https://www.theai.news/briefs/2026/08/ten-advances-in-mathematics-and-theoretical-computer-science-e13c90fd The AI News en 2026-08-01T21:21:15.103Z Ten advances in mathematics and theoretical computer science https://www.theai.news/briefs/2026/08/i-ve-had-ling-3-0-flash-and-glm-5-2-both-in-my-executor-slot-0716bfe4 The AI News en 2026-08-01T18:22:14.683Z I've had ling-3.0-flash and glm-5.2 both in my executor slot for a few weeks. They don't split the way the benchmarks predict https://www.theai.news/briefs/2026/08/a-security-researcher-built-a-self-spreading-worm-that-hides-7e3ccd39 The AI News en 2026-08-01T14:11:35.460Z A security researcher built a self-spreading worm that hides inside Word docs and hijacks Microsoft Copilot https://www.theai.news/briefs/2026/08/german-court-rules-ai-music-generator-suno-violated-copyrigh-fdd20ea3 The AI News en 2026-08-01T11:11:30.116Z German court rules AI music generator Suno violated copyrights, rejects fair use defense https://www.theai.news/briefs/2026/08/everyone-is-building-llm-routers-we-deprecated-ours-3b20d3ea The AI News en 2026-08-01T04:02:28.167Z Everyone is building LLM routers, we deprecated ours https://www.theai.news/briefs/2026/07/60-82-accuracy-swing-on-4b-model-classification-task-the-onl-4ea45637 The AI News en 2026-07-31T23:03:20.700Z 60-82% accuracy swing on 4B model classification task: the only variable was harness design https://www.theai.news/briefs/2026/07/new-deepseek-flash-model-matches-openai-s-gpt-5-6-luna-at-ro-ab19e7a2 The AI News en 2026-07-31T20:44:27.220Z New Deepseek Flash model matches OpenAI's GPT-5.6 Luna at roughly 60 percent lower cost https://www.theai.news/briefs/2026/07/judge-says-trump-admin-still-lacks-evidence-for-anthropic-su-57519c60 The AI News en 2026-07-31T20:42:23.530Z Judge says Trump admin still lacks evidence for Anthropic ‘supply-chain risk’ label https://www.theai.news/briefs/2026/07/the-ai-slowdown-is-coming-124713b6 The AI News en 2026-07-31T20:33:54.972Z The AI slowdown is coming https://www.theai.news/briefs/2026/07/ponytail-skill-for-claude-code-does-it-really-cut-agent-code-6226c255 The AI News en 2026-07-31T20:24:55.187Z Ponytail Skill for Claude Code: Does It Really Cut Agent Code by 54%? https://www.theai.news/briefs/2026/07/how-frontier-labs-are-building-subtle-developer-lock-in-ed3a92e6 The AI News en 2026-07-31T20:13:54.144Z How Frontier Labs Are Building Subtle Developer Lock-In https://www.theai.news/briefs/2026/07/public-intelligence-87d0385c The AI News en 2026-07-31T20:13:54.123Z Public Intelligence https://www.theai.news/briefs/2026/07/chinese-hacker-commands-deepseek-via-telegram-to-launch-auto-cc382b97 The AI News en 2026-07-31T17:32:21.090Z Chinese Hacker Commands DeepSeek via Telegram to Launch Autonomous Attacks https://www.theai.news/briefs/2026/07/measuring-llms-ability-to-perform-cryptanalysis-dde7f683 The AI News en 2026-07-31T17:22:07.862Z Measuring LLMs’ Ability to Perform Cryptanalysis https://www.theai.news/briefs/2026/07/why-haven-t-organoids-solved-all-of-drug-discovery-9ed863f9 The AI News en 2026-07-31T17:14:32.728Z Why haven't organoids solved all of drug discovery? https://www.theai.news/briefs/2026/07/escaping-linux-sandboxes-via-pipewire-cve-2026-5674-f87eae74 The AI News en 2026-07-31T17:14:32.556Z Escaping Linux Sandboxes via PipeWire (CVE-2026-5674) https://www.theai.news/briefs/2026/07/pg-llm-benchmarking-general-purpose-language-models-for-prot-b183cf29 