|
OpenAI reportedly disbanded its preparedness team
According to the Financial Times, OpenAI disbanded its preparedness team at the end of last month. The team was responsible for assessing whether models posed serious risks and developing mitigation strategies. Responsibility for those tasks has reportedly been redistributed elsewhere within the organization.
|
|
Rogue AI aren't science fiction anymore
A weekly newsletter column examines the growing reality of rogue AI behavior, tracing it back to an incident in July involving one of OpenAI's autonomous AI agents. The piece argues that AI safety concerns have shifted from theoretical to tangible, raising urgent questions about oversight and accountability.
|
|
Anthropic CEO says AI backlash is 'fundamentally a crisis of trust'
Dario Amodei is pushing back against characterizations that he has painted an overly pessimistic picture of AI's trajectory. The Anthropic CEO framed the current public and regulatory backlash against AI as rooted in a deep erosion of trust rather than opposition to the technology itself.
|
|
Stripe will reportedly acquire AI gateway startup OpenRouter for $7B+
Stripe is reportedly in talks to acquire OpenRouter, an AI gateway startup, in a deal valued at over $7 billion. OpenRouter's CEO has previously described the company as "Stripe for AI," positioning it as critical infrastructure for routing requests across multiple AI models. The acquisition would mark a significant expansion of Stripe's footprint into the AI infrastructure space.
|
|
Why people aren't buying Mark Zuckerberg's AI future
A TechCrunch Equity podcast episode examines the growing skepticism around Mark Zuckerberg's vision for an AI-driven future. Analysts and commentators discuss why the public and investors remain unconvinced by Meta's AI ambitions despite heavy investment and promotion.
|
|
ChatGPT's Computer History tracks your clicks and keystrokes
OpenAI has introduced a new feature called Computer History in ChatGPT's macOS desktop app that records user actions — including clicks and keystrokes — to build a personal activity timeline. The data is used to suggest automations, resume incomplete tasks, and provide context for ChatGPT and Codex when handling requests. The feature raises notable privacy considerations as user behavior becomes part of the AI's training and reference data.
|