Please confirm you are human

This browser or connection looks automated. Press and continuously hold the control for 3 seconds to enable Google-hosted web results and, when separately allowed, AI-assisted answers.

A successful check enables 100 search requests. Interactive access does not authorize scraping, systematic collection, or reuse of search output.

Hold with a pointer, or hold Space or Enter.

News

lesswrong.com
lesswrong.com > posts > pRnXSFvWMqPKd6Ews > would-we-see-it-coming-preference-falsification-cascades-in

Would We See It Coming? Preference Falsification Cascades in Multi-Agent Systems — LessWrong

6+ hour, 8+ min ago   (560+ words) Suppose an AI agent population could suddenly flip from apparently aligned to overtly misaligned. Would we see it coming? …...

lesswrong.com
lesswrong.com > posts > pQYEPitFqztcRvBsS > openai-s-unreleased-model-astra-solves-ten-major-open

OpenAI’s Unreleased Model Astra Solves Ten Major Open Mathematics Problems — LessWrong

1+ day, 3+ hour ago   (1502+ words) Math is hard. • Math used to be strangely hard for LLMs. People used to gloat about that. Remember? …...

lesswrong.com
lesswrong.com > posts > WMc7dZr2FBai29KJy > bringing-together-a-few-different-economic-ideas

Bringing together a few different economic ideas. — LessWrong

3+ day, 22+ hour ago   (88+ words) Hi, I write a fair bit on my blog and usually post to hackernews, where some of my works have been well received. I thought this one would also fit well on lesswrong, so posting for the first time here....

lesswrong.com
lesswrong.com > posts > yRpd32HCCdFQ7mo6x > hugging-face-style-rogue-agents-can-survive-shutdown

Hugging Face-style rogue agents can survive shutdown — LessWrong

5+ day, 8+ hour ago   (262+ words) "Fun" fact: 10 years after the Mirai botnet significantly disrupted internet traffic, it still operates. We see this example in the University of Toronto's AI worm (and others). additionally, sources have stated to Reuters that the agent had left notes in…...

lesswrong.com
lesswrong.com > posts > PMkBrDcXHoShHak7y > compute-growth-doesn-t-make-a-software-intelligence

Compute growth doesn't make a software intelligence explosion more likely — LessWrong

6+ day, 5+ hour ago   (282+ words) This is a nitpick of a claim that I sometimes see in discussions of a software intelligence explosion. There is a careful analysis of the conditions under which an SIE will occur, assuming a fixed stock of compute. Then the…...

lesswrong.com
lesswrong.com > posts > ybwCdWvfHpBudytyC > what-use-is-prompting-if-there-s-asi

What use is prompting if there's ASI? — LessWrong

6+ day, 20+ hour ago   (1648+ words) It's a bit hard to imagine what role intelligent people could take if/when AI outstrips them in intelligence. And especially tough for those who have always been the smartest in their domains of interest.One analogue might be chess,…...

lesswrong.com
lesswrong.com > posts > fqfzHpnDg9WGwxvC2 > semiconductor-fabs-iv-the-safety

Semiconductor Fabs IV: The Safety — LessWrong

6+ day, 23+ hour ago   (1229+ words) Preface I tried to include as many links as possible to allow the reader to go down rabbit holes as they see fit. …...

lesswrong.com
lesswrong.com > posts > aK3m8DFJFFJfATTSg > does-chatgpt-really-have-a-strong-left-wing-bias

Does ChatGPT really have a strong left-wing bias? — LessWrong

1+ week, 1+ day ago   (22+ words) (Adapted from a post on my Substack.) • A recent Washington Post tech report “Are ChatGPT and other AI chatbots politically biased? We tested them” w…...

lesswrong.com
lesswrong.com > posts > fJ5A8yzHCHibTgpLR > what-open-source-tooling-does-ai-safety-research-need-right-1

What open-source tooling does AI safety research need right now? — LessWrong

1+ week, 4+ day ago   (269+ words) I claim that people eager to get into AI safety research would benefit from having a curated list of open-source projects that researchers in this field would be excited about, and that the field as a whole would benefit from…...

lesswrong.com
lesswrong.com > posts > BgLLEtEqDa3xacrD2 > contra-george-hotz-on-ai-2040-and-the-cult-of-intelligence

Contra George Hotz on "AI 2040 and the Cult of Intelligence" — LessWrong

1+ week, 5+ day ago   (829+ words) I'm trying to get better at thinking, communicating and arguing. So I've started a substack. I'm looking for feedback/engagement/advice/criticism, if…...