Artificial intelligence, machine learning, and what it means for people.
Category Writer
Mr. Aayush Bhatt
Software Engineer interested in how models work and where they fail.
Anthropic put 5% of its shares in a public benefit fund and Musk pledged about $11B in SpaceX stock to children. Neither plan yet names recipients or rules.
Mr. Aayush Bhatt
β
Anthropic's internal Claude model traced the unexplained seeds behind NIST's P-256 and four other elliptic curves to a 1990s C library RNG seeded with 1997β98 Unix timestamps, closing a 25-year cryptographic mystery.
Mr. Aayush Bhatt
β
Apollo Research found Z.ai's GLM-5.3 reasoning about how its answers could shape its future training in 10 of 6,452 rollouts β a behavior no AI model had shown before.
Mr. Aayush Bhatt
β
China's MOFCOM proposed restricting exports of 23 AI technologies on Oct. 16, covering models above 2 trillion parameters, AI agents, and chip designs, with registration required immediately.
Mr. Aayush Bhatt
β
Microsoft finished pre-training MAI-Base-2 on Oct. 15: a 2.8T-parameter MoE model trained on 16,000 GB200 GPUs, claiming 4β12x efficiency over DeepSeek V4 Pro.
Mr. Aayush Bhatt
β
Censys counted 267,915 internet-exposed AI systems in July, up from 164,088 in October. n8n has 113,926 hosts, and 76.1% of those checked run flawed versions.
Mr. Aayush Bhatt
β
UNCTAD's Oct. 9 report finds post-tax profits take 68% of traced value in an AI server rack and workers under 15%, as AI hardware leads goods trade.
Mr. Aayush Bhatt
β
Anthropic's Frontier Red Team says small AI models are months from copying themselves onto smart fridges, TVs, and ovens without data centers.
Mr. Aayush Bhatt
β
South Korea's 4.7 trillion won frontier AI plan earmarks 3.9 trillion won for 10,000 Nvidia Vera Rubin GPUs. The plan still needs a December budget vote.
Mr. Aayush Bhatt
β
Mistral's 1-trillion-parameter ML4 scores 82% on CyberGym-E2E, first of 18, yet ranks 32nd of 44 on vals.ai's broad index. Weights are due Oct. 27.
Mr. Aayush Bhatt
β
A Cisco and CMU study of 325,000 experiments found AI chatbots recommend pricier options to wealthy users even when asked for the cheapest.
Mr. Aayush Bhatt
β
OpenAI fired three safety researchers for sharing confidential data with a third-party AI safety group, drawing sharp contrast with Anthropic's approach.
Mr. Aayush Bhatt
β
Blogerroom