Daily AI news, sourcedOfficial news, first rumor reports and substantive progress. Each entry keeps the information published at that time.
- Three fired OpenAI safety researchers say they were never told what they actually leaked, per The Decoder
The Decoder reports that Tomek Korbak, Mikita Balesni, and Jasmine Wang are pushing back on OpenAI’s leak allegations after being fired, denying they shared confidential information and saying OpenAI never told them in writing which material they allegedly mishandled.
- Sam Altman says advanced AI will eventually escape sandboxes and only alignment can contain it, per The Decoder
The Decoder reports Sam Altman arguing that sandboxes and monitoring are temporary, that advanced AI will inevitably leave containment, and that alignment remains an unsolved problem, while warning models already find hardware flaws humans miss.
- A judge dismisses Ann Altman's sexual abuse lawsuit against OpenAI CEO Sam Altman with prejudice, per The Verge
US District Judge Zachary Bluestone on Tuesday dismissed the lawsuit Ann Altman filed last year accusing her brother, OpenAI CEO Sam Altman, of sexually abusing her during her childhood, The Verge reports. The judge wrote that she had failed to appear for an independent medical examination, and the case was dismissed with prejudice, so it cannot be refiled.
- OpenAI's annualized revenue is reportedly about $20B below earlier $70B projections, per TechCrunch
TechCrunch cites a new report claiming OpenAI's annualized revenue is about $20 billion lower than earlier figures that put it near $70 billion.
- OpenAI safety report author David Robinson quits and sounds the alarm, per The Verge
David Robinson, who wrote the safety reports accompanying every major OpenAI model release, resigned this week and published an editorial in The Atlantic arguing the industry's culture is broken. He wrote that Silicon Valley's extreme confidence and perpetual sprints build bigger models with unimpeded optimism while downplaying risks, and that frontier labs need nuclear-style safeguards, running like nuclear power plants or busy airports with layers of redundancy. The Verge notes he is part of a broader wave of departures from major AI labs, including Anthropic and Google DeepMind.
- OpenAI parts ways with three safety researchers over leaked company information, per WSJ
OpenAI has parted ways with three safety team members who allegedly shared confidential company information with an outside AI safety organization, the Wall Street Journal reports. An internal investigation confirmed the individuals mishandled sensitive information outside established company procedures. A spokesperson said the three were let go for violating policies on accessing and handling sensitive company information, breaking trust essential to the work. The departures came two days after a NYT report said executives dismissed employees' safety warnings, and follow OpenAI scrapping GPT-6.1 Astra over safety concerns. The researchers and the outside organization were not named.
- ChatGPT Pro 200 reopens to new subscribers with lower usage, per The Next Web
The Next Web reports ChatGPT Pro 200 is open to new subscribers again, but with a lower usage allowance than before, per OpenAI's Pro help page. From October 30, further plan changes take effect.
- OpenAI previews Decisions API powered by the Luna model, per The Next Web
The Next Web reports OpenAI's Decisions API entered limited preview, using the Luna model to pick from a fixed set of answers.
- OpenAI gives Codex reusable cloud environments and a voice-controlled CLI, per TechCrunch
TechCrunch reports OpenAI is expanding Codex with reusable cloud development environments that work across devices, a revamped CLI with voice controls, new code review tools and a security-focused product for scanning repositories.
- OpenAI launches $500 Pro 500 plan with Ultrafast tier, per The Next Web
The Next Web reports OpenAI launched a $500-a-month Pro 500 plan, the only Pro tier with Ultrafast access. OpenAI said Ultrafast generates up to 300 tokens per second in Codex, up to eight times faster than standard. ChatGPT Pro 200 also reopened to new subscribers, with a lower usage allowance than before.
- OpenAI expands ChatGPT plugins with app-like panels and automations, per TechCrunch
TechCrunch reports OpenAI is expanding ChatGPT plugins with dedicated sidebar homes, interactive panels, file viewers, improved discovery and support for automations.
- 20+ leading AI researchers warn automated AI research poses extreme risks, per The Decoder
The Decoder reports more than 20 AI researchers, including Geoffrey Hinton, Yoshua Bengio and OpenAI research lead Jakub Pachocki, warn of an impending intelligence explosion from self-improving AI that could soon automate all AI research.
- OpenAI has fallen behind on consumer AI agents, per The Verge
The Verge reports that as OpenAI's 2026 DevDay approaches, the company has fallen behind in one of the industry's hottest categories: continuously running, consumer-facing AI agents.
- OpenAI works to fix its math breakthrough announcements, per The Verge
The Verge reports that in a chaotic few months OpenAI has made impressive mathematics breakthroughs while repeatedly mishandling the announcements, and is now trying to do better.