China Telecom, a state-owned business, says it has used chips designed in-house to create two large language model AIs trained to consider a trillion variables at a time.
Tag: AI
Post
ENDOR LABS OFFERS TEST FOR OPEN-SOURCE AI MODELS
Endor Labs, a Silicon Valley company specializing in secure software development, has created a screen to test the quality of open-source software models available on Hugging Face and other sites where the models can be downloaded.
Post
MAJOR AI SYSTEMS FAIL EUROPE’S AI ACT COMPLIANCE TEST
A number of AI models have work to do if they’re to comply with key provisions of the European Union’s AI Act, passed by the union’s parliament last May.
Post
GOOGLE COMMISSIONS MINI NUCLEAR POWER PLANTS TO RUN DATA CENTERS
Google has signed what it says is "the world's first corporate agreement to purchase nuclear energy from multiple small modular reactors.”
Post
OPENAI CREATES SAFETY CHECKLIST FOR AI
Engineers at OpenAI have designed a suite of tests to determine whether a future AI should be considered “artificial general intelligence” (AGI)—an AI that not only can outperform humans in most mental abilities but also can evolve new versions of itself without human intervention.
Post
THIS WEEK IN SURVEILLANCE
A new Department of Defense (DoD) directive issued in late September very arguably runs afoul of longstanding prohibitions on the U.S. military being used in domestic operations and surveillance against American citizens.
Post
NEEDED: A NATURAL HUMAN REVOLUTION
In a 10 October X post, Peter Diamandis, co-founder of Singularity University and protege of preeminent transhumanist Ray Kurzweil, endorsed good AI fighting bad AI as the only way humanity would be protected in the future.
Post
META DEBUTS FILMMAKING AI
Meta has unveiled Movie Gen, its AI that can create realistic videos from written instructions. The system will play catch-up with similar offerings already out, including OpenAI’s Sora and developer Runway’s version.
Post
MOZILLA URGES “PUBLIC AI”
Mozilla, the nonprofit that operates the Thunderbird email program and Firefox browser, is urging AI developers to create a parallel AI industry devoted solely to public benefit—one that creates AI platforms just for nonprofits, government agencies, and companies emphasizing social goals.
Post
OPENAI TO RESTRUCTURE AS PUBLIC BENEFIT CORPORATION
OpenAI is considering transforming itself into a profit-seeking corporation that also is legally committed to socially beneficial works.