Deception Channeling: Training Models to Always Verbalize Alignment Faking
Published on February 17, 2026 10:28 PM GMTDeception Channeling: Training Models to Always Verbalize Alignment…
Category: Opinion
Rephrasing Reduces Eval Awareness…
Published on February 17, 2026 10:23 PM GMT...or at least SOME forms of it.Note: This…
You’re an AI Expert – Not an Influencer
Published on February 17, 2026 9:03 PM GMTYour hot takes are killing your credibility.Prior to…
“We are confused about agency”
Published on February 17, 2026 7:51 PM GMTEpistemic status: Speculation on agent foundations research culture…
Maybe benchmarks should be broken?
Published on February 17, 2026 7:49 PM GMTI asked GPT-5 mini to implement me a…
The brain is a machine that runs an algorithm
Published on February 17, 2026 7:36 PM GMTSome people say “the brain is a computer”.…
TV Detector Vans
Published on February 17, 2026 6:29 PM GMTIn the UK, you need a license to…
Notes on International Klein Blue
Published on February 17, 2026 5:51 PM GMTOne day in 2022, I visited the San…
Correlation Does in Fact Imply Causation
Published on February 16, 2026 9:17 PM GMT(Cross-posted from my personal website)Epistemic status: I've become…
Memory Decoding Journal Club: The Songbird as a Model for the Generation and Learning of Complex Sequential Behaviors
Published on February 16, 2026 8:46 PM GMTJoin Us for the Memory Decoding Journal Club!A…