Vertical Tabs in Chrome
When you have a lot of browser tabs open it can be very hard to tell them apart, especially if they're from the same site: Since…
When you have a lot of browser tabs open it can be very hard to tell them apart, especially if they're from the same site: Since…
Alice: Are there any examples of X? Bob: Here. Alice: Okay, but are there any examples of X'? Bob: Yes, here. Alice: That has feature Y,…
(Cross-posted to EA Forum)Rhone and I, both rising sophomores, are restarting Penn's AI Safety group, which was quite active until late 2024. Please join us, or…
Decomposing emergent misalignment into behavior-per-role and role-elicitation.This project was done as part of BlueDot's Technical Safety Project.Code available: https://github.com/unrulyabstractions/bluedot-tais-project-2026. Results are preliminary.SummaryWe calculate persona vectors and…
IntroductionIf anyone builds superintelligence, everyone dies. That is, it is extremely reckless to build AI systems more intelligent than all humans, given anything like our current…
A four-day, in-person hackathon to improve the AI safety resource hub.WhenThursday 17th September – Sunday 20th September 2026WhereCEEALAR (the EA Hotel) in Blackpool, EnglandCostFree, including accommodation…
For a long time AISafety.com has listed upcoming events and training programs as they are announced. But the information it contained was limited and its presentation…
An Introductory ExampleConsider the following unwritten rule:Keep the kitchen clean. Clean after you cook.Consider the following three characters: the unreflective individual, the conscious egoist, and the…
There seems to be one type of work that truly matter: Work that enables people. Most work can be fitted into this category, such as teaching,…
Statement on AI use: AI models (mostly Fable 5 and Opus 5) were extremely helpful in 1. iterating through lots of variations on the game theory…