DL towards the unaligned Recursive Self-Optimization attractor - LessWrong

Consider this abridged history of recent ML progress: …

x DL towards the unaligned Recursive Self-Optimization attractor — LessWrong Optimization AI Safety Public Materials AI Risk AI Frontpage 32 DL towards the unaligned Recursive Self-Optimization attractor by jacob_cannell 18th Dec 2021 5 min read 22 32 Consider this abridged history of recent ML progress: A decade or two ago, computer vision was a field that employed dedicated researchers who designed specific increasingly complex feature recognizers (SIFT, SURF, HoG, etc.) These were usurped by deep CNNs with fully learned features in the 2010's [1] , which subsequently saw success in speech r

related reading

• Dario Amodei — The Urgency of Interpretabilitydarioamodei.com
• On Optimism for Interpretabilitygoodfire.ai
• Transformer Circuits Threadtransformer-circuits.pub
• An Ambitious Vision for Interpretability — AI Alignment Forumalignmentforum.org
• A Pragmatic Vision for Interpretability — AI Alignment Forumalignmentforum.org
• An Extremely Opinionated Annotated List of My Favourite Mechanistic Interpretability Papers v2 — AI Alignment Forumalignmentforum.org
• Intentionally Designing the Future of AIgoodfire.ai
• In (highly contingent!) defense of interpretability-in-the-loop ML training — AI Alignment Forumalignmentforum.org
• Neel Nanda on Mechanistic Interpretability: Progress, Limits, and Paths to Safer AI — EA Forumforum.effectivealtruism.org
• The Building Blocks of Interpretabilitydistill.pub
• Andrej Karpathy — AGI is still a decade awaydwarkesh.com
• An Extremely Opinionated Annotated List of My Favourite Mechanistic Interpretability Papers v2 — AI Alignment Forumalignmentforum.org