CharlesW's favorites | Hacker News

1.		Meta's New A.I. Superstars Are Chafing Against the Rest of the Company (nytimes.com)
		24 points by bookofjoe 8 hours ago \| 3 comments
2.		OpenAI are quietly adopting skills, now available in ChatGPT and Codex CLI (simonwillison.net)
		250 points by simonw 6 hours ago \| 155 comments
3.		Chatbot-powered toys rebuked for discussing sexual, dangerous topics with kids (arstechnica.com)
		15 points by smurda 17 hours ago \| 5 comments
4.		From text to token: How tokenization pipelines work (paradedb.com)
		115 points by philippemnoel 1 day ago \| 19 comments
5.		Something Ominous Is Happening in the AI Economy (theatlantic.com)
		42 points by jonbaer 1 day ago \| 5 comments
6.		AI Doesn't Need More Compute – It Needs Less Entropy (medium.com/yttrium39pt)
		8 points by rutheok 1 day ago \| 1 comment
7.		AI coding is sexy, but accounting is the real low-hanging automation target
		33 points by bmadduma 1 day ago \| 22 comments
8.		[flagged] Why AGI Will Not Happen (timdettmers.com)
		46 points by dpraburaj 2 days ago \| 50 comments
9.		NeurIPS 2025 Top Paper Breakdown: LLMs Are Collapsing into Artificial Hivemind (interviewready.io)
		10 points by smurda 3 days ago \| discuss
10.		Transformers know more than they can tell: Learning the Collatz sequence (arxiv.org)
		128 points by Xcelerate 9 days ago \| 45 comments
11.		Apple's slow AI pace becomes a strength as market grows weary of spending (yahoo.com)
		359 points by bgwalter 3 days ago \| 422 comments
12.		Agentic AI Foundation (block.xyz)
		127 points by thinkingkong 3 days ago \| 28 comments
13.		Alignment is capability (off-policy.com)
		106 points by drctnlly_crrct 4 days ago \| 90 comments
14.		The consumption of AI-generated content at scale (sh-reya.com)
		32 points by ivansavz 11 days ago \| 31 comments
15.		AI should only run as fast as we can catch up (higashi.blog)
		197 points by yuedongze 4 days ago \| 181 comments
16.		The "confident idiot" problem: Why AI needs hard rules, not vibe checks (steerlabs.substack.com)
		323 points by steerlabs 8 days ago \| 379 comments
17.		The AI wildfire is coming. it's going to be painful and healthy (ceodinner.substack.com)
		120 points by LordAtlas 5 days ago \| 207 comments
18.		Google Titans architecture, helping AI have long-term memory (research.google)
		583 points by Alifatisk 5 days ago \| 185 comments
19.		The Reverse Centaur's Guide to Criticizing AI (pluralistic.net)
		40 points by ArmageddonIt 6 days ago \| 4 comments
20.		Running Claude Code in a loop to mirror human development practices (anandchowdhary.com)
		51 points by Kerrick 6 days ago \| 9 comments
21.		Autism's confusing cousins (psychiatrymargins.com)
		350 points by Anon84 6 days ago \| 317 comments
22.		We gave 5 LLMs $100K to trade stocks for 8 months (aitradearena.com)
		398 points by cheeseblubber 8 days ago \| 306 comments
23.		A Technical Tour of the DeepSeek Models from V3 to v3.2 (sebastianraschka.com)
		23 points by ibobev 8 days ago \| 1 comment
24.		Show HN: Marmot – Single-binary data catalog (no Kafka, no Elasticsearch) (github.com/marmotdata)
		103 points by charlie-haley 10 days ago \| 21 comments
25.		Japanese game devs face font dilemma as license increases from $380 to $20k (gamesindustry.biz)
		337 points by zdw 10 days ago \| 202 comments
26.		Show HN: I built a dashboard to compare mortgage rates across 120 credit unions (finfam.app)
		391 points by mhashemi 9 days ago \| 130 comments
27.		Python Data Science Handbook (jakevdp.github.io)
		312 points by cl3misch 10 days ago \| 61 comments
28.		Claude 4.5 Opus’ Soul Document (lesswrong.com)
		341 points by the-needful 10 days ago \| 244 comments
29.		Apple Releases Open Weights Video Model (starflow-v.github.io)
		451 points by vessenes 11 days ago \| 169 comments
30.		It's Always the Process, Stupid (promp.td)
		272 points by DocIsInDaHouse 13 days ago \| 119 comments
		More