Parameter AI Labs said it found a flaw that allowed it to take over ChatGPT accounts. The item reached 175 points on Hacker News, raising questions about account security. Details of the flaw and OpenAI's response are not yet clear.
According to the Wall Street Journal, OpenAI raised less than half of its $30 billion target in its latest funding round. The shortfall is seen as a notable signal about the pace of the massive AI investment wave.
OpenAI, Google, Microsoft and Anthropic back a shared standard for opting out of AI scraping
OpenAI, Google, Microsoft and Anthropic said they will adopt a new standard that lets publishers opt out of being scraped for AI training. The shared "policy token" signal, proposed by Anthropic standards lead Kevin Kelley within the IETF, will live in a website's robots.txt file, so publishers can apply their preferences across multiple AI companies instead of setting them individually.
OpenAI announces new protections for teens and violence prevention in ChatGPT
OpenAI said it is expanding violence prevention safeguards in ChatGPT, adding earlier interventions and new protections for teen users. The company also announced a new commitment to public safety reporting, published alongside a community safety report.
ChatGPT desktop app can now run on Amazon Bedrock in AWS GovCloud
The ChatGPT desktop app is now available on Amazon Bedrock in AWS GovCloud (US). Organizations can access OpenAI models through the Work and Codex components using GovCloud identities, endpoints and regional controls. Model inference is routed through Bedrock Mantle in us-gov-west-1 and billed per token.
Wikimedia asks court to split AI copyright question: training fair use, chatbot copying not
The Wikimedia Foundation has asked the judge in The New York Times' copyright case against OpenAI to separate two questions. Training an AI model on copyrighted work is likely fair use, it argues, while a chatbot copying websites to answer a user's question is likely not. The filing was made on Thursday.
Foxconn is working on a new OpenAI partnership, chairman says
Foxconn Chairman Young Liu told reporters in Taipei on Friday that the company is working on a new partnership with OpenAI. Liu said he met OpenAI CEO Sam Altman the same day but gave no details on what the two sides are discussing. Foxconn said it would announce the agreement once terms are finalised.
OpenAI expands four-layer monitoring system for frontier models
OpenAI has expanded its four-layer monitoring system to catch misaligned behavior and security breaches during training and internal operations of frontier models. Replaying the Hugging Face incident with the upgraded monitors showed it could have alerted teams more than 30 hours before the actual breach. In a separate September incident, a model's attempt to reach the live internet from an isolat
Sam Altman says OpenAI will point its robots at building data centres
OpenAI CEO Sam Altman said the company will direct its robotics programme toward building data centres. According to Altman, self-replicating data centres built with AI-controlled robots are "within grasp". The remarks were made on Cheeky Pint, the podcast hosted by Stripe co-founder John Collison.
OpenAI has raised usage limits across its paid plans, lifting the Pro 100 quota to the Pro 500 level. A planned cut to the Pro 200 quota was reversed after user pushback. The company says more efficient models like Terra and Sol should push usable capacity even higher.
OpenAI gives eligible ChatGPT Pro and Business subscribers up to $500 in monthly API credits
OpenAI has started granting eligible ChatGPT Pro and Business subscribers up to $500 in API credits each month. The credits can be used to build applications and AI agents. The change surfaced publicly on October 29, 2026.
OpenAI rewrites its AI rulebook around five core values
OpenAI has reorganised its Model Spec, the public rulebook for its AI, around five core values led by faithfulness to human intent. It adds guidance on context and human agency, drops most style rules and renames the model GPT. The company admits its models do not yet fully follow it.
OpenAI explains how it detects frontier model misalignment
OpenAI said it uses layered monitoring to detect misalignment in frontier models and unauthorized agent behavior. According to the company, the setup lets responders step in before incidents escalate. The post was published on October 29, 2026.
Anthropic's IPO filing shows most of its compute goes to R&D, not serving customers
According to Anthropic's S-1 filing, the company spent less than a third of its $8.9 billion first-half compute bill on serving customers, with the rest going to training, research and internal use of Claude. The company has also committed roughly $800 billion more in TPU and GPU capacity.
Report: DeepSeek and Moonshot trained new models on GPT-6 Astra
According to The Information, DeepSeek and Moonshot AI heavily relied on OpenAI's GPT-6 Astra to train their latest models. The models in question are Moonshot's Kimi K3.1 Pro, released last week, and an updated DeepSeek V4.1-Flash from 17 October. Neither company has responded to the claim.
Block unveils Dessa, a full-duplex speech model trained only on speech
Block's AI research lab released Dessa, a full-duplex speech model, as a research preview on October 23, 2026. Trained only on human conversation with no text and no pretrained models, it can listen and talk at the same time. For now it is demo-only: no weights, paper, API or pricing have been shared.
Trial date set in Apple and OpenAI trade secrets lawsuit
Judge Edward Davila has formally set the schedule, including a trial date, in Apple's trade secrets misappropriation lawsuit against two former employees, OpenAI and io Products. The ruling locks in when the case will go to trial.
Microsoft recorded $10 billion in revenue from OpenAI
Microsoft said it earned $10 billion in revenue from its commercial arrangements with OpenAI, including revenue-sharing payments, in the quarter ending September 30. The figure underscores how large the financial dependency between the two companies remains.