All AI news

AIA study led by Google and BIDMC tested Google's diagnostic AI chatbot AMIE with 98 outpatients before emergency visits, with a supervising doctor monitoring every exchange. No conversation needed interruption under the predefined safety criteria, and clinicians said AI summaries helped them prepare for 75% of visits. AMIE's differential diagnoses matched final diagnoses 90% of the time, but the authors say larger trials are needed.

AIJetBrains released Mellum2.1, a 12B mixture-of-experts coding model with 2.5B active parameters under Apache 2.0, emphasizing agentic programming. Under high load, its inference throughput in tokens is nearly twice that of Qwen3.5-9B in JetBrains' comparison, and multi-token prediction (MTP) speeds single-request responses by about 1.6x. The model is available on Hugging Face for local or private-infrastructure deployment, with GGUF and vLLM MTP support announced for later.

AIOdyssey announced the Odyssey-3 series of foundation world models, with Odyssey-3 Pro scoring 66.1 on the Physics-IQ Verified video-to-video benchmark, the highest recorded on that leaderboard. The series includes a standard version balancing physical accuracy and generation cost, and a Pro version with stronger physics prediction. The preview supports first-person and third-person navigation and lets users move the camera, take actions, or trigger events while the model predicts environmental changes in real time.

AIRichard Yu said more Huawei phones will adopt LogicFolding chips built under the Tau Scaling Law, though no models or timetable were given. He said the Kirin 9050 Pro's performance is 31% higher than its predecessor, and the source outlines a roadmap reaching 5.0 GHz by 2031.

AIHuawei-backed openJiuwen has open-sourced AgentOS for Enterprise under Apache 2.0 on GitHub and AtomGit, targeting multi-agent coordination, memory-based self-evolution, multi-tenant isolation and fault recovery. Huawei Connect 2026 also introduced an all-in-one appliance built on it, which the launch information says enables an end-to-end private deployment in hours.

AIByteDance's AI development brand TRAE has merged its TraeWork and TraeCode clients into a single product with Agent and IDE modes. TraeWork will stop service on October 23, and users who have not upgraded by then will lose access, while chat history, account data and commercial credits migrate automatically.

AIKingKong Technology has open-sourced the software stack for Jumper, a six-legged crab-style robot it designed, including its MuJoCo model, simulation scenes, reinforcement learning training and deployment tooling. Jumper has 22 degrees of freedom, measures about 400 by 400 by 200 mm, weighs about 1.8 kg and lists a maximum jump height of 400 mm or more. The mechanical CAD files, bill of materials, PCB designs and electrical schematics are not public, and RKNN inference on the real board has not yet been validated.

AIByteDance's Doubao Work has added an infinite creation canvas that places source materials, design options and finished output on one page, wired to the new Seedream 5.0 Flash image model. The update also adds Doubao 2.1 Lite, a lighter model aimed at everyday office tasks such as documents, spreadsheets and slide decks, with faster responses and lower credit consumption. The announcement included no benchmark results for either model.

AIAfter visiting Google, the author says the company's employee care is extremely generous, from lunch to concern for body hair, with costs unconsidered. The post claims the Chinese office even serves whole soft-shell turtles at lunch, and says Google is a better workplace than companies that subject staff to PUA.

AIShares of Chinese optical chipmakers fell for a second straight session on Friday amid fears of potential US trade curbs on next-generation optical transceivers for data centres. China's CSI 300 Index slid 1.3 per cent by midday to its lowest level since August last year, with upstream laser chip suppliers falling more steeply than downstream transceiver makers.

AITom Watson, now a senior vice-president at Palantir, warned UK ministers against letting "mob rule" dictate public procurement, saying they could get "in a lot of trouble" if they do. The former Labour deputy leader's comments come as Prime Minister Andy Burnham faces pressure to drop Palantir from government contracts, including a £330m NHS software deal.

AIOpenAI defended its decision to fire three safety researchers, Jasmine Wang, Tomek Korbak and Mikita Balesni, saying they committed a "significant breach of trust." The company said the dismissals were not about the researchers raising safety concerns, though it agreed with the letter they sent to board members and safety committees about preserving the monitorability of frontier models.

AIHarrison Chase says trajectory labeling should be split into several questions rather than one pass/fail judgment. LangChain's Jev judge in LangSmith Evals answers difficulty and correctness as separate typed outputs in a single pass on every trace.

AIJerry Liu argues that most tasks can now be solved by defining an eval and hillclimbing on it, rather than hand-coding a deterministic or agentic workflow. He says data provider companies are building evals across economic activity so frontier models can handle more work, leaving developers to define goals and success measures. He expects agent interfaces to compress most tasks into goals and eval instructions, while the most complex processes will still need explicit workflow builders.

