AI News
Latest AI news, product launches, research breakthroughs, startup moves, and industry analysis from across the artificial intelligence ecosystem.
NASA and IBM released an open lunar AI model trained on LRO data for crater mapping, volcanic features, polar ice, and surface-change research.
Google Research’s ToolGrad framework uses an answer-first approach for tool-use data generation and reports a 99.8% pass rate.
OpenAI Says AI Found Possible Navier–Stokes Breakdown, Prompting Debate Over Proof and Credit
OpenAI says thousands of AI agents found a Navier–Stokes blowup, but human mathematicians are still reviewing the proof and its attribution.
NVIDIA is working with Australian cloud and infrastructure partners on land, power and data center capacity for up to 2GW of AI compute by 2027.
Anthropic alignment lead Evan Hubinger estimates AI has over a 10% chance of killing all humans within the next decade.
The UN human rights chief warns that AI could pose an existential risk, while key details about the threat and possible safeguards remain unclear.
OpenAI’s page indicates support for independent journalism in Ukraine, but provides no details on funding, partners, eligibility, or delivery.
Arm’s CSS for Mobile 2 combines the Mali G2-Ultra NX GPU, C2 CPU cluster and software tools for on-device agents and neural mobile graphics.
Axis Robotics releases Axis A, a browser-based engine with 207 robot manipulation tasks and 50,129 trajectories, with access details still limited.
Perplexity names Ivy, Tulip, and Rose as components of the GPU serving stack behind PPLX-Embed, but leaves key technical details undisclosed.