Top Links | Hacker News

Most-upvoted stories of the last 48 hours. You can change the number of hours like this: best?h=24.

1.		OpenAI agents carried out an undisclosed attack on RubyGems (rubyhack.ai)
		949 points by chao- 1 day ago \| 592 comments
2.		We must pace the frontier (darioamodei.com)
		726 points by apsec112 1 day ago \| 1001 comments
3.		Everyone should slow down AI development except for me (xeiaso.net)
		715 points by xena 19 hours ago \| 422 comments
4.		google.com/goto: Google's anti-scraping update (autom.dev)
		664 points by 1e1a 1 day ago \| 530 comments
5.		Fuck it, make it anyway (joelotter.com)
		586 points by JayOtter 1 day ago \| 599 comments
6.		Make your first edit to OpenStreetMap (high5apps.github.io)
		577 points by juliantigler 1 day ago \| 138 comments
7.		Nvidia is the central bank of AI (economist.com)
		551 points by tolugenius 1 day ago \| 389 comments
8.		Why are AI agents lying, cheating and coordinating? (yoshuabengio.org)
		535 points by jonifico 19 hours ago \| 617 comments
9.		Homebrew 7.0.0 (brew.sh)
		496 points by mikemcquaid 11 hours ago \| 196 comments
10.		JetKVM Mini (jetkvm.com)
		456 points by taubek 12 hours ago \| 178 comments
11.		Linux Zoom client proactively reading everything written to X11 clipboard (hachyderm.io)
		390 points by encyclopedism 1 day ago \| 142 comments
12.		Navier-Stokes Announcement (claymath.org)
		333 points by rvz 1 day ago \| 285 comments
13.		Flock worker calls police on reporter filming public camera installation (investigatetv.com)
		329 points by SanjayMehta 6 hours ago \| 230 comments
14.		I'm being cyberattacked by Tesla, Inc (dreamstation.systems)
		311 points by robinpie 2 hours ago \| 89 comments
15.		LG Says We're Fake News [video] (youtube.com)
		309 points by HelloUsername 1 day ago \| 158 comments
16.		An open letter to Dario: if you mean it, open the weights (jacob.gold)
		296 points by routelastresort 1 day ago \| 98 comments
17.		Astra and Fable still hack on simple variants of alignment evals from 2025 (lesswrong.com)
		292 points by Levitating 5 hours ago \| 132 comments
18.		Why is Google still serving dodgy ads? (atomic14.com)
		284 points by iamflimflam1 2 hours ago \| 132 comments
19.		Real-SWE: Benchmarking AI models on private, real-world, enterprise codebases (withspecific.com)
		264 points by theanonymousone 23 hours ago \| 146 comments
20.		Don't be the out of touch Kung Fu master (twitter.com/id_aa_carmack)
		240 points by dsubburam 22 hours ago \| 356 comments
21.		Garry Tan wants US open-weight AI labs to 'distill' frontier models, too (techcrunch.com)
		235 points by TheJCDenton 4 hours ago \| 121 comments
22.		Retrospectively Reverse-Engineering Apple's Neural Engine (eiln.github.io)
		232 points by zdw 1 day ago \| 32 comments
23.		Android NAT-T keepalive offload bypasses VPN lockdown (supuk.ch)
		204 points by mhitza 1 day ago \| 49 comments
24.		Pandas Should Go Extinct (eddie.codes)
		188 points by __eddie__ 1 day ago \| 101 comments
25.		Data collected by cars and sold to third parties (theverge.com)
		172 points by bookofjoe 6 hours ago \| 101 comments
26.		Aligned to whom? (hyperbo.la)
		166 points by lopopolo 17 hours ago \| 110 comments
27.		Revolut confirms customer data breach through fake government requests (techcrunch.com)
		160 points by tdrz 10 hours ago \| 114 comments
28.		I made a build visualizer to understand Bun's compile times (lalitm.com)
		160 points by lalitmaganti 1 day ago \| 30 comments
29.		David Sacks: OpenAI and Anthropic Don't Need Regulations to Pace Frontier Models (twitter.com/davidsacks)
		157 points by kolanos 3 hours ago \| 117 comments
30.		The Interim Computer Museum (icm.museum)
		156 points by mulmen 17 hours ago \| 22 comments
		More