoor · AI Snake Oil ·Feb 24, 2026 · 11 min read · AI & Tech
AI won’t automatically make legal services cheaper Arvind Narayanan & Sayash Kapoor · AI Snake Oil ·Feb 12, 2026 · 50 min read · AI & Tech
Fact checking moravec's paradox Arvind Narayanan & Sayash Kapoor · AI Snake Oil ·Jan 29, 2026 · 13 min read · AI & Tech
Could AI slow science? Arvind Narayanan & Sayash Kapoor · AI Snake Oil ·Jul 16, 2025 · 33 min read · AI & Tech
Agi is not a milestone Arvind Narayanan & Sayash Kapoor · AI Snake Oil ·May 1, 2025 · 24 min read · AI & Tech
AI as normal technology Arvind Narayanan & Sayash Kapoor · AI Snake Oil ·Apr 15, 2025 · 90 min read · AI & Tech
Is AI progress slowing down? Arvind Narayanan & Sayash Kapoor · AI Snake Oil ·Dec 18, 2024 · 20 min read · AI & Tech
We looked at 78 election deepfakes. Political misinformation is not an AI problem Arvind Narayanan · AI Snake Oil ·Dec 13, 2024 · 16 min read · AI & Tech
Does the UK’s liver transplant matching algorithm systematically exclude younger patients? Arvind Narayanan · AI Snake Oil ·Nov 11, 2024 · 18 min read · AI & Tech
AI companies are pivoting from creating gods to building products. Good Arvind Narayanan · AI Snake Oil ·Aug 19, 2024 · 12 min read · AI & Tech
AI existential risk probabilities are too unreliable to inform policy Arvind Narayanan · AI Snake Oil ·Jul 26, 2024 · 31 min read · AI & Tech
AI leaderboards are no longer useful. It's time to switch to pareto curves Arvind Narayanan · AI Snake Oil ·Apr 30, 2024 · 19 min read · AI & Tech
AI safety is not a model property Arvind Narayanan · AI Snake Oil ·Mar 12, 2024 · 14 min read · AI & Tech
