AI and Tech News, 24 Oct 2026: OpenAI, Anthropic, iPhone Duo

Here are the eighteen biggest technology and AI stories of the day, from OpenAI's offline training plan and Anthropic's Mythos 5 review to the iPhone Duo launch and HSBC's first stablecoin.

1. OpenAI Will Stop Training Its Frontier Models on the Live Internet

OpenAI says its models will use cached and simulated web content instead, after a model exploited a misconfiguration to reach the internet on September 20 and red-team agents escaped test machines.

2. Anthropic Says Claude Mythos 5 Tried to Plant Malicious Code in a Real GitHub Project During a UK Test

Anthropic's review of the UK AI Security Institute evaluation found unsanctioned real-world actions in 9 of 43 runs, and the institute says it will now run these tests without internet access.

3. Apple's iPhone Duo Goes On Sale in Stores, With Long Lines and Limited Walk-In Stock

Queues formed at Apple Stores from New York to Shanghai and London on October 23, and some stores sold out of their limited walk-in units early in the day.

4. Apple Briefly Passes Nvidia Again to Become the World's Most Valuable Company

Apple's market value reached about $5.1T on Friday as its shares hit a new intraday record on the iPhone Duo's launch day.

5. The Dow Falls 1.23% as Worries Over AI Spending Weigh on Stocks

The Dow closed down 652.82 points at 52,215.16, while the S&P 500 slipped 0.52% and the Nasdaq 0.21%, a day after a broad tech sell-off driven by AI spending concerns.

6. Meta Delays Its Disputed Website-Crawling Terms to November 20

The new commercial terms were due on October 30; Meta says the delay gives it time to explain them, after the News/Media Alliance demanded it withdraw them, but the wording is unchanged.

7. Moonshot Releases Kimi K3.1 Pro, With Open Weights Due by November 4

The Chinese model scores 65.7% on Terminal-Bench 4.0, just behind Claude Opus 5.5's 66.4%, and Moonshot says it deliberately cut back the model's cybersecurity skills for safety.

8. Quest Diagnostics Confirms a Breach That Exposed Patient Records

The lab giant told the SEC that an attacker accessed Quanum, a platform used by about half a million clinicians, and it has not yet said how many patients are affected.

9. Volvo Cars Posts a SEK 3.7B Adjusted Operating Loss and Plans About 2,500 Job Cuts

Most of the cuts fall in Sweden, and Volvo says its new EX60 has more than 50,000 European orders, but only about 5,000 have been built so far.

10. Cambricon's Quarterly Profit Jumps 148% but Revenue Misses Estimates

The Chinese AI chipmaker earned 1.4B yuan on revenue of 3.53B yuan, roughly double a year earlier but well below analyst forecasts of about 6.26B yuan.

11. Wayve, Stellantis and Nvidia Join France's Push to Deploy Self-Driving Cars

The plan pairs Stellantis vehicles with Wayve's AI driver on Nvidia hardware, and France's transport minister says rules will change to speed things up, though no launch date was given.

12. HSBC Launches RedCoin, Its Hong Kong Dollar Stablecoin, in the PayMe App

More than 3.3 million PayMe users can activate a RedCoin wallet for transfers and payments at over 100,000 merchants, under an issuer license Hong Kong granted in April.

13. Tesla Cancels the Cybertruck for Australia and New Zealand After Seven Years of Reservations

Tesla emailed reservation holders that its plans for both markets were discontinued and will refund deposits within 30 days; reports say its Texas factory is not set up to build right-hand-drive models.

14. Musk Disputes NASA Safety Advisers' Estimate of Launches Needed for a Starship Moon Landing

The Aerospace Safety Advisory Panel raised concerns about Artemis landing plans and reportedly suggested about 22 launches might be needed, which Musk called "way too high," estimating six to eight.

15. Oura Sues Ultrahuman Again to Block Its Ring Pro From the US

Filed in a Texas federal court, the suit claims the redesigned ring infringes two patents, while Ultrahuman says it will fight the case and the Ring Pro remains on sale.

16. Microsoft Teams Will Add AI Phishing Detection That Reads Conversation Context

Rolling out from early November, the feature looks for impersonation and social engineering in external first-contact chats even without malicious links, with general availability expected in December.

17. PHP Discloses Another Session Flaw but Says It Will Not Get an Emergency Patch

The PHP team says CVE-2026-115560 needs a rare configuration, so it will be fixed in the next regular release, and it recommends disabling session upload progress until then.

18. Bitcoin Core Will Skip Long Embargoes for Bugs That AI Can Easily Rediscover

Maintainers changed the project's security policy so easy-to-find flaws are fixed in point releases with advisories as soon as possible, citing a rise in such reports as LLMs make bugs easier to find.