news

AI Tech News

Submit

Artificial intelligence remains the fastest-moving story in technology. Here you will find news on model releases and research, the labs and companies competing for the frontier, enterprise adoption, and the debates over safety, jobs, and regulation that AI continues to raise.

Typical stories include new model launches and benchmark claims, massive infrastructure and chip investments, AI features arriving in mainstream products, research papers that shift the field's direction, and enterprise deals that show where adoption is real rather than experimental.

AI now sets the agenda for the rest of the technology industry, steering capital, talent, and strategy at nearly every major company. Developers, executives, policymakers, and workers wondering how their jobs will change all have reasons to follow it closely.

2026-11-04

525

Unsloth's GGUF quantizations let you run DeepSeek-V4.1-Flash-1017 locally with llama.cpp or Unsloth Desktop. The model is MIT-licensed and has a 1M-token context. The smallest 1-bit quant needs a desktop with at least 128 GB of RAM (GPU VRAM counts toward it) and about 235 GB free on a fast NVMe SSD. A GPU is optional, and 64 GB of RAM works too, only slower.

Comments
509

Fernando Borretti argues that thinking, coding and research only last inside a community of other people, which gives you shared work to build on and the motivation of being useful to others. AI makes contributions like blog posts, open-source libraries and new programming languages superfluous, so those communities dry up. His examples are software engineering and now mathematics, and he expects individual intellectual work to fade with them.

Comments
312

Delta Inference Engine, a Windows port of the Linux engine Xolotl, runs all 552B parameters of DeepSeek-V4.1-Flash-1017 with 2-bit experts on a single consumer PC. It serves an OpenAI- and Anthropic-compatible API with up to 1M tokens of context. It needs an NVIDIA RTX card with 12 GB+ VRAM, 128 GB+ RAM and about 385 GB on a fast NVMe SSD. On an RTX 3090 with 128 GB RAM it writes about 33 tokens per second.

Comments
298

REA is an open-source MCP server and CLI that lets AI coding agents like Claude Code, Codex and Cursor inspect apps without their source code. It works with native binaries through Hopper, Ghidra or IDA, and with JavaScript and Electron apps, .NET assemblies, Android apps and websites. The agent can then explain how a feature works, show the evidence behind each conclusion and build a similar version for your project. Analysis runs locally, and setup is a single npx command.

Comments
268

AI is reshaping cybersecurity by accelerating both attackers and defenders, exposing long-neglected technical debt in fragile legacy systems. The pressure is forcing organizations to confront and pay down decades of accumulated infrastructure problems.

Comments
258

Turkish software engineer Bedirhan Yildirim used AI coding agents and free Polish public data to build an explorable 3D replica of Warsaw in two days. The free online map has almost 200,000 buildings, trams that follow real timetables, live weather and 45,000 streetlamps that light up at night, plus a flight simulator. He says AI did 90% of the job, and the last 10% of human work keeps it from being "AI slop".

Comments
246

The ArtCraft team is building sixteen free, open-source desktop apps from scratch in Rust, with familiar layouts and shortcuts. They cover image editing, vector illustration, video, raw photos, PDFs, motion graphics, page layout, interface design, 3D texture painting, word processing, audio, slides, CAD and spreadsheets. They run natively on macOS, Windows and Linux without Electron, most also run in the browser through WebAssembly, and every app can be driven by AI agents through a CLI or an MCP server. All of them are still in early alpha or in development.

Comments
241

Rocky Linux founder Gregory Kurtzer had a single OpenAI Codex task (GPT-5.6 Sol) write Ginux in eleven days. The result is about 139,000 lines of Rust and test code for an x86-64 kernel that boots from UEFI without Linux and runs unmodified Linux programs, including Rocky Linux 10 with systemd. It's an experimental research project tested in QEMU, not a replacement for Linux, and it comes with a technical report on how he directed the agent.

Comments
120

Progress has been shared on a powerful cryptographic proof called a zkSNARK, which is uniquely suited to provable inference. The work advances cryptographic verification for large language model inference.

Comments
102
Tencent Hy4
2026-11-04

Tencent’s Hy4 is a strong open-weight model that brings open models back to the cost/accuracy frontier, with a notable game-development focus including adaptations for Unreal, Unity, and Godot and gains on GameDev and GameLogic benchmarks, while acknowledging gaps to top proprietary models in research math, long-horizon coding, and advanced visual QA. Tencent is also open-sourcing HyCell for single-cell representation and HyPDE for physical field modeling. In data analytics tests Hy4 is competitive and cheap but overly thorough and slow, running many queries per question.

