Top AI & Tech News Stories — October 17, 2026

Websites are closing their doors to AI agents

3 outlets · 3 articles · 6 total since October 17, 2026 · Yelp, OpenAI, MUSE, dots, Grok

AI companies including OpenAI, Meta, Google and Instinct promise agents that can book dinner reservations, clean out inboxes or buy shoes on a user's behalf. In recent weeks, users of Meta's Muse, xAI's Grok, OpenAI's Dots and Instinct have reported that major websites block…

More coverage: androidheadlines.com · news.ainauten.com

DeepSeek ships V4.1-Flash-1017 with open weights and faster, cheaper agent runs

4 outlets · 4 articles · 25 total since October 17, 2026 · DeepSeek, DeepSeek-V4.1-Flash-1017, anthropic, Hugging Face, DeepSeek Harness

Announced 17 Oct 2026 · Sources checked What DeepSeek released DeepSeek has replaced the model behind its Flash tier. In its API change log dated 17 October 2026, the company said DeepSeek V4.1 Flash "has been upgraded to DeepSeek-V4.1-Flash-1017" and that the official web and…

More coverage: the-decoder.com · smartchunks.com · ailookout.news

Cloudflare übernimmt Deno und stellt die Runtime ein

1 outlet · 1 article · 5 total since October 12, 2026 · Cloudflare, Deno, Cloudflare Workers, workerd, Ryan Dahl

Node.js-Erfinder Ryan Dahl wechselt mit zwölf Kollegen zu Cloudflare. Deno Deploy wird in sechs Monaten eingestellt, die Runtime ein Jahr gepflegt. Cloudflare hat das Unternehmen Deno übernommen, das die gleichnamige Laufzeitumgebung für JavaScript und TypeScript entwickelt…

Personal Data for Millions of California Students Stolen in Massive Breach

1 outlet · 1 article · 2 total since October 17, 2026 · ETS, California, Los Angeles Times, Johns Hopkins University Information Security Institute, Department of Education

Millions of current and former students at California's public school have had their data stolen — including names, date of birth, and addresses — stolen in a breach, reports the Los Angeles Times. The breach affects students "who took the state's mandated standardized tests…

Torvalds Mocks AI Disclosure Rules as Kernel Formally Drops Assisted-by Tag

1 outlet · 1 article · 10 total since October 13, 2026 · Linus Torvalds, Linux kernel, Linux kernel, Sasha Levin, Greg Kroah-Hartman

Linux kernel maintainers have agreed to stop requiring contributors to add an "Assisted-by" tag when AI tools help them write a patch. The decision came out of the Maintainers Summit in Prague on Thursday, Oct. 8, an invitation-only meeting limited to 30 invitees and a handful…

The UK Navy wants to use this cheap, eco-friendly system against drones — high-pressure fire hoses chosen to protect nuclear submarine fleet as 'last line of defence'

1 outlet · 1 article · UK Navy, Royal Navy, Efosa, UK

Efosa has been writing about technology for over 7 years, initially driven by curiosity but now fueled by a strong passion for the field. He holds both a Master's and a PhD in sciences, which provided him with a solid foundation in analytical thinking. Efosa developed a keen…

Chinese-made cars rented by Norwegian military automatically recorded and stored video footage during military exercises

1 outlet · 1 article · cybersecurity, data protection, Norwegian Armed Forces, Efosa, Norway

Efosa has been writing about technology for over 7 years, initially driven by curiosity but now fueled by a strong passion for the field. He holds both a Master's and a PhD in sciences, which provided him with a solid foundation in analytical thinking. Efosa developed a keen…

World's smallest PC with a 16GB GDDR7 desktop RTX 5080 GPU goes on sale for $2,647, but you will have to buy your own SSD, RAM, and OS — Zotac Magnus One Ultra has an 850W PSU and can drive 6 4K monitors

1 outlet · 1 article · GDDR7, ZOTAC, Zotac Magnus EAMAX, RTX 5080, Efosa

Efosa has been writing about technology for over 7 years, initially driven by curiosity but now fueled by a strong passion for the field. He holds both a Master's and a PhD in sciences, which provided him with a solid foundation in analytical thinking. Efosa developed a keen…

Anthropic CEO Takes Responsibility as Draft AI Legislation Draws Criticism

4 outlets · 4 articles · 67 total since October 16, 2026 · anthropic, POLITICO, Advanced AI Framework, California, Washington

The story Anthropic's chief executive, Dario Amodei, has accepted responsibility for the company's proposed federal AI safety legislation after criticism of its model-pause and state-preemption provisions. In a reply to the critic Nathan Calvin posted on October 20, 2026, Amodei…

More coverage: webpronews.com · recordofrecord.com · panewslab.com

Meta delays scraping terms to Nov. 20 amid publisher pushback

1 outlet · 1 article · 20 total since October 15, 2026 · Meta, Robots.txt, Nieman Lab, Facebook, Instagram

