Uber gives all 30,000 employees OpenAI's workplace AI agent ChatGPT Work

Uber and OpenAI expanded their partnership. Uber's whole global workforce, about 30,000 people, will now use ChatGPT Work, an AI agent for business teams that OpenAI launched in July. Uber first tested it with 1,000 employees, and about 80% of them were not engineers. Uber did not say what the rollout costs. It says it will track how staff use the tool before deciding how to deploy AI more widely.

Microsoft says its AI diagnosis tool beats generalist doctors on hard medical cases

Microsoft says its Diagnostic Orchestrator, which runs on GPT-5, uses AI agents that act like a panel of experts. It claims the tool diagnoses conditions up to four times more accurately than doctors. TechCrunch reports that a study in Science found the system beat generalist doctors on hard cases. TechCrunch also notes that most of the gain seems to come from the AI models underneath, not from Microsoft's own system. Microsoft AI CEO Mustafa Suleyman told The Verge he expects something indistinguishable from human performance in the next couple of years.

Design app Canva releases Craft, its first in-house AI image model

Canva released Canva Craft, an image generation model built by its own research team. TechCrunch reports it was trained on human-designed templates and is available to all users, with plans to support third-party chatbots. The Verge reports it uses a two-stage design that plans a layout before generating the image. Canva says this gives more precision, consistency and control at lower cost.

OpenAI reportedly nears up to $30 billion in funding at a $1.25 trillion valuation

CNBC, citing sources, reports that OpenAI is close to a funding round of up to $30 billion, led by investors from the UAE and Coatue. The valuation before the new money would be $1.25 trillion, and the round is described as a bridge ahead of a stock market listing. The Financial Times reports that OpenAI cut its target from the $1.4 trillion it pitched earlier this month after resistance from some investors. At over $1 trillion, it would be the most highly valued private company ever.

Chinese AI lab DeepSeek raised about $15 billion, Bloomberg's sources say

Bloomberg, citing sources, reports that DeepSeek raised about 100 billion yuan, or roughly $15 billion, in its latest round. That would make it one of the largest fundraisings ever in China. Battery maker CATL and Tencent are said to be among the biggest contributors. TechCrunch notes the round comes days after the US ordered federal agencies to remove DeepSeek's technology over national security concerns.

Memory maker Micron drops a smaller memory cell design after 20 years of research

Micron says it will not bring 4F² DRAM to market, even though it says it invented the cell design and has researched it for about two decades. Its technology chief Scott DeBoer shared the decision at a company webinar on October 21. On paper, 4F² uses vertical transistors to shrink each memory cell's footprint by about a third compared with today's 6F² layout. DeBoer said real density is limited by how tightly capacitors can be packed, called the advantage 'a bit of a myth', and said the design costs more and hits scaling limits. TechPowerUp reports Micron is betting on 3D DRAM instead.

Attackers exploited two flaws in the PHP web language days before they were disclosed

Researchers from Calif.io, Wiz, OpenAI and Quarkslab disclosed two flaws in PHP, known together as php2shell. They affect web servers running PHP 5.4 or later. Specially crafted web requests can make PHP misuse memory after an error, which can crash the server, leak data or run the attacker's code. Amazon says it saw exploitation on October 17, before public disclosure, and Ars Technica reports that attackers had a four-day head start. Cloudflare says details leaked early. It added firewall protection on all its plans about 9 hours before disclosure.

Microsoft says China-linked hackers used an AI hacking tool to breach ConnectWise

Reuters reports that Microsoft says a Chinese state-aligned hacking group used ARTEX, an AI-driven penetration testing tool, to compromise IT services firm ConnectWise. Microsoft says the breach was part of a campaign targeting hundreds of organizations. TechCrunch reports that the hackers relied on AI agents to quickly break into thousands of ConnectWise's customers and spy on them.

Indian certificate issuer eMudhra quits the web certificate business after a key leak

eMudhra says its board decided to withdraw from the global business of issuing SSL/TLS certificates, which make the padlock in your browser work. This follows an incident in which, Tom's Hardware reports, private keys for eight certificate authority certificates leaked, letting an attacker pose as any website. eMudhra says it revoked the affected certificates within 14 hours, that no customer certificates were issued from them and that no data was breached. It says the business is about 1% of its revenue. Customers will get free replacement certificates from partners.

Australian judge orders Apple to allow other app stores on iPhones from March 2027

An Australian federal judge ruled that Apple must let iPhone users in Australia install apps from third-party app stores starting March 1, 2027. The ruling follows a years-long antitrust fight with Epic Games. Apple plans to appeal. Epic says Fortnite returns to the App Store in Australia on Thursday.

Everything else