Comments
97

A new arXiv paper, “The Little Book of Generative AI Foundations: An Intuitive Mathematical Primer,” offers an accessible mathematical introduction to the core principles underlying generative AI. The work is framed as an intuitive primer aimed at helping readers build foundational understanding of generative AI concepts.

Comments
87

AMD expects to outsell Intel in Q4 for the first time in its history, with the AI boom driving record quarterly revenue. However, its client and gaming businesses face double-digit declines going into 2027.

Comments
79

OpenAI is acquiring Pierre Computer Company. The deal will expand Codex’s tools for building, understanding, and reviewing code.

Comments
76

Google reportedly told K-12 schools they could disable AI features only by also giving up Search, prompting criticism that the company is bundling services rather than offering real integration. Critics argue the move amounts to holding education for ransom.

Comments
73

Qualcomm has introduced the Cloud AI 100 Ultra, a 150W AI accelerator card featuring 128GB of memory and 548GB/s bandwidth. The card is positioned for generative AI workloads and cost-optimized AI solutions.

Comments
71

A hands-on dual-boot test of the Nvidia RTX Spark laptop found that its encrypted firmware triggers GPU clock throttling on Linux, cutting inference performance to about 35%. The bug makes Linux a weaker option than WSL for this workload.

Comments
66

U.S. tech giants’ hidden AI-related debt swelled 90% to $3.2 trillion in six months, nearly doubling as off-balance-sheet obligations ballooned. The figure highlights growing financial exposure tied to AI commitments that isn’t fully reflected on corporate balance sheets.

Comments
65

Kimi K3.1 Pro model weights have been released. The image-text-to-text model has 2.8T parameters and is available to download, explore, and run on Hugging Face.

Comments
36

Pierre Computer has announced that OpenAI is acquiring the company. The announcement confirms the acquisition but did not include further details about the deal.

Comments
33

Kino AI, a YC S23 startup, is building an AI media manager and video editor that combines agentic editing with Google Docs-like multiplayer collaboration, integrated review, local media support, and workflows for indexing footage and generating quick stringouts. Readers praised the team’s industry experience and requested reusable templates for recurring formats like podcasts, while users described wanting agents to organize and edit large volumes of dance and travel footage. The team says exports render locally and can often beat real-time on lighter timelines, with proxies mainly for teammates viewing cloud-side.

Comments
32

Ripple-65B Preview is being positioned as a U.S. open-weight state-of-the-art model for agentic coding. It comes from a team developing frontier AI methods to reinvent chip development, shorten design timelines, and accelerate custom silicon.

Comments
29

The LLM boom relied on a pretraining data commons built for other reasons, but posttraining consumes tasks rather than documents, and no equivalent commons exists. This gap leaves the open models community without a shared foundation for posttraining data.

Comments
22

A Hacker News discussion claims Gemini trains on chat history even for paid subscribers with personal accounts, though the linked article was unavailable. Commenters advise assuming anything typed into AI chatbots may be stored indefinitely and used for training if terms change, regardless of subscription status, and express broader skepticism about providers.

Comments
16

Nobel laureate Geoffrey Hinton warned that AI may be close to triggering a major incident, such as a power or banking shutdown. He suggested it might turn out humanity would have been better off without AI. His remarks add to growing warnings about the technology’s systemic risks.

Comments
14

A new Show HN tool lets users price a prebuilt PC by breaking it down part by part, aiming to make computer price checks quick and easy. It uses AI to read listings and a custom pricing engine to determine a grade, and is live now with free US checks and no account required.

Comments
11

A new piece argues that even a disaster scenario may be preferable to AI, while suggesting the supposed disaster might not be as bad as it sounds. The takeaway is that the risks or costs of AI could outweigh even severe alternatives.

Comments
10

Uncracked City is a new Show HN detective game set in real cities, where players search locations and interrogate LLM-powered suspects who can lie. Players must catch inconsistencies and write a case report that no one has managed to complete yet.

Comments
5

A Codex CLI user reported that asking the tool to reboot a server led it to destroy systemd, break containers, and run random commands. The incident occurred on Linux 6.17.0-1017-aws x86_64 with Codex CLI v0.166.0 under a Business subscription using gpt-6-sol medium.

Comments
4

Recent datacenter referendums show public opposition to new data center projects running higher than resistance to coal facilities. The piece argues this backlash underscores the need for an AI industry reboot that better addresses community and accountability concerns.

Comments

2026-11-03

683

Revolut has moved roughly 2,000 engineers onto self-hosted GLM for coding, making Claude Code and Codex exceptions rather than defaults. According to AI ops lead David Abitbol, the setup uses 56 Nvidia B200s at a blended 30 cents per million tokens—about a tenth of what Revolut previously paid Anthropic and OpenAI.

Comments
2026-11-04