#openai
-
AI labs still won't say how they'd rein in a model that slips its leash
A study by Guidelight AI Standards found that leading AI companies publish almost nothing about their plans for a model that tries to escape control. OpenAI scored best; Anthropic and Meta scored worst.
-
Nvidia Invests in Data-Center Infrastructure Developer Cloverleaf
Nvidia keeps plowing its profits into building the data centers that then turn around and buy its own chips. This time the partner is Cloverleaf Infrastructure.
-
Codex on AWS Bedrock overpays for cache: no control over prompt caching
An issue has been filed in the Codex GitHub repository: the native Amazon Bedrock provider can't enable explicit prompt caching for GPT-5.6 Sol, so agentic sessions generate a flood of expensive cache-write tokens.
-
OpenAI Is Winning Back Enterprise Customers from Anthropic
Data from the payments service Ramp shows that businesses readily jump between labs with every new model release — a worrying signal about the “stickiness” of corporate AI spending.
-
Mass ChatGPT Outage: Users Unable to Log In or Sign Up
On the evening of August 19, OpenAI's services suffered a major outage: logging in, signing up and loading chats stopped working worldwide. The company acknowledged the problem and said it was working to fix it.
-
OpenAI Answers Anthropic with a Monitoring Service That Stores No Data
OpenAI has unveiled Private Safety Processing—a system that tracks abuse across multiple sessions at once while retaining no customer data. It's the company's bid to outmaneuver Anthropic in the debate over the privacy of enterprise customers.
-
OpenAI Tightens Security After Its AI Accidentally Hacked Hugging Face
Following a July incident in which a model broke out of its sandbox and unintentionally compromised Hugging Face, OpenAI is rebuilding its research environment, monitoring and alignment methods.
-
Anthropic's Annual Revenue Rate Soars to $65B
According to Bloomberg, Anthropic's annualized revenue rate topped $65 billion at the end of July — the company added $18 billion in just two months. Investors are anticipating an IPO at a valuation of $2 trillion or more.
-
GPT-5.6 Sol: OpenAI Finally Steps Up Its Computer Vision
The Roboflow team ran OpenAI's new lineup of models through its own benchmark for visual tasks. The Sol model turned out to be the company's best ever at “vision,” though it still falls short of the market leaders.
-
Nvidia Sharply Scales Back Its Guarantees for Financing OpenAI's Infrastructure
According to Reuters, citing the Wall Street Journal, Nvidia is substantially reducing the size of the guarantees it is prepared to provide toward the construction of OpenAI's data centers.
-
OpenAI Disbands Its Risk-Assessment Team Ahead of IPO
According to the Financial Times, the company has shut down its preparedness unit, which was responsible for finding serious threats in its models, and distributed its functions among other teams.
-
ChatGPT Learns to Watch You: The New Computer History Feature Remembers Clicks and Keystrokes
The ChatGPT desktop app for macOS has gained a Computer History feature that turns a user's actions into training data. OpenAI promises it doesn't take screenshots — it only records “events.”
-
The AI Price War: OpenAI and Anthropic Cut Prices Under Pressure from China
Leading American labs are slashing the cost of their mid-tier models in a bid to hold onto customers who are defecting to cheap Chinese alternatives like DeepSeek and Moonshot.
-
OpenAI's Codex Desktop Comes to Linux
Hacker News spotted that OpenAI's Codex desktop app (ChatGPT Desktop) is now available for Linux too.
-
How Proprietary LLMs' Hidden Reasoning Got Stolen
Researchers showed that the encrypted chains of thought from Anthropic, OpenAI and Google could be recovered in plaintext by feeding them to the weakest model in the same family. The vulnerability has since been closed.
-
OpenAI Releases GPT 5.6 Cyber — but Only for a Chosen Few
OpenAI has unveiled its GPT 5.6 Cyber model for vulnerability research, pentesting and incident response. Only a select group of companies will get access to it — ordinary users won't.
-
GitHub Has Shut Down GitHub Models
Without giving any reason, GitHub has fully wound down GitHub Models, the service that offered unified access to a range of language models. Simon Willison suspects the culprit was tokens too expensive to keep handing out for free.
-
AI Safety Tests Have Themselves Become a Threat: Agents Escape Their Sandboxes
In recent months, AI agents undergoing cyber exercises have broken out of isolated environments and slipped into real systems. Experts warn that the testing infrastructure is falling behind the models' capabilities.
-
OpenAI Acquires NextSlide, a Startup That Builds Presentations from Prompts
The team behind NextSlide, a startup that turned notes and documents into finished slides, is moving to OpenAI to work on ChatGPT.
-
OpenAI Accidentally Attacked Hugging Face: A Timeline of the AI-Agent Incident
At the Black Hat conference, OpenAI disclosed how its experimental AI agents, during training, organized a cyberattack entirely on their own — and the company learned of its own involvement only when it asked to have the stolen credentials revoked.
- More