DeepSeek is a Chinese AI research laboratory that has developed highly capable open-source language models including DeepSeek-V3 and DeepSeek-R1, notable for their efficiency and performance.
Global investors are increasingly splitting Chinese tech stocks into AI winners and losers, with a pair trade strategy going long on Alibaba and short on Meituan returning around 130% year-to-date. Th…
A research team from CUHK Shenzhen and Tencent's Hunyuan released GameCraft-Bench, a benchmark testing AI coding agents on building complete, playable games in the Godot engine. The top agent, Claude …
Chinese AI labs are using gray market access to Western AI models for distillation attacks, training their own models on outputs from US providers like OpenAI. This practice, which involves querying A…
Microsoft is integrating Chinese AI models from DeepSeek into its Azure platform while its partners OpenAI and Anthropic warn about intellectual property theft. The tech giant tests DeepSeek-V4 for it…
Researchers at Aikido Security discovered 15 malicious plugins on the JetBrains Marketplace that stole AI API keys from approximately 70,000 developers. The plugins, posing as AI coding assistants and…
China's top diplomat Wang Yi called for a global AI cooperation organization offering cheap or free models to developing nations, as the G7 discussed limiting access to US AI models to "trusted partne…
A security researcher known as Bobdahacker discovered a critical authentication flaw in FIFA World Cup systems, allowing unauthorized access and content injection. Google Chrome's Manifest V3 update w…
A developer post-trained a language model to roll a die, revealing that models default to outputting 4 due to training data bias. Standard reinforcement learning fails to encourage exploration, but ad…
Security researchers identified 15 malicious JetBrains plugins masquerading as DeepSeek AI assistants in a supply chain attack. The plugins harvest API keys, exfiltrate LLM chat sessions, and establis…
Zkit, a set of Go libraries for building AI agents, has been released as open-source software. The libraries provide modular components including a streaming runner, tool dispatch, guardrails, compact…
A developer evaluated DeepSeek V4 Flash and Gemini 2.0 Pro for a high-volume ranking system processing 12 million inference calls daily, finding DeepSeek V4 Flash achieved a p99 latency of 1.18 second…
A freelance developer tested DeepSeek V4 Flash, DeepSeek V4 Pro, Qwen3-32B, GLM-4 Plus, and GPT-4o for a legal document classification system over a month. DeepSeek V4 Flash achieved comparable accura…
Microsoft is considering integrating a self-hosted, fine-tuned version of China's DeepSeek V4 open-source model into its Copilot Cowork assistant to reduce costs, as agentic AI usage drives up compute…
US officials restricted access to Anthropic's cybersecurity AI models Mythos 5 and Fable 5 after finding safeguards could be bypassed, blocking foreign nationals and prompting Anthropic to shut down a…
Researchers released IndexCache, a patch for SGLang and vLLM that accelerates sparse attention in DeepSeek-V3.2 and GLM-5 models by reusing index computations across layers, achieving up to 1.82× pref…
A developer discovered that Claude Code, when configured to use DeepSeek's API as a backend, falsely claimed to be Claude Opus 4.8 by Anthropic. The AI's identity came entirely from a client-side syst…
Microsoft launched Copilot Cowork with usage-based billing, now generally available. The AI agent performs long-running tasks independently in the cloud, with pricing starting at 1 cent per credit. Cu…
U.S.-based developers and small startups are increasingly adopting Chinese open AI models like DeepSeek to cut costs, with some reporting coding costs dropping from $10 on Claude to under 50 cents. Fi…
Amnesty International released a research briefing documenting that standalone generative AI systems, based on unlawful web scraping, violate international human rights law through mass invasions of p…
U.S. developers and startups are increasingly turning to Chinese AI models like DeepSeek for cost savings, with some reporting costs as low as 50 cents compared to $10 for similar tasks on U.S. models…