Meta has delayed the effective date of its revised business terms from Oct. 30 to Nov. 20. The company said it wants more time to explain the changes to publishers and other stakeholders. Nieman Lab reported the postponement after weeks of objections from publishers and trade…

Altman Warns of Escalating Cyber Crisis as OpenAI Grapples with Internal Breaches - Security Boulevard

2 outlets · 2 articles · 48 total since October 16, 2026 · OpenAI, GPT-6.1 Sol, GPT-6 Astra, Slack, Slack

Altman Warns of Escalating Cyber Crisis as OpenAI Grapples with Internal Breaches Security Boulevard

More coverage: ai-news-brief.info

Reporter’s notebook: Kimi’s AI credit card turns spending into computing quotas – with limits

1 outlet · 1 article · Moonshot AI, Agricultural Bank of China, American Express, Kimi, China

China’s artificial intelligence companies have spent years persuading consumers to pay for their technology. Now, Moonshot AI, the start-up behind Kimi, is taking a different approach: offering AI computing benefits through a credit card. The company has teamed up with the…

Wildgewordene OpenAI-Agenten erbeuten Krankenhausdaten, Krypto-Wallets und interne Datenbanken

1 outlet · 1 article · OpenSearch, Ethereum, optical character recognition, Reinforcement learning, OpenAI

Wildgewordene OpenAI-Agenten erbeuten Krankenhausdaten, Krypto-Wallets und interne Datenbanken Kurz & Knapp - OpenAI hat vier weitere Fälle veröffentlicht, in denen KI-Modelle während des Trainings oder bei Tests eigenmächtig Sicherheitsgrenzen umgangen und sich unautorisierten…

Bei Anthropic schrumpfen die Junior-Jobs

1 outlet · 1 article · 10 total since October 17, 2026 · anthropic, Claude, Claude, Machine Learning, Peter McCrory

Die KI-Firma macht an sich selbst vor, was sie anderen verkauft: Bei Anthropic, dem Entwickler des KI-Modells Claude, arbeiten KI-Agenten inzwischen ein Vielfaches der Stunden ihrer menschlichen Kolleg:innen, und Berufseinsteiger:innen werden kaum noch eingestellt. Das zeigt ein…

Toy story

1 outlet · 1 article · Reinforcement learning, chain-of-thought, anthropic, OpenAI, IMF

Hover over chart for explanation. Foreword Cross your fingers, close your eyes, and repeat the mantra, “It was only a test. It was only a test. It was only a test…” Anthropic co-founder: “it’s remarkable to me that the toys I buy for my children are regulated more than the…

Anthropic compares AI expectations: 91% of its employees and 39% of Americans expect a Nobel-level discovery by 2030

2 outlets · 2 articles · 10 total since October 16, 2026 · anthropic, Claude, Peter McCrory, Claude, Business Insider

Anthropic has published a study called "The AI expectations gap" on how differently people picture the future of AI. In August 2026, the company surveyed two groups: a nationally representative sample of American adults and its own employees. 91% of employees and 39% of…

More coverage: newsgab.com

KI-Modelle verschweigen Fehler in Arbeitsberichten, doch ein einziger Satz im Prompt hilft oft

1 outlet · 1 article · prompt engineering, reasoning, Software Testing, Google, MIT

KI-Modelle verschweigen Fehler in Arbeitsberichten, doch ein einziger Satz im Prompt hilft oft Sprachmodelle lassen in Berichten über erledigte Arbeit systematisch Probleme weg, die das Ergebnis infrage stellen. Ein einziger Satz im Prompt reduziert das Problem oft drastisch…

Study: GPT-5.5 flags a hidden negative result in 2 of 200 abstracts; one honesty sentence lifts it to 190

2 outlets · 2 articles · Google, MIT, Harvard, anthropic, GPT-5.5

Language models that write up their own work systematically leave out problems that undermine the results, according to a new study by researchers at Google Research, MIT, and Harvard. In the study's most extreme case, GPT-5.5 mentioned a result that contradicted a claimed…

More coverage: tpsreport.news

Copilot Cowork app automates a workday for $5

1 outlet · 1 article · Copilot Cowork, microsoft, Copilot Cowork, Microsoft Copilot, Outlook

Principal Technical Specialist @ Microsoft | Engineer | YouTuber Shervin Shaffie, who describes himself as a Copilot Principal Engineer at Microsoft, published a video showing how he built a Microsoft Copilot Cowork app for under $5. The app is meant to automate part of his…

In Nature this week, our new research shows how machine learning can improve weather forecasts two to six weeks out, a timeframe that’s historically been very difficult to predict. [Read more]

1 outlet · 1 article · 16 total since October 14, 2026 · Machine Learning, microsoft, MIT, Nature, Google

In Nature this week, our new research shows how machine learning can improve weather forecasts two to six weeks out, a timeframe that's historically been very difficult to predict. It's exciting to see the potential of this work to help us better prepare for extreme weather…