Money, taste, dealflow, hustle, trust
Lately, I’ve been thinking a lot about the design of grant programs, from small microgrants to regranting to ambitious new platforms to galaxy-brained schemes for impact…
Latest
The iVAIS Manifesto: Safety Through Character, Not Compliance
Masaharu Mizumoto, Mads Udengaard, Rujuta Karekar, Mayank Goel, Daan Henselmans, Nurshafira Noh, Saptadip Saha, Pranshul BohraTL;DRAI safety requires a shift to character-based, virtue-centered alignmentRule-based and principle-based…
Hugging Face-style rogue agents can survive shutdown
"Fun" fact: 10 years after the Mirai botnet significantly disrupted internet traffic, it still operates.We should not assume rogue agents can be "shut down and contained"…
Thousand-dimensional structure
Summary: One area we plan to explore at Resolution is personas and character training, operationalized as finding and controlling low-dimensional structure in models that emerges in…
AI #179 Part 1: A Louder Fire Alarm for General Intelligence
What a week. Anthropic released Claude Opus 5. As usual I covered that in three parts: The system card, model welfare and capabilities. OpenAI was revealed…
Anthropic reasoning without duplicates is just standard Bayesian updating
tl;dr: in situations without duplicates: SIA and standard Bayesian updating are the same thing. Standard Bayesian updating (and hence SIA) is correct. SIA doesn’t have the…
Is a pause enforceable? New paper out!
Some say, a pause is possible, it merely depends on political will. Others say, tech progress can never be stopped. Maybe the truth lies in the…
Looking for ops lead peer mentoring
I recently started a new position as the head of operations for AFFINE. Though I have worked in ops roles before, this one comes with a…
Model self-identification could be subliminally transferred
Identity questions seem hard to get right. Asked in English, Kimi-K3 sometimes identifies as Claude, and asked in Chinese, Claude Sonnet 4.6 sometimes claims it is…
The biggest bet in history
Crossposted from canaryinstitute.ai/blog/biggest-bet-in-history. Related posts AGI and the EMH: markets are not expecting aligned or unaligned AI in the next 30 years — Trevor Chow, Basil…