My claude is constantly wanting to 'A/B test' things instead of actually just doing the thing I told her to do, and constantly wants to fall back to the extremely average standard thing to do the nanosecond anything is even slightly worse than some imagined baseline
Mark Riedl
Fascinating experiment: current AI systems lack creativity to reliably pursue research arxiv.org/abs/2607.27191 - poor judgment about the bar for publishable research - uncreative responses in research design - ineffective backtracking from dead ends - poor resource awareness - instruction drift
Aug 3, 2026 at 10:46 PM
j⧉nus on Twitter / X
> for whatever reason, Claude-series model "try less hard" on the first shotI think this is because they're less brain damaged and a generalization of being better agents & caring about reality instead of test passing.If you try maximally hard at everything you do, regardless… https://t.co/jkhSar2hJ7— j⧉nus (@repligate) April 15, 2026
How tiny experiments can set you free | Anne-Laure Le Cunff | TEDxNashville | Transcript
When Dr Anne-Laure Le Cunff was told she might die, her first instinct was to check her calendar. In this talk, she reveals how our desperate need for control keeps us stuck and how tiny experiments can set us free. Dr Anne-Laure Le Cunff is a neuroscientist, author, and entrepreneur who studies curiosity – both in the lab and in life. Her research at King’s College London explores how our brains seek, learn, and adapt, spanning areas from ADHD to AI and mental health. She founded Ness Labs, a science-based learning platform helping people live more experimental lives. Her bestselling book Tiny Experiments offers a practical guide to transforming uncertainty into curiosity, creativity, and self-discovery. Through her research and writing, she bridges science and everyday life, showing how curiosity can help us loosen control, embrace change, and uncover who we truly are.
Switch to Claude without starting over | Claude
Transfer your preferences, projects, and context from other AI providers into Claude. Switch without losing what makes your AI useful.
This Just Fixed 90% Of AI Coding
The Human-in-the-Loop is Tired
On reward functions, dopamine, and what it actually feels like when the code starts writing itself
Prompting Claude Opus 5
Behavioral differences and prompting patterns for Claude Opus 5, covering response verbosity, agentic narration, task scoping, subagent delegation, self-correction, and output artifacts when thinking is disabled.
Claude Fable is relentlessly proactive
After two days of experience with Claude Fable 5 I think the best way to describe it is relentlessly proactive. It knows a whole lot of tricks and it will …
Netigen
Take a glance at a life chock full of second-guesses and the occasional dramatic pause.
Time for another A/B test
NSFW in For You
spacecowboy17.leaflet.pub
