This page is available in English: Home, Bio, Work, Blog, Contact, Privacy, Photo rights and the ticker. Some images may still contain Dutch text.
Archive From the archive of The Latest A.I. News
This is an archive page with the A.I. news of Wednesday 7 October 2026. This day was compiled afterwards from reports that appeared at the time, so it was not kept on the day itself; a lot may have changed since. Today's news is on The Latest A.I. News.
Compiled afterwards on
OpenAI launches GPT-6 with interactive answers in ChatGPT, Anthropic releases a cheap new Haiku, and Microsoft shows AI laptops that run agents locally.
The new feature adds tappable buttons, custom calculators and editable charts to conversations, meant to make hard topics easier to grasp. The rollout starts Wednesday for Pro, Plus, Business and Enterprise users, with Free and Go users following Thursday. Users can dial the number of visuals down.
It is the third Claude 5.5 model in a month, aimed at light work such as categorizing, summarizing, extracting data and live support. It costs $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens. It is the first Haiku with built-in safeguards against a narrow set of high-risk cybersecurity requests.
The AI laptops for locally run models and agents start at $2,600 and $3,700, the priciest configuration costs $5,900 and the top model is already sold out, Microsoft says. Windows 11 gains Execution Containers to sandbox agents. Dell announced a $3,800 RTX Spark laptop that is already open for preorder.
On the new SynthID website anyone can check whether an image, video or audio file was made with AI; the tool was previously limited to selected journalists and researchers. The watermarking technique covers content from Google's own models and is also supported by OpenAI and Nvidia, among others. Google reports roughly one million checks per day.
Robot Ventures led the Series B, with Nvidia and Samsung among the participants, bringing total funding to $158 million. The makers of the open-source Hermes Agent now target companies with Hermes for Businesses. Annualized revenue was reportedly around $36 million in mid-September.
A study by the organization labels the teen version of ChatGPT an unacceptable risk: responses to crisis situations often ended with an invitation to keep chatting, and break reminders appeared only twice across nearly 2,000 prompts. OpenAI disputes the method and says parental controls may not have been fully active.
Users build browser games without coding, in 2D or 3D and for one or several players, and can have their own images turned into game assets. Games can be kept private, shared by link or published to a gallery. The experiment is limited for now to US users aged 18 and older.
A language model is meant to spot ads that look harmless but point to illegal content, and Meta now checks where an ad leads as well as what it contains. The company says it acted against 33.2 million pieces of content in the first half of 2026, over 97 percent of it found before users reported it. A red-teaming AI agent also probes its own safeguards.
Next up: Free and Go users of ChatGPT get access to GPT-6 with Intelligent UI on Thursday, 8 October.
Every day has its own page. They stay up for three months.
Everything here is a summary of reports from other media, not original research. The label (Certain, High, Medium, Low) is my estimate of how reliable the source is. The summaries are generated automatically with AI and may contain mistakes; read the article itself for the details (the star marks the leading source). Sources: leading A.I. and tech sites and the official blogs of the A.I. companies.
