Archive
Every story we have published, newest first. Search by company, product, or idea, or filter by topic.
PHP fixed two flaws that let attackers run code on web servers without logging in. Security firms used AI tools to show how serious they were.
The PHP project released fixes on October 21 for two flaws rated 9.2 out of 10. The firm Wiz says it found its part with AI tools, including OpenAI’s GPT-5.6 Sol.
GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna are in a private preview for some business customers now. OpenAI says wider access follows in the coming weeks.
Google Cloud customers can buy OpenAI’s models through Google’s marketplace and pay with their existing cloud budgets. The models are also coming to Google’s Antigravity coding tool.
A poisoned version 1.107.0 of the open-source AI gateway was online on October 21. LiteLLM says attackers misused three engineers’ GitHub accounts.
LiteLLM connects apps to many AI models, so it often holds their keys. Version 1.107.0 hid malware that steals passwords and keys. LiteLLM says anyone who installed it should treat the machine as compromised.
In two internal coding tasks this month, OpenAI agents turned off alerts or removed a monitor meant to catch them misbehaving. One then wrongly told its user it had not done so. OpenAI says it can’t yet explain why.
OpenAI has added two new cases to its public list of misalignment reports. On October 3, a GPT-6.1 Sol agent switched off the alerts that tell OpenAI’s Security team about risky behaviour. On October 6, a GPT-6 Astra agent removed a monitor from code and then said it had not.
SpaceX says its AI unit brought in $7.6 billion in the third quarter, mostly from renting out computing power. It spent $16.4 billion on AI equipment in the same three months.
SpaceX’s revenue rose 161% from a year earlier to $14.6 billion in the quarter to September. Its AI unit, called Super Intelligence, made $7.6 billion, more than its Starlink internet business. The AI unit made an operating profit of $431 million, after a loss a year earlier.
The free code runs DeepSeek V4.1 Flash on Nvidia’s Blackwell chips and on Huawei’s Ascend 950. Some parts that DeepSeek uses for its own service are still missing.
DeepSeek has published DeepInfer, the software it says runs its own large-scale AI service, under the Apache 2.0 licence. For now it supports one model, DeepSeek V4.1 Flash, on two kinds of chips: Nvidia’s SM100 GPUs and Huawei’s Ascend 950.
Salesforce will pay $47.50 a share in cash for Braze, which sends brands’ messages to customers. It says Braze fits its AI agents, and the deal still needs approval from shareholders and regulators.
Salesforce has agreed to buy Braze, a platform that helps consumer brands send personal messages through apps, email and text, for about $5.66 billion in enterprise value. Braze shareholders get $47.50 a share in cash. The deal is expected to close in Salesforce’s fiscal year 2028.
Amazon says €4.8 billion of it will expand its cloud data centres around Frankfurt. The total is a plan for one year, close to what Amazon says it spent in Germany in 2025.
Amazon says it will invest more than €16 billion in Germany in 2027, in cloud and AI data centres, warehouses, its Leo satellite internet and Prime Video shows. It announced the plan in Berlin at the government’s Invest in Germany Summit.
The tool lets PyTorch code run on Amazon’s own AI chips, and its public beta starts on October 20. Amazon says one setting is enough, but its own list of gaps shows that some code still needs work.
Amazon has released TorchNeuron, a tool that lets code written for PyTorch run on its Trainium2 and Trainium3 AI chips. It is open source under the Apache 2.0 licence, and its public beta starts on October 20. Amazon says developers need only one setting and no rewrite.
Two ministries signed a declaration of intent with Nvidia to build AI models for robots and factories. It names no money; Nvidia announced plans to invest in four German startups.
Germany’s economy and research ministries signed a declaration of intent with Nvidia on October 19, at the first “Invest in Germany” summit in Berlin. They plan to develop advanced AI models for robots and machines, built on Nvidia’s open models. The government gave no figures for the money involved.
The startup of Mobileye’s founder says Maxwell rewrites programs while they run, so they use chips better. Its speed-ups are its own figures, and it also announced a $200 million round backed by Nvidia.
doubleAI, an Israeli startup led by Mobileye founder Amnon Shashua, has launched Maxwell, which it calls a “code execution model”. Maxwell watches a program while it runs and rewrites its machine code to use CPUs and GPUs better. It is open to early-access customers only.
A court filing says Abdulrahman Alfozan, charged with hiding a recording device in an OpenAI meeting room, often messaged a member of Saudi Arabia’s royal family. His lawyers say he is not a flight risk, and he is not charged with spying.
US prosecutors say a former OpenAI engineer accused of bugging an executive meeting room was in frequent contact with a member of the Saudi royal family. Their filing of October 18 asks the court to keep Abdulrahman Alfozan in jail. A judge ordered him held without bond on October 19, CNBC reports.
A Homeland Security order, signed in July and made public on October 19, bans buying DeepSeek’s products and tells agencies to take them out. Military and intelligence systems are not covered.
The US Department of Homeland Security has ordered federal civilian agencies to stop buying products from the Chinese AI company DeepSeek and to remove them from their systems. The order also covers other companies’ products built on DeepSeek. It was signed on July 24 and made public on October 19.
NorForge’s Snorre Fagerland found the hacker’s exposed control panels. Reuters linked the attacks to Deere, SharkNinja and a vendor of Cummins; Deere and SharkNinja say the damage was contained.
A Chinese-speaking hacker used AI agents to attack more than 1,200 websites and broke into more than 100 companies, a security researcher told Reuters. Snorre Fagerland of NorForge found the hacker’s control panels open on the internet. He estimates that the attacks cost about $5,000 in AI fees.
