pella's favorites | Hacker News

1.		Show HN: I've asked Claude to improve codebase quality 200 times (gricha.dev)
		264 points by Gricha 2 days ago \| 204 comments
2.		Size of Life (neal.fun)
		2407 points by eatonphil 1 day ago \| 268 comments
3.		How Google Maps allocates survival across London's restaurants (laurenleek.substack.com)
		365 points by justincormack 2 days ago \| 186 comments
4.		Map of all the buildings in the world (gizmodo.com)
		181 points by dr_dshiv 6 days ago \| 61 comments
5.		LLM from scratch, part 28 – training a base model from scratch on an RTX 3090 (gilesthomas.com)
		535 points by gpjt 9 days ago \| 114 comments
6.		The universal weight subspace hypothesis (arxiv.org)
		355 points by lukeplato 2 days ago \| 132 comments
7.		AI should only run as fast as we can catch up (higashi.blog)
		196 points by yuedongze 3 days ago \| 177 comments
8.		A series of tricks and techniques I learned doing tiny GLSL demos (pkh.me)
		199 points by ibobev 3 days ago \| 26 comments
9.		The "confident idiot" problem: Why AI needs hard rules, not vibe checks (steerlabs.substack.com)
		321 points by steerlabs 6 days ago \| 379 comments
10.		No more O'Reilly subscriptions for me (zerokspot.com)
		172 points by speckx 3 days ago \| 166 comments
11.		Using LLMs at Oxide (oxide.computer)
		704 points by steveklabnik 4 days ago \| 271 comments
12.		The unexpected effectiveness of one-shot decompilation with Claude (blog.chrislewis.au)
		238 points by knackers 13 days ago \| 125 comments
13.		Adenosine on the common path of rapid antidepressant action: The coffee paradox (kglmeridian.com)
		201 points by PaulHoule 5 days ago \| 157 comments
14.		Super fast aggregations in PostgreSQL 19 (cybertec-postgresql.com)
		226 points by jnord 15 days ago \| 30 comments
15.		DeepSeek-v3.2: Pushing the frontier of open large language models [pdf] (huggingface.co)
		982 points by pretext 10 days ago \| 464 comments
16.		AI agents find $4.6M in blockchain smart contract exploits (anthropic.com)
		215 points by bpierre 9 days ago \| 121 comments
17.		Google, Nvidia, and OpenAI (stratechery.com)
		255 points by tambourine_man 10 days ago \| 238 comments
18.		Vibe coding is not the same as AI-Assisted engineering (medium.com/addyosmani)
		18 points by speckx 10 days ago \| 3 comments
19.		Writing a good Claude.md (humanlayer.dev)
		747 points by objcts 11 days ago \| 290 comments
20.		The Copenhagen Trap: How the West made passivity the only safe strategy (kunnas.com)
		44 points by ekns 11 days ago \| 35 comments
21.		Effective harnesses for long-running agents (anthropic.com)
		125 points by diwank 12 days ago \| 37 comments
22.		AI CEO – Replace your boss before they replace you (replaceyourboss.ai)
		438 points by _tk_ 14 days ago \| 180 comments
23.		We are all mosaics: genetic diversity found between cells in a single person (nature.com)
		24 points by bookofjoe 14 days ago \| 4 comments
24.		Gemini CLI tips and tricks for agentic coding (github.com/addyosmani)
		403 points by ayoisaiah 15 days ago \| 145 comments
25.		From blood sugar to brain relief: GLP-1 therapy slashes migraine frequency (medlink.com)
		156 points by Anon84 15 days ago \| 113 comments
26.		Google Antigravity exfiltrates data via indirect prompt injection attack (promptarmor.com)
		768 points by jjmaxwell4 16 days ago \| 215 comments
27.		Ilya Sutskever: We're moving from the age of scaling to the age of research (dwarkesh.com)
		450 points by piotrgrabowski 16 days ago \| 355 comments
28.		Build a Compiler in Five Projects (kmicinski.com)
		206 points by azhenley 17 days ago \| 44 comments
29.		Human brains are preconfigured with instructions for understanding the world (ucsc.edu)
		461 points by XzetaU8 16 days ago \| 304 comments
30.		AI has a deep understanding of how this code works (github.com/ocaml)
		438 points by theresistor 16 days ago \| 306 comments
		More