⚡ BREAKING

Fields Medal Finalist Joins OpenAI: "Mathematics as a Profession May Disappear"

2026 Fields Medal laureate Jakob Zimmermann announced his transition from pure mathematics to AI safety and joining OpenAI. He believes that within a few years, AI will surpass humans in solving mathematical problems, so he will no longer take on graduate students to avoid training them for a non-existent profession.

Безопасность ИИ 🇨🇳 24.07 · 15:02 InfoQ 中国 OpenAIOpenAI
Regulation 🇺🇸

Who Is Afraid of Chinese Models?

Ben Thompson suggests legalizing AI model distillation to help American open models compete with Chinese ones. Alibaba released Qwen 3.8 Max with open weights, possibly influenced by Xi Jinping's speech on openness.

Alibaba/QwenAlibaba/Qwen
Simon Willison24.07 · 18:04

WAIC 2026: main announcements and why China is turning artificial intelligence into national infrastructure

At the WAIC 2026 conference in Shanghai, Chinese companies unveiled new AI models and platforms, while authorities announced plans to transform AI into national infrastructure. Key announcements: the Kimi K3 model from Moonshot AI, a video development platform AGENT from Sohu, and initiatives to integrate AI into industry and education.

Moonshot AIMoonshot AI SohuSohu
GNews RU — ИИ24.07 · 18:03
Applications 🇷🇺

Google AI & Economy ATLAS: AI is used in 68% of occupations but only in 21% of tasks

Google released its first study, AI & Economy ATLAS, based on 15 million interactions with Gemini. AI is used in 68% of occupations, but in the median occupation it affects only 21% of tasks. The majority of queries are related not to full automation but to collaboration between humans and models — ideas, strategy, information retrieval, learning.

Google/DeepMindGoogle/DeepMind
Habr — хаб ИИ24.07 · 18:02
Agents 🇷🇺

Troubleshooting AI Bug Hunting: How Not to Become a Free Triage for Your Agent

AI agents generate many false positives, turning bug hunters into triagers. The article examines approaches to AI bug hunting, model payment methods, and verification architecture using rules, RAG, and MCP. It includes data from HackerOne and Bugcrowd on the increase in reports and the share of valid findings.

AnthropicAnthropic CursorCursor HackerOneHackerOne OpenAIOpenAI OllamaOllama LM StudioLM Studio vLLMvLLM AWS BedrockAWS Bedrock Google Vertex AIGoogle Vertex AI BugcrowdBugcrowd
Habr — хаб ИИ24.07 · 16:03
Fresh news