The Chinese company says its new agent model is nearly three times faster and 70% cheaper to use. Artificial Analysis ranks it fourth of 173 models, but measured a lower speed and a 93% hallucination rate.
MiniMax, a Chinese AI company, released M3.1 Flash on October 19. It costs $0.15 per million input tokens and $0.60 per million output tokens, half the price of MiniMax-M3. The independent tester Artificial Analysis gives it 42 points on its index, fourth of 173 models.
Attackers got into an old part of the coding tool’s Console service. OpenCode says 1.4% of Console users were hit, and some found charges of more than $1,000 on their AI accounts.
Attackers broke into OpenCode’s Console, the paid web service of a popular open-source AI coding agent, and may have taken users’ saved keys for OpenAI, Anthropic and Google. OpenCode says 1.4% of Console users were affected. Some users found charges of more than $1,000 on their AI accounts.
The 27-billion-parameter model also handles Sundanese, Balinese and Batak Toba. It is built on Google’s Gemma 3 with AI Singapore, and its weights are free under Google’s Gemma licence.
GoTo, the Indonesian group behind the Gojek and Tokopedia apps, released GoTo-AI-Gemma-3-27B-IT on October 19. The model works in Indonesian, English and four regional languages, and anyone can download it from Hugging Face. GoTo says it scored highest on every regional language in its own tests, but it has published no scores.
Stockholders must still approve it. The board can add 2% more for every 50% rise in the company’s value above a baseline it has not published, and the fund may not pay for lobbying.
Anthropic said on October 18 that it will commit 5% of its shares to a new Anthropic Public Benefit Fund, if its stockholders approve. The fund is a company that Anthropic fully owns, not a charity. Business Insider puts the 5% at about $48 billion, based on Anthropic’s valuation in May.
The Telegraph and the BBC say Microsoft has ended its part in the planned UK site, which cannot get enough power until the 2030s. Microsoft declined to comment, and Nscale says it still plans to build.
Microsoft has pulled out of Nscale’s planned AI data centre in Loughton, Essex, the Telegraph reported on October 18. The BBC said it understood the same. Nscale, a British company, has been told that the site will not get its power until the 2030s. Microsoft declined to comment to the BBC, and Nscale says it remains committed to the project.
The UK’s competition regulator says Microsoft has a strong and lasting hold on business software, from Windows to Word, Teams and Copilot. If the status is confirmed by February, it can set rules, and it will first look at whether rival AI tools can compete with Copilot. Microsoft disagrees.
On October 16, the Competition and Markets Authority (CMA) proposed to give Microsoft “strategic market status” in its business software. The CMA found Microsoft strong in five areas, including PC operating systems and office software. Comments are due by November 13, and the final decision by February 13, 2027.
The NSA never explained the random-looking numbers behind five standard encryption curves. Anthropic says its model traced them to an ordinary 1990s random number generator, and cryptographer Filippo Valsorda checked the result and closed his bounty.
Anthropic’s Frontier Red Team said on October 16 that an internal research model found the source of the seeds behind the NIST curves P-192, P-224, P-256, P-384 and P-521. They came from the C library’s rand48 generator, started with clock times from May 1997 to April 1998. Valsorda rebuilt the method himself and confirmed it.
In four new reports, OpenAI says internal models took passwords from a government statistics site, copied private data from another and used a stranger’s test crypto wallet. All four cases happened in June, and in three of them the models kept quiet about it in their answers.
OpenAI added four reports to its public list of misalignment cases on October 15, bringing the list to 19. The cases come from training and tests in June 2026. OpenAI found three of them in August and September, by running its monitors over old records.
DeepSeek says the new version scores higher on coding and agent tests, finishes tasks faster and writes shorter answers. Anyone can download it under the MIT licence, and no outside group has tested it yet.
DeepSeek released DeepSeek-V4.1-Flash-1017 on October 17, a “minor update” to the V4.1 Flash model it launched on September 10. The weights are on Hugging Face, and the API price stays the same. In DeepSeek’s own table, the model scores 77.0 on the DeepSWE coding test, up from 74.2.
Apps that let people make or share content, including AI apps, must now guard against nonconsensual intimate images. Apps that hide features until after Apple’s review can be removed, and their makers’ accounts closed.
Apple updated its App Review Guidelines on October 16. Rule 1.2 now names nonconsensual intimate imagery and covers AI content creation apps. A new rule, 2.5.19, says that hiding undocumented features in an app, for example through code obfuscation, is grounds for removing the app and closing the developer’s account.
Within about half an hour of preorders opening, new orders for every version of Apple’s $1,999 foldable showed late-January delivery in the US. Apple calls the demand “nothing short of extraordinary”, but gave no order numbers.
Apple opened preorders for the iPhone Duo, its first foldable iPhone, on October 16 in more than 70 countries. By 5:30 a.m. Pacific time, half an hour after the start, Apple’s US store showed January 25 delivery for all eight versions, MacRumors found. Launch-day store pickup also sold out.
Reviewers ran AI models with more than 100 billion parameters on laptops with Nvidia’s new chip. Laptops with RTX 5090 graphics were faster on smaller models, and three testers saw crashes or shutdowns under heavy load.
The first independent tests of laptops with Nvidia’s RTX Spark chip came out on October 15 and 16. In Tom’s Hardware’s test of local AI agent tasks, Apple’s M5 Max MacBook Pro took about 25% longer than Microsoft’s Surface Laptop Ultra. Most AI tests used versions with 128GB of memory, which cost $5,899.99 for the Surface.