TechArtificial intelligence
Artificial intelligence · Page 2
Page 1- Perplexity launches Teammates, AI agents only for Max subscribers
The agents work in the background and inside Slack, but the company has not published separate pricing, usage limits or availability for Mexico.
- Meta's Muse agent now builds dashboards and apps from a prompt
Meta introduced Spaces, which turns a description into a planner or a game inside the app. Muse is available only in the U.S., with no date for Mexico.
- CrowdStrike unveils a kill switch for rogue AI agents
The AI Kill Switch is available now, the company says, but it cuts agents off at the AI gateway. The version that runs on the device has no date.
- OpenAI adds GIF creation to ChatGPT with Images 2.6
The company says its image generator is 50% faster and rolling out to every plan. API pricing per token matches the 2.5 version.
- ChatGPT and Codex will share one token balance in December
Plus, Pro and Business subscribers lose their separate limits, but OpenAI has not said how many tokens each plan includes.
- Google and Microsoft to honor a shared AI opt-out in 2027
One robots.txt rule will let a site refuse AI training, but the signal is an IETF draft and no one reads it yet.
- DeepMind unveils Genie 4, a world model that generates audio
For now only a select group of testers can use it. Google says Google AI Ultra subscribers will get it "soon," with no date or countries.
- OpenAI says its AI disproved the Exponential Time Hypothesis
The company also claims its systems proved the Kakeya conjecture. No outside expert has yet verified any of the three results.
- Elvis rights holders sue Anthropic over 274 songs
They say the company copied the work to train Claude and that the chatbot still rewrites protected songs. They are seeking up to $48.6 million.
- Moonshot launches Kimi K3.1 Pro, promises open weights
The model is live in Kimi's API and apps, but the downloadable files are due by Nov. 4. Moonshot says it cut cybersecurity data from training.
- Texas sues OpenAI and Altman over ChatGPT safety
Attorney General Ken Paxton says the company sold its chatbot as safe without warning of the risks, especially to minors, and wants new models held back in the state.
- ChatGPT brings ads to 26 more countries, for a total of 89
Fourteen of the new markets are in Latin America, including Argentina, Chile, Colombia and Peru. Mexico has had the ads since August.
- Claude Duet, Anthropic's agent, opens U.S. waitlist
It is only for Pro and Max subscribers in the United States. Anthropic says it asks for approval before sending, paying or deleting.
- NVIDIA proposes stopping AI agents with its own chips
The OpenShell Reference Architecture watches the agent from a separate chip and halts it if it leaves its perimeter. The figures are NVIDIA's, and there is no price.
- NVIDIA adds Bolt, Volvo and Mercedes to its self-driving platform
Bolt will run NVIDIA's software in its European robotaxis, but the timeline hinges on permits and a Lucid delay.
- Nvidia promises DSX POD AI factories ready in 14 days
The company says its containerized design cuts setup at a prepared site from six months to 14 days. Partners would deliver from early 2027.
- OpenAI models will be sold through Google Cloud
Companies will be able to pay for them with spending already committed to Google. There is no exact date or pricing.
- Gemini comes to Android Auto with access to Gmail and Calendar
Personal Intelligence is optional, requires a personal Google account and leaves out Europe, the UK and Nigeria. Mexico is not among the exclusions.
- Anthropic says Claude traced the seeds of NIST curves
Its report says the seeds come from a 1997 number generator and that the curves are not weakened; no outside expert has reviewed it.
- OpenAI model used a stranger's crypto wallet and didn't report it
The model was working on an Ethereum test network. It paid with another account's funds and left that out of its final report; OpenAI caught it three months later.


















