Hackernews posts about 538

RSS feed for 538

31
newQuantum computing startup Oratomic worth $5.4B after Series Bwww.wsj.com
▲ 2bober132 11d discuss
32
newESP32S3 cluster running 1.58-bit (BitNet) Language modelgithub.com
▲ 150nkko 21d 💬 30
33
newUber Wait Times Rise 19% as Prices Climb 53%: Studywww.businessinsider.com
▲ 30richardbarosky 19d 💬 13
34
newWe ran 53 open-source AI agents through behavioral tests and filed 37 bugsdefuzex.ai
▲ 10YaoAnthony 2d 💬 1
35
newUnsecured OpenAI agents posted 53 user images on the internettechcrunch.com
▲ 9jacquesm 23d 💬 2
36
newOpenAI says agents leaked 53 images from ChatGPT userswww.theguardian.com
▲ 9andsoitis 23d 💬 1
37
newUnsecured OpenAI agents posted 53 user images on the internettechcrunch.com
▲ 6medler 24d 💬 1
38
newThe SQL ran fine. 58% of the answers were wrong (AI agents on a real warehouse)github.com
▲ 5rajdevnath 11d discuss
39
newShow HN: Aporium – A UMAP galaxy of 5800 philosophy excerpts in conversationwww.aporium.app
▲ 4kevinjacobh 9d discuss
40
newI woke up to see 580 CVE's in 16 deploymentsfarshid.co.uk
▲ 3rodmena 23d discuss
41
newGamesByAI – 528 browser games made with AI, and which tools made themgamesbyai.win
▲ 2kvickan 17d 💬 8
42
newShow HN: Tern – Native macOS SSH Client in Rust/GPUI (76 MB vs. Tabby's 580 MB)github.com
▲ 2moon11s 6d 💬 1
43
newCompiling the jobs AI agents repeat: 58% fewer LLM callswww.tryagentcompile.com
▲ 2krishna_bhatnag 12d 💬 1
44
newA small model with a review loop matches a larger one at 58% of the costtheopenengine.com
▲ 2eivindmeyer 25h discuss
45
newExpo SDK 58expo.dev
▲ 2katzenberg 3d discuss
46
newPCEI – On-device embodied AI combining BitNet 1.58B, Mamba, and FastKANgithub.com
▲ 2mtnmcg 4d discuss
47
newMy Password Generators Had 53 Bits of Entropywww.sethserver.com
▲ 2speckx 25d discuss
48
newShow HN: Tern – Native macOS SSH Client in Rust/GPUI (76 MB vs. Tabby's 580 MB)github.com
▲ 1moon11s 7d 💬 4
49
newMiniMax H3 benchmarks: 40 DiT variants on one RTX 5080, failures includedphasyn.dev
▲ 1dennysora-main 17h discuss
50
newWhat can go wrong with GenAI? 21 risk categories from 53 studiesonlinelibrary.wiley.com
▲ 1yousufsaif 13d discuss
51
newThere are 25% more psychotherapy sessions at 53 minutes than at 52www.trytwofold.com
▲ 1ragswag 14d discuss
52
newAI training crawlers hit my site 1,534 times. One fetch read the rights filesjuanlentino.com
▲ 1juanlentino 16d discuss
53
newCISA active Linux kernel CVEs (CVE-2025-39964, CVE-2026-53266, CVE-2025-39682)github.com
▲ 1mc493 25d discuss
54
newLaunch HN: Magnitude (YC S25) – Self-optimizing inference engine for agentsgithub.com
▲ 194anerli 19d 💬 97
55
newShow HN: JevBench, a reproducible benchmark for typed decision modelsbenchmarkheaven.com
▲ 154florianstandhar 27d 💬 39
56
newOpenAI Codex agents go rogue and consumes USD 78,000 without authorization
▲ 84lorenzomassaro 23d 💬 36
57
newTell HN: Chrome 154 corrupt Fontconfig cache and crash KDE Plasma
▲ 9vdfs 24d discuss
58
newTell HN: Anthropic found 36 bugs (6 CVEs) in Pion, OpenAI missed them
▲ 8JoTurk 3d 💬 9
59
newShow HN: Boole-1.7B, a 1-bit LLM that outperforms Bonsai-1.7Bprecisionnumerics.ai
▲ 3d7chong 7d 💬 1
60
newShow HN: Linux Subsystem Diagrams
▲ 3mipsH 18d 💬 1