sonabinu's submissions

1.		Using AI, Mathematicians Find Hidden Glitches in Fluid Equations (quantamagazine.org)
		4 points by sonabinu 1 day ago \| past \| discuss
2.		Distinct AI Models Seem to Converge on How They Encode Reality (quantamagazine.org)
		5 points by sonabinu 2 days ago \| past \| discuss
3.		The $160k Mechanic Job That Ford Can't Fill (wsj.com)
		8 points by sonabinu 4 days ago \| past \| 7 comments
4.		MHC: Manifold-Constrained Hyper-Connections (arxiv.org)
		1 point by sonabinu 8 days ago \| past \| discuss
5.		Measuring Agents in Production (arxiv.org)
		3 points by sonabinu 9 days ago \| past \| discuss
6.		Cryptographers Show That AI Protections Will Always Have Holes (quantamagazine.org)
		2 points by sonabinu 18 days ago \| past
7.		Bloom: An open source tool for automated behavioral evaluations (anthropic.com)
		2 points by sonabinu 19 days ago \| past
8.		Oral Exams Improve Engineering Student Performance, Motivation (2023) (ucsd.edu)
		1 point by sonabinu 30 days ago \| past
9.		Ask HN: Why does scale values between -1 and 1 return nothing on Google search
		2 points by sonabinu 45 days ago \| past
10.		Algorithms for Decision Making (algorithmsbook.com)
		1 point by sonabinu 54 days ago \| past
11.		Google Colab Is Coming to VS Code (googleblog.com)
		1 point by sonabinu 58 days ago \| past
12.		Sarah Jessica Parker's Year of Judging the Booker Prize (nytimes.com)
		2 points by sonabinu 61 days ago \| past
13.		A Personal History of the Tidyverse (hadley.github.io)
		2 points by sonabinu 66 days ago \| past
14.		In a First, AI Models Analyze Language as Well as a Human Expert (quantamagazine.org)
		4 points by sonabinu 71 days ago \| past
15.		Large Language Models Get All the Hype, but Small Models Do the Real Work (wsj.com)
		6 points by sonabinu 71 days ago \| past \| 1 comment
16.		The Math Trick Hidden in Your Credit Card Number (scientificamerican.com)
		5 points by sonabinu 83 days ago \| past \| 3 comments
17.		The Art of Scaling Reinforcement Learning Compute for LLMs (arxiv.org)
		2 points by sonabinu 86 days ago \| past
18.		America's Favorite Apple Is a Farmer's Nightmare (wsj.com)
		3 points by sonabinu 88 days ago \| past
19.		Expected Attention: KV Cache Compression by Estimating Attention (arxiv.org)
		20 points by sonabinu 3 months ago \| past \| 3 comments
20.		Delta-Code: How Does RL Unlock and Transfer New Programming Algorithms in LLMs? (arxiv.org)
		1 point by sonabinu 3 months ago \| past
21.		One AI Model Creates a Physical Intuition of Its Environment (quantamagazine.org)
		3 points by sonabinu 3 months ago \| past
22.		How to live to 117? Researchers find clues in the oldest woman (washingtonpost.com)
		3 points by sonabinu 3 months ago \| past
23.		No more scribbling: Indian court tells doctors to fix their handwriting (bbc.com)
		4 points by sonabinu 3 months ago \| past \| 1 comment
24.		Understanding RL for model training, and future directions with GRAPE (arxiv.org)
		33 points by sonabinu 3 months ago \| past \| 1 comment
25.		To Understand AI, Watch How It Evolves (quantamagazine.org)
		2 points by sonabinu 3 months ago \| past
26.		The surprisingly lucrative business of making a list of 500 stocks (npr.org)
		2 points by sonabinu 3 months ago \| past \| 2 comments
27.		RL in Name Only? (arxiv.org)
		1 point by sonabinu 3 months ago \| past
28.		9/16/25, celebrate a date of mathematical beauty (npr.org)
		4 points by sonabinu 3 months ago \| past
29.		Why Language Models Hallucinate (arxiv.org)
		3 points by sonabinu 4 months ago \| past \| 1 comment
30.		Gordon Bowker, Who Came Up with the Idea for the First Starbucks, Dies at 82 (wsj.com)
		1 point by sonabinu 4 months ago \| past
		More