AI ALIGNMENT FORUM
Tags
AF

Fundamental Controllability Limits

•

Applied to Why Recursive Self-Improvement Might Not Be the Existential Risk We Fear by Nassim_A 1mo ago

•

Applied to Formalize the Hashiness Model of AGI Uncontainability by Remmelt Ellen 2mo ago

•

Applied to Lenses of Control by WillPetillo 2mo ago

•

Applied to What if Alignment is Not Enough? by WillPetillo 10mo ago

Remmelt Ellen v1.56.0Dec 15th 2023 GMT (+27/-71) LW1

The research field ~~AGI~~Fundamental Controllability Limits ~~of Engineerable Control & Safety Impossibility Theorems (AGILECSIT)~~ has the purpose of verifying (both the empirical soundness of premises and validity of formal reasoning of):

•

Applied to Help me solve this problem: The basilisk isn't real, but people are by canary in the machine 1y ago

•

Applied to Projects I would like to see (possibly at AI Safety Camp) by Linda Linsefors 1y ago

•

Applied to Inference-Time Intervention: Eliciting Truthful Answers from a Language Model by likenneth 2y ago

•

Applied to The Control Problem: Unsolved or Unsolvable? by Remmelt Ellen 2y ago

•

Applied to On the possibility of impossibility of AGI Long-Term Safety by Remmelt Ellen 2y ago

•

Applied to Challenge to the notion that anything is (maybe) possible with AGI by Remmelt Ellen 2y ago

•

Applied to Presumptive Listening: sticking to familiar concepts and missing the outer reasoning paths by Remmelt Ellen 2y ago

•

Applied to How 'Human-Human' dynamics give way to 'Human-AI' and then 'AI-AI' dynamics by Remmelt Ellen 2y ago

•

Applied to List #3: Why not to assume on prior that AGI-alignment workarounds are available by Remmelt Ellen 2y ago

•

Applied to The limited upside of interpretability by Remmelt Ellen 2y ago

•

Applied to Why mechanistic interpretability does not and cannot contribute to long-term AGI safety (from messages with a friend) by Remmelt Ellen 2y ago