01
OpenAI ↗
The King's Trust and OpenAI launched a partnership on November 16, backed by a £1 million donation from OpenAI. The program is meant to help up to 25,000 young people in the UK build practical AI skills for work over 18 months, with priority for those not in education, employment or training. It sits inside a wider set of nationally recognized AI qualifications the charity plans to launch.
02
Google DeepMind ↗
Google DeepMind said AlphaProof Code agents formally verified key security properties of the code that implements the Signal protocol. The agents generated more than six million lines of Lean proofs, with human experts setting the specifications and reviewing about one percent of the proof that the system flagged. DeepMind said the work took four weeks, and that reintroducing two previously fixed bugs produced machine-checked proofs that those versions were insecure. The verification artifacts are being open-sourced.
03
The Verge ↗
Uber riders in Las Vegas can now be matched with Amazon-owned Zoox robotaxis that have no steering wheel, at no extra charge on several Uber ride types. The cars cover tourist areas including the Strip, downtown and the Arts District, and Zoox will keep its own app. Uber and Zoox also signed a multiyear deal with MGM Resorts, and they plan a Los Angeles Uber integration in 2027 after Zoox starts its own service there.
04
TechCrunch ↗
Dreamteam, founded by former Freshworks executives, raised $53 million across seed and Series A rounds led by Together Fund and Accel, with Premji Invest also participating. The San Francisco company says its CRM captures context from email and customer conversations and uses four agents for research, follow-ups, deal risks and workflow setup. It says more than 100 customers in 13 countries have signed up in three months of active selling, and it did not disclose a valuation or current revenue.
05
BBC ↗
Han Won-seok, who scored 397.5 out of 400 on the first round of South Korea's certified tax accountancy exam, confessed to photographing questions with a microcamera, sending them to a generative AI system and checking answers on a phone. He said he cheated on both rounds and has given a written confession to police. The government said people who used AI to cheat on the exam will lose the qualification and be banned from retaking it for five years.
06
The Verge ↗
Apple opened enrollment for the Apple Health AI Usage Study, run with the clinical AI platform OpenEvidence. Participants use a new iPhone app to ask health questions and rate answers for accuracy, detail and relevance. Apple is seeking up to 200,000 people aged 21 or older in the US, on an iPhone 15 Pro running iOS 26 or later, and says collected data will not be used to train AI models.
07
ZDNet ↗
Microsoft's FORGE Lab said its MDASH scanner produced 59 upstream Linux kernel bug reports, including one flaw at least 21 years old, while the kernel CVE team averaged 27 CVEs a day through November 2. Memory-management reviewer Lorenzo Stoakes said on November 10 that he was stepping back from general review because AI-generated patch series had become about half of what he sees. Linus Torvalds wrote that, on the Berkeley Packet Filter side of Linux 7.4 rc2, fixes may now be arriving faster than new reports, while other areas have not.
08
Alibaba Cloud ↗
Alibaba Cloud released Qwen4-Max through Model Studio on November 16, with a one-million-token context window and text, image and video input. Its Singapore list price is $1.50 per million input tokens and $4.50 per million output tokens, 25 percent below Qwen3.8-Max. Thinking is on by default, and thinking tokens share the output budget with the final answer. Fine-tuning and custom deployment are not offered in the documented regions.
09
Bloomberg ↗
Baidu reported its sixth straight quarterly revenue drop. Revenue fell 1 percent year over year to about $4.6 billion, and net income was about $210 million. Bloomberg said the company is trying to commercialize AI to offset a weaker search business, and that AI businesses now account for more than half of revenue.
10
arXiv ↗
A new paper introduces a benchmark of 70 workplace requests written by professionals across 13 domains. Each request comes with an expert rubric of binary checks and the workbook or CSV file it refers to. Together the files hold 2.18 million populated cells, and the rubrics contain 791 criteria. Agents work in a container with spreadsheet tools, and a separate agent that can open the file applies the rubric.