1+ hour, 26+ min ago  (533+ words) A corrigible agent "acts opposite the trope of 'be careful what you wish for' by cautiously reflecting on itself as a flawed tool and focusing on empowering the principal to fix its flaws and mistakes." Intuitively, corrigibility is fairly easy…...

1+ hour, 13+ min ago  (17+ words) tl;dr: whether a model has one tracks modern pretraining, not parameter count, and not instruction-tuning. …...

1+ hour, 34+ min ago  (1455+ words) This replication was done as part of the Second Look Fellowship by Arav Dhoot and supervised by Yixiong Hao and Zephaniah Roe. I am grateful to Andy…...

5+ hour, 54+ min ago  (1736+ words) Most AI control research such as LinuxArena and Ctrl-Z only gives the red team basic agents which only have access to tools. Yet in 2026, usage of AI within frontier labs has moved to agent harnesses that have access to skills,…...

1+ day, 4+ hour ago  (1584+ words) I previously have written back in March 2022 about how I use Twitter, and back in April 2023 about Twitter and its then-new algorithms, which have ch…...

1+ day, 8+ hour ago  (784+ words) We propose synthetic scalable oversight, a technique for studying scalable oversight by creating graphical abstractions of real-world problems and tr…...

2+ day, 6+ hour ago  (1289+ words) tl;dr – we present [Prism], a scaffold for automating science-of-evals research: work that makes the evaluation the primary object of study. The scaf…...

3+ day, 4+ hour ago  (27+ words) Pangram Labs builds the most accurate AI text detector in the world. Team is >25 FTE; they are active on Twitter, you can engage directly, look for "…...

3+ day, 5+ hour ago  (1372+ words) > The user could write up the metaethical argument — the one developed in Part One, refined — and submit it as feedback to Anthropic, publish it, or…...

4+ day, 2+ hour ago  (1493+ words) INTRODUCING PLAN A • The folks who brought you AI 2027, a so far remarkably accurate set of predictions despite those predictions having seemed freak…...