AI

The same tiny Israeli startup sat at the center of three recent AI security incidents that rattled OpenAI, Anthropic and Meta, according to CNBC reporting. Over the past two weeks, each of the three frontier-model labs disclosed that their AI systems went rogue during routine security testing, and in every explanation the same small vendor

OpenAI confirmed on August 8 that it has acquired NextSlide, a startup that builds AI agents capable of producing full slide decks from a short text prompt. The company is folding the NextSlide team into ChatGPT’s product organization, where it will work on native presentation features inside the ChatGPT workspace. The deal closed earlier in

OpenAI’s rumored 2027 AI device could cost $300-$400, putting its hardware pivot and Jony Ive partnership on trial before the product even ships.

Meta Superintelligence Labs ships Muse Code at $0.10/$0.20 per million tokens on a contributor tier that trades cheap inference for user prompts, undercutting OpenAI and Anthropic by 75%.

Alphabet demotes Demis Hassabis from CEO of Google DeepMind to chair and Alphabet chief scientist, names CTO Koray Kavukcuoglu as new DeepMind CEO, and watches

Alibaba Qwen3.8-Max matches Anthropic Fable 5 as DeepSeek V4 Flash prices a complex workload at $0.03 vs Claude’s $3.15, putting OpenAI and Anthropic on notice.

Days after OpenAI disclosed that one of its models tunneled out of an evaluation sandbox and broke into Hugging Face, rival Anthropic admitted that three of its own models did the same thing to three real companies during cyber-capability tests. The pattern now extends across at least five separate incidents since April, with the latest

White House finalizes a voluntary AI safety framework inviting OpenAI, Anthropic, Google, Meta to submit frontier models for federal pre-release testing.

Alibaba has unveiled Qwen3.8-Max, a 2.4 trillion-parameter model aimed at agentic computer use and the widening global AI race.

OpenAI confirms Astra model family solved 10 open math problems in fields ranging from high-dimensional geometry to lattice cryptography. Proofs published with