Extinction risk is not the right first sentence
Money wins quiet fights but votes win louder ones. What actually moves the public to act on AI, and why it decides which regulations and safeguards…
Latest
The Banality of Takeoff
"In wickedness the haughty man and the weakling meet. But they misunderstand one another. I know you." — NietzscheBack in the day, there was some point…
The Conservation Ethic in AI 2040
Summary: Halfway through, AI 2040 argues for an extreme conservation success story. How? Why? We should seek to answer these questions now so we don't make…
Can Frontier Models Autocomplete Safety Research?
We pose the following research question: how can we measure the "research taste" of language models in experiment planning? What parts of planning taste remain intrinsic…
WSJ Article Claiming China Has Matched Anthropic Is Obvious Nonsense
The Wall Street Journal printed an outright false headline and heavily misleading story claiming this, which of course was uncritically amplified by the usual suspects. I…
Easy Whole Set Dances With a Hook
I've lately ended up calling a bunch of parties and I've been happy with calling mostly longways whole-set dances. These are ones that are shaped like…
KISS AI Safety
One of the most important principles in engineering is the KISS principle — Keep It Simple, Stupid. The best engineers are the ones that rigorously adhere…
The current bottleneck is political will, not research
Abstract:We already know enough to act. I wish we were in a world where research was the bottleneck, but the main constraint on AI safety is…
Introduction for and Reactions to Plan A
Introducing Plan A The folks who brought you AI 2027, a so far remarkably accurate set of predictions despite those predictions having seemed freaky to many…
Theories of Deep Learning
This field has been blessed with exponential empirical success in the form of architectures and algorithms that simply worked through scaling, while the theory lagged behind[1].…