Dan H

AISN #38: Supreme Court De­ci­sion Could Limit Fed­eral Abil­ity to Reg­u­late AI Plus, “Cir­cuit Break­ers” for AI sys­tems, and up­dates on China’s AI industry

AI Safety Newslet­ter #37: US Launches An­titrust In­ves­ti­ga­tions Plus, re­cent crit­i­cisms of OpenAI and An­thropic, and a sum­mary of Si­tu­a­tional Awareness

AISN #36: Vol­un­tary Com­mit­ments are In­suffi­cient Plus, a Se­nate AI Policy Roadmap, and Chap­ter 1: An Overview of Catas­trophic Risks

AISN #35: Lob­by­ing on AI Reg­u­la­tion Plus, New Models from OpenAI and Google, and Le­gal Regimes for Train­ing on Copy­righted Data

• Dan HMay 5, 2024, 3:31 PM

  7 points

  0 ∶ 0

  in reply to: Zach Stein-Perlman’s comment on: In­tro­duc­ing AI Lab Watch

  I’ve heard OpenAI employees talk about the relatively high amount of compute superalignment has (complaining superalignment has too much and they, employees outside superalignment, don’t have enough). In conversations with superalignment people, I noticed they talk about it as a real strategic asset (“make sure we’re ready to use our compute on automated AI R&D for safety”) rather than just an example of safety washing. This was something Ilya pushed for back when he was there.

• Dan HMay 4, 2024, 9:04 PM

  3 points

  0 ∶ 3

  on: In­tro­duc­ing AI Lab Watch

  OpenAI has made a hard commitment to safety by allocating 20% compute (~20% of budget) for the superalignment team. That is a huge commitment which isn’t reflected in this.

• Dan HMay 3, 2024, 5:16 AM

  4 points

  0 ∶ 1

  in reply to: Zach Stein-Perlman’s comment on: In­tro­duc­ing AI Lab Watch

  I mean Google does basic things like use Yubikeys where other places don’t even reliably do that. Unclear what a good checklist would look like, but maybe one could be created.

AISN #34: New Mili­tary AI Sys­tems Plus, AI Labs Fail to Uphold Vol­un­tary Com­mit­ments to UK AI Safety In­sti­tute, and New AI Policy Pro­pos­als in the US Senate

• Dan HMay 2, 2024, 6:45 AM

  13 points

  5 ∶ 0

  on: In­tro­duc­ing AI Lab Watch

  To my understanding, Google has better infosec than OpenAI and Anthropic. They have much more experience protecting assets.

AISN #33: Re­assess­ing AI and Biorisk Plus, Con­soli­da­tion in the Cor­po­rate AI Land­scape, and Na­tional In­vest­ments in AI

AISN #32: Mea­sur­ing and Re­duc­ing Hazardous Knowl­edge in LLMs Plus, Fore­cast­ing the Fu­ture with LLMs, and Reg­u­la­tory Markets

AISN #31: A New AI Policy Bill in Cal­ifor­nia Plus, Prece­dents for AI Gover­nance and The EU AI Office

AISN #30: In­vest­ments in Com­pute and Mili­tary AI Plus, Ja­pan and Sin­ga­pore’s Na­tional AI Safety Institutes

AISN #29: Progress on the EU AI Act Plus, the NY Times sues OpenAI for Copy­right In­fringe­ment, and Con­gres­sional Ques­tions about Re­search Stan­dards in AI Safety

AISN #28: Cen­ter for AI Safety 2023 Year in Review

AISN #27: Defen­sive Ac­cel­er­a­tionism, A Ret­ro­spec­tive On The OpenAI Board Saga, And A New AI Bill From Se­na­tors Thune And Klobuchar

AISN #26: Na­tional In­sti­tu­tions for AI Safety, Re­sults From the UK Sum­mit, and New Re­leases From OpenAI and xAI

AISN #25: White House Ex­ec­u­tive Order on AI, UK AI Safety Sum­mit, and Progress on Vol­un­tary Eval­u­a­tions of AI Risks

AISN #24: Kiss­inger Urges US-China Co­op­er­a­tion on AI, China’s New AI Law, US Ex­port Con­trols, In­ter­na­tional In­sti­tu­tions, and Open Source AI

AISN #23: New OpenAI Models, News from An­thropic, and Rep­re­sen­ta­tion Engineering