Unreal Numbers(lcamtuf.substack.com)

38 points by surprisetalk 8 months ago | 7 comments

Sierpiński Triangle? In My Bitwise and?(lcamtuf.substack.com)

204 points by guiambros 1 year ago | 49 comments

Approximation Game(lcamtuf.substack.com)

14 points by surprisetalk 8 months ago | 3 comments

How has mathematics gotten so abstract?(lcamtuf.substack.com)

106 points by thadt 1 year ago | 155 comments

The Mathematician Who Sculpted the Shape of Space(quantamagazine.org)

96 points by digital55 3 years ago | 19 comments

Alexander Grothendieck Revolutionized 20th-Century Mathematics(quantamagazine.org)

124 points by anujbans 5 months ago | 26 comments

It's okay to make something nobody wants(zhangluyao.com)

411 points by levi0214 3 years ago | 171 comments

A personal blog essay by 张路遥 arguing it's fine to build products for your own excitement rather than for users, opening with a Rick Rubin quote.

Loyal workers are selectively and ironically targeted for exploitation(sciencedirect.com)

341 points by ck45 3 years ago | 219 comments

Abstract page for a Journal of Experimental Social Psychology paper finding that loyal workers are selectively targeted for exploitation by managers.

Hallucinations in code are the least dangerous form of LLM mistakes(simonwillison.net)

339 points by ulrischa 1 year ago | 264 comments

Simon Willison blog post arguing that LLM hallucinations in code are the least harmful kind of mistake because running the code exposes them immediately.

Why I'm still bearish on LLMs after Navier-Stokes(dank.systems)

426 points by jaykru 1 month ago | 552 comments

Personal blog essay by Jay Kruer arguing LLMs remain overhyped despite headline results like Navier-Stokes, citing poor generalization, reward hacking, and specification costs.

Is legal the same as legitimate: AI reimplementation and the erosion of copyleft(writings.hongminhee.org)

541 points by dahlia 7 months ago | 539 comments

Blog essay arguing that AI-assisted clean-room reimplementation, as with chardet 7.0 relicensed from LGPL to MIT, may be legal but erodes the legitimacy of copyleft.

LLMs use a surprisingly simple mechanism to retrieve some stored knowledge(news.mit.edu)

336 points by CharlesW 2 years ago | 121 comments

MIT News article on research showing LLMs often retrieve stored facts via simple linear functions, and a technique to probe what a model knows about subjects.