OpenAI and Anthropic Post Rival AI Proofs Disproving the Exponential Time Hypothesis
OpenAI released three AI-generated proofs on Monday spanning mathematics, physics and computer science, including a disproof of the Exponential Time Hypothesis that an internal model found on 16 October. Anthropic posted its own independent disproof of the same conjecture hours later. The model and prompts remain undisclosed, and the scientists who proposed the hypothesis are now organising an independent review.
Read full story →Microsoft Offers Voluntary Buyouts to 60,000 Employees, a Quarter of Its Workforce
Microsoft said Monday it is offering voluntary buyouts to about 60,000 staff, roughly 27 percent of its workforce, across four countries, even as it pours billions into AI data centres. The offer lands the same week a JLL survey of 1,321 organisations found 89 percent shifting facilities budgets from labour toward automation, with client robotics adoption projected to nearly double by 2030.
Read full story →Chinese Startup NeodraDynamics Unveils 13-Foot Neo1 Robot Claiming 500kg Lift
Beijing's NeodraDynamics showed Neo1, a four-metre, 1.2-tonne hydraulic humanoid it says can carry 500-kilogram loads, at the China International Industry Fair this month. The company aims the machine at mining, heavy assembly and disaster zones, but has not demonstrated its claimed payload or 0.01mm precision. The launch fits China's five-year plan placing AI robots at the centre of its economy.
Read full story →India Picks Sarvam AI to Build Military Tools for Army, Navy and Air Force
India's Ministry of Defence signed a memorandum with startup Sarvam AI to build decision-support, target-acquisition and data-fusion tools across the three armed services, the company said Tuesday. The work goes to Chanakya, a firewalled national-security division using Sarvam's own 105B foundation model. The deal keeps sensitive military data inside a domestically controlled AI stack, though no contract value or timeline was disclosed.
Read full story →