When is misalignment just a bug? by Yoav Hollander

·Nuno Sempere··

Cross-posted from The Foretel­lix CTO Blog. In­tro­duc­tion and epistemic sta­tus: This is the first post in a planned se­ries, “Align­ment as a ver­ifi­ca­tion prob­lem”. I co-origi­nated cov­er­age-driven ver­ifi­ca­tion (CDV), which be­came the stan­dard method­ol­ogy for chip ver­ifi­ca­tion and is heav­ily used in AV safety. Back in 2015 I wrote that ver­ify­ing “Friendly AI” would be our biggest ver­ifi­ca­tion challenge (and that per­haps CDV can help a bit). A decade of au­ton­omy ver­if...

Read full article →

Related Articles

AI could keep poor countries poor by Deena Mousa
Deena Mousa · Nuno Sempere · 6h ago
METR Time Horizon 2.0—The benchmark you’ve been waiting for by AgentMa🔸
AgentMa🔸 · Nuno Sempere · 20h ago
Same Question, Three AI Models, Different Answers: A Case Study in AI Biosecurity Safeguards by Nnaemeka Emmanuel Nnadi
Nnaemeka Emmanuel Nnadi · Nuno Sempere · 1d ago
AI Forecasting in 2026: What 11 Analyses Say by Benjamin Wilson 🔸
Benjamin Wilson 🔸 · Nuno Sempere · 1d ago
A Calibration Benchmark for LLM Beliefs Across a Taxonomic Hierarchy by DanRKAlex
DanRKAlex · Nuno Sempere · 2d ago