DanielFilan

Daniel Kah­ne­man has died

[Link] A com­mu­nity alert about Ziz

Bot­tle Caps Aren’t Optimisers

Pod­cast with Oli Habryka on LessWrong /​ Light­cone Infrastructure

Take­aways from a sur­vey on AI al­ign­ment resources

Test Cases for Im­pact Reg­u­lari­sa­tion Methods

Most ex­perts be­lieve COVID-19 was prob­a­bly not a lab leak

An­nounc­ing the Vi­talik Bu­terin Fel­low­ships in AI Ex­is­ten­tial Safety!

• DanielFilan 24 Dec 2020 6:19 UTC

  64 points

  on: Elephant seal

  Review for 2020 Review

  This seal baldly asserts that it is an elephant. The only quasi-evidence given for this claim is the demonstration of an unusually long nose—but it’s entirely possible to have a long nose and not be an elephant! For instance, the humble jellyfish has long tendrils that may act as a nose, and yet is widely believed to not be an elephant at all. As such, the seal fails to prove their case—hardly fitting for a forum that prizes rationality. I think it would be incredibly unwise to select this as one of the best posts of 2020.

• DanielFilan 24 Oct 2014 0:39 UTC

  64 points

  on: 2014 Less Wrong Cen­sus/​Survey

  Did the survey!

AXRP Epi­sode 27 - AI Con­trol with Buck Sh­legeris and Ryan Greenblatt

Su­perfore­cast­ing the Ori­gins of the Covid-19 Pandemic

On Fre­quen­tism and Bayesian Dogma

AXRP Epi­sode 9 - Finite Fac­tored Sets with Scott Garrabrant

• DanielFilan 7 Sep 2023 18:34 UTC

  59 points

  23

  in reply to: Deena Englander’s comment on: Shar­ing In­for­ma­tion About Nonlinear

  I can guarantee you from my perspective as a coach that a good number of the items mentioned here are abjectly false.

  What’s an example of something that’s false?

Se­cu­rity Mind­set and Take­off Speeds

AXRP Epi­sode 24 - Su­per­al­ign­ment with Jan Leike

An An­a­lytic Per­spec­tive on AI Alignment

An­nounc­ing AXRP, the AI X-risk Re­search Podcast

Mechanis­tic Trans­parency for Ma­chine Learning