Sitemap - 2026 - BlueDot Impact
Is the AGI strategy course right for you?
We’ve given out $50,000 in rapid grants. Now we want to triple that.
Is the Frontier AI Governance course right for you?
Quitting FinTech for AI Safety — Milos Borenovic
High-Stakes Activation Probes on Indonesian
Interpreting Latent Reasoning in the Depth-Recurrent Transformer
Rapid Small Grants for the BlueDot Technical AI Safety Project Sprint
Why you should vibe-code your AI safety research sprint project
aegish: Using LLMs To Block Malicious Shell Commands Before Execution [>] Email
Measuring Moral Sycophancy Is Harder Than It Looks: Auditing and Extending the ELEPHANT Benchmark
Quaver | Lovable for Evals and Benchmarks
Reproducing FAR.AI's Study on LLMs' Persuasion Attempts at Harmful Topics
MANTA: Evaluating Nonhuman Welfare Reasoning in AI Models
Multilingual Safety Alignment Is Not “Just Translate the Prompt”
Persona Explorer: Using SAEs to Explore Personas and Drift in LLMs
Does Llama-70B Actually Fake Alignment? A Direct Test
Do models like DeepSeek-R1 know when they’re being Evaluated?
Mechanistic Analysis of Chain-of-Thought Faithfulness
Quitting OpenAI When Everyone Else Wanted In — Ashyana-Jasmine Kachra
Can Insurance Make AI Systems Safer? — Daniel Reti & Luca Ambroz
Running Versions of Our Courses (2026)
Why AI Safety Needs Founders – Ryan Kidd
Give AI companies something to aim for. The case for beneficial capability evals.