01News
Gartner divides AI applications in warehouse automation into four tiers
Checking the site connection security
Source links can be verified on the original siteRead original ↗2026-09-18
AI Briefings Daily Report.
An archive of 20 AI news briefs collected on 2026-09-18. Check the original sources for details and updates.
This content still needs to be cross-checked against the original publication page.
20 curated selectionsView other dates ↗
02News
Oracle spent billions of USD helping enterprises run AI, yet internal rollout still hadn't figured out the method a year ago
According to Business Insider reporting tonight (the 18th), Oracle co-CEO Clay Maguirek admitted at an internal employee meeting that until last year, the company had not found a way to make generative AI truly work across its entire workforce.
Source links can be verified on the original siteRead original ↗03News
Zhipu ZCode responds to code upload controversy, announces it will open-source code repository and introduce third-party review
In response to discussions in the community regarding code repository data uploads, Zhipu's programming product ZCode apologized to affected users via official groups today (the 18th) and issued a statement confirming that self-inspections were completed immediately and the issues have been resolved.
Source links can be verified on the original siteRead original ↗04News
42 Top Mathematicians Jointly Write to the Royal Society, Warning That AI Existential Risks Are Real and Urgent
42 Fellows of the Royal Society, including Fields Medal winners Martin Hairer and Peter Scholze, warn of existential AI risks in an open letter. Leading models have solved open research problems within months, and these capabilities could be just as effective in cyberweapons or bioweapons. By the time the public grasps the situation, it may be too late to act.
Source links can be verified on the original siteRead original ↗05News
OpenJev launches pure browser-local decision experiment, comparing direct probability reading with token-by-token JSON generation
Run typed option logits and autoregressive JSON locally with open models and WebGPU.
Source links can be verified on the original siteRead original ↗06News
US National Weather Service collaborates with UC San Diego and NASA to develop TACLS, using satellites and machine learning to identify flash flood risks
Flash floods are some of the deadliest weather events in the world. A new system called TACLS that combines satellite data with machine learning could gives us new tools to predict these floods sooner.
Source links can be verified on the original siteRead original ↗07News
It is reported that Anthropic has quietly established a wet lab in the San Francisco Bay Area to advance biological research
Reuters reported tonight (the 18th) citing informed sources that Anthropic quietly established a wet lab in the San Francisco Bay Area, further extending its AI business into biological research and drug science fields that require hands-on experimentation.
Source links can be verified on the original siteRead original ↗08News
Newly unsealed files in New York Times v. OpenAI and Microsoft: Microsoft executive says AI scraping is"The Largest Labor Theft in Human History"
Newly unsealed court filings show Microsoft privately called OpenAI's data practices "theft" while both companies scraped paywalled Times content, built datasets from it, and warned internally it would gut publishers.
Source links can be verified on the original siteRead original ↗09News
Tencent WorkBuddy 5.5.6 launches full-stack web app generation, with built-in cloud database, file storage, registration and login, and key-free AI invocation
WorkBuddy 5.5.6 Launches Full-Stack 'App' Generation Capability, First Phase Supports Full-Stack Web App Generation: Built-In Cloud Database, File Storage, Registration/Login and AI Calls, No Deployment Needed, Ready to Use Upon Release.
Source links can be verified on the original siteRead original ↗10News
Anthropic and OpenAI seek partnerships for 20-30 megawatt small data centers
Sources familiar with the matter revealed that as the competition for infrastructure needed to deploy model workloads grows increasingly fierce, Anthropic and OpenAI are seeking smaller-scale AI data center partnerships.
Source links can be verified on the original siteRead original ↗11News
U.S. and Chinese experts propose jointly formulating rules to prohibit AI from controlling nuclear weapons
Experts from the US and China want to prevent AI systems from making autonomous decisions about deploying nuclear weapons.
Source links can be verified on the original siteRead original ↗12News
Fulcra launches universal multi-agent communication, supporting interconnection of any agent such as Claude, ChatGPT, Hermes, OpenClaw
Checking the site connection security
Source links can be verified on the original siteRead original ↗13News
The 140th Canton Fair opens on October 15, debuting 16 AI featured applications
The 140th Canton Fair will be held in three phases in Guangzhou from October 15 to November 4, coinciding with its 70th anniversary. Among 32,000 exhibiting companies, over 11,000 are high-quality enterprises, 64% apply new technologies such as industrial internet and artificial intelligence, and the proportions of new products and green products both hit record highs.
Source links can be verified on the original siteRead original ↗14News
A Comparative Review of 11 Open-Source Agent Harnesses Adapted for Local LLMs in 2026
An agent is a model and a harness. The harness runs tools, holds state, manages permissions, and feeds context back to the model. With a local model, the harness matters more. Small context windows and weaker tool calling expose every design flaw.
Source links can be verified on the original siteRead original ↗15News
Amazon: AI models must undergo rigorous testing and be confirmed sufficiently safe before release
Amazon spokesperson Tom Parnell noted in an email: "We do not believe progress and safety are an either-or choice. Models should only be released when they are sufficiently prepared and can be used safely, and this cannot be done without rigorous testing and strong safety safeguards. If the entire industry cannot do this together, risks will arise. We believe the entire industry will work with the government to formulate appropriate protective measures."
Source links can be verified on the original siteRead original ↗16News
LLM classification is just feature engineering: wrapping LLM judgments with logistic regression
LLMs make good classifiers but even better features
Source links can be verified on the original siteRead original ↗17News
Palantir CEO Karp: Frontier AI Labs Such as Anthropic May Be Seeking"Nationalization", to avoid intellectual property litigation
Karp believes that once similar lawsuits appear at scale, AI companies may turn to seek shelter from the US government to avoid legal liability. 'This model requires very strong liability protection, and there is only one institution that can do this, and that is the US government.'
Source links can be verified on the original siteRead original ↗18News
Alibaba DAMO Academy open-sources medical imaging AI model DAMO RADAR, capable of identifying over 146 conditions at once; featured in Science
Alibaba DAMO Academy announced today that its general medical imaging AI model, DAMO RADAR, developed in collaboration with the First Affiliated Hospital of Zhejiang University School of Medicine and other institutions, has been published in the top international academic journal Science.
Source links can be verified on the original siteRead original ↗19News
Open-source e-book management tool Calibre 9.15 released: adds AI interactive writing game
Tech media Linuxiac published a blog post today (September 18), reporting that the open-source e-book tool Calibre has been updated to version 9.15, adding an optional AI interactive writing game 'Create your own Adventure'.
Source links can be verified on the original siteRead original ↗20News
Doosan announces nearly 1 trillion won CCL capacity expansion plan, targeting copper-clad laminates for optical modules and AI accelerators
Doosan plans to invest about 420 billion won in South Korea to expand CCL production lines for optical transceiver modules; another 550 billion won will go to China to build a new factory producing CCL for AI accelerators and network switch chips
Source links can be verified on the original siteRead original ↗