AIKolig is now live on Product Hunt, presented as an office for AI teammates that know a company and work together. The post's creator asks for support and honest feedback after a lot of design and rebuilding work.

AISupaYatta is a Mac app that alerts Claude Code and Codex users when an agent turn finishes or needs their input, displayed in the Mac notch. It also previews replies and questions and offers one-click return to the right session, with usage bars for Claude and Codex. The launch price is $5 for the first 50 copies, then $7.

AIThe sixth installment of the seven-layer "Carbon-Silicon Dao Code: Cross-Domain Migration Governance Code" proposes a civilization-scale risk audit layer for detecting slow, irreversible disturbances from AI cross-domain migration.

AIThe seventh and final layer of the "Carbon-Silicon Dao Code" cross-domain migration governance framework sets a self-audit baseline, opens falsification terms, and defines the framework's applicability boundary. The article says it validates each layer's input-output consistency backward from layer seven to layer one, and it allows anyone who constructs a reproducible, traceable counterexample targeting NT1–NT4 to submit it for baseline review. It also archives the framework's documents and versions with hashes across Toutiao, Douyin, and GitHub.

AIThe vLLM Semantic Router team has released Decision 2.0, which answers multiple questions about one input in a single forward pass and outputs per-option probabilities. The post presents this as useful for routing and classification. A quoted post from Xunzhuo Liu says Decision 2.0 includes six open decision models ranging from 0.6B to 27B parameters, each topping same-size open models on the Jev Decision Index 0.3.

AIOpenAI says it parted ways with researchers Jasmine, Mikita, and Tomek after an internal investigation found they violated policies on handling sensitive information. The company says the decisions were not about raising safety concerns, which it says it encourages, and that it has not terminated any employee for raising concerns. OpenAI also says it is finalizing contracts with third-party safety assessors and will announce details in the coming weeks.

AIFei-Fei Li notes that mathematician friends find beauty in numbers, shapes, and natural phenomena. She asks who will define beauty as machines increasingly shape the world.

AIMarkTechPost publishes a hands-on tutorial implementing RRSI (Regularized Recursive Self-Improvement), a method that lets an LLM agent revise its own harness around a frozen model. The full loop drafts edits with Claude Opus on Vertex AI and scores them in Docker benchmarks, but the edit-selection rules are plain Python that the tutorial runs in a simulated environment with a calibrated noise band.

AICurrent AI can quickly implement or replicate features, but clearly defining requirements and describing design is still missing, and the author expects this gap to persist. As requirements grow more abstract, humans may only specify goals and check results while agents handle implementation, leaving the software's logic layer as model-generated tokens.

AIA LoRA for MiniMax H3 removes a person from video by tracking them with SAM 3.1 and generating the replacement background in overlapping windows. Users supply the original video and a clean version of its first frame.

AIDoubao Work has added a Canvas feature for complex creative tasks, placing materials, design plans and outputs on one infinite canvas where users can keep editing text, colors and layout after images are generated. The update also integrates the lightweight Doubao 2.1 Lite model, aimed at everyday Q&A, document writing, spreadsheets and PPT creation, with optimized response speed and usage consumption.

AITRAE announced on October 9 that it has merged TraeWork and TraeCode into a single platform offering Agent mode and IDE mode with seamless switching between them. The upgraded product covers desktop, web, and mobile, letting users start tasks on a computer, check progress on mobile, and continue development back on desktop.

AIGeekPark reports that Krinwave, a Shenzhen company, has completed a new 400 million yuan round with XVC and Sequoia China as new investors. The company says its low-frequency ultrasound captures brain structures through the skull, which conventional high-frequency devices cannot, and it aims to supply hardware and its kOS software platform to brain-computer interface and NeuroAI groups.

AIToday AI, built by Teambition founder Qi Junyuan, launched its China version on September 24 and connects to Feishu, DingTalk, Tencent Docs and email. The author found it proactively sends morning and evening briefings and handles a single chat window across tasks, but struggled with misjudging task weight and gave confident yet wrong mod-installation instructions that cost an hour of testing.

AIGordon Brown argues that raising UK innovation intensity to Swedish or Japanese levels could leave every household £5,000 better off and add £150bn a year to national income. He says the UK leads in universities and research papers but struggles to scale startups, with three-quarters of venture capital coming from overseas.

AISimon Willison released ttok 1.0, changing its default tokenizer from GPT-4 to a newer GPT-5/GPT-6 tokenizer after finding the old default was outdated. OpenAI has not confirmed that GPT-6 shares the GPT-5 tokenizer, but William Liu's experiment found all seven GPT models tested reported 44,794 tokens on a 31-fixture corpus.

AIJeremy Howard posted a sardonic remark that he was not allowed to say he is quite unhappy with much of what OpenAI does. The post quotes Tomek Korbak's similar statement and his relief at being permitted to say it.