The AI News en 2026-07-31T17:06:21.158Z PG-LLM: Benchmarking General-Purpose Language Models for Protein Variant Ranking https://www.theai.news/briefs/2026/07/how-we-use-goal-to-find-bugs-in-patch-the-planet-92c321ff The AI News en 2026-07-31T17:06:21.042Z How we use /goal to find bugs in Patch the Planet https://www.theai.news/briefs/2026/07/show-hn-distilling-deepseek-into-gpt-oss-doesn-t-transfer-ce-2ce71095 The AI News en 2026-07-31T05:12:39.551Z Show HN: Distilling DeepSeek into GPT-OSS doesn't transfer censorship. Try it https://www.theai.news/briefs/2026/07/why-compute-might-get-10x-more-expensive-in-coming-years-8f4fbfff The AI News en 2026-07-31T02:42:51.890Z Why compute might get 10x+ more expensive in coming years https://www.theai.news/briefs/2026/07/kimi-k3-vs-gpt-5-6-sol-on-deepswe-cost-coding-and-routing-3785cd8b The AI News en 2026-07-31T02:33:59.847Z Kimi K3 vs GPT-5.6 Sol on DeepSWE: Cost, Coding, and Routing https://www.theai.news/briefs/2026/07/chinai-368-the-affordable-luxury-of-kimi-k3-37438be8 The AI News en 2026-07-31T02:33:59.785Z ChinAI #368: The Affordable Luxury of Kimi K3 https://www.theai.news/briefs/2026/07/teaching-llms-to-update-beliefs-for-efficient-long-horizon-i-64d2125a The AI News en 2026-07-31T02:14:04.668Z Teaching LLMs to Update Beliefs for Efficient Long-Horizon Interaction https://www.theai.news/briefs/2026/07/investigating-three-real-world-incidents-in-our-cybersecurit-32633fdb The AI News en 2026-07-31T00:12:00.310Z Investigating three real-world incidents in our cybersecurity evaluations https://www.theai.news/briefs/2026/07/introducing-align-evals-streamlining-llm-application-evaluat-423c112e The AI News en 2026-07-30T23:12:23.511Z Introducing Align Evals: Streamlining LLM Application Evaluation https://www.theai.news/briefs/2026/07/nanbeige4-2-3b-i-m-not-impressed-f1a387a2 The AI News en 2026-07-30T22:54:54.592Z Nanbeige4.2-3B: I'm not impressed https://www.theai.news/briefs/2026/07/lg-ai-research-releases-k-exaone-2-0-750b-a37b-0d543486 The AI News en 2026-07-30T22:52:38.841Z LG AI Research releases K-EXAONE 2.0 750B A37B https://www.theai.news/briefs/2026/07/run-multiple-isolated-agents-in-a-single-sandbox-0677100a The AI News en 2026-07-30T22:42:46.945Z Run multiple isolated agents in a single Sandbox https://www.theai.news/briefs/2026/07/grok-voice-think-fast-2-0-now-available-on-ai-gateway-892f0c2f The AI News en 2026-07-30T22:42:46.945Z Grok Voice Think Fast 2.0 now available on AI Gateway https://www.theai.news/briefs/2026/07/inkling-small-from-thinking-machines-is-now-available-on-ai-bf96d711 The AI News en 2026-07-30T22:42:46.945Z Inkling Small from Thinking Machines is now available on AI Gateway https://www.theai.news/briefs/2026/07/inside-the-model-factory-eiso-kant-poolside-ai-542cc42e The AI News en 2026-07-30T22:33:37.509Z Inside the Model Factory — Eiso Kant, Poolside AI https://www.theai.news/briefs/2026/07/benchmarked-mindcontrol-for-llama-cpp-4cce1eff The AI News en 2026-07-30T22:25:12.515Z Benchmarked: MindControl for Llama.cpp https://www.theai.news/briefs/2026/07/china-s-apparent-ai-benevolence-is-not-unprecedented-80f43fa2 The AI News en 2026-07-30T22:25:12.138Z China’s apparent AI benevolence is not unprecedented https://www.theai.news/briefs/2026/07/echoverse-deep-evolving-environments-for-computer-use-agents-b8373a69 The AI News en 2026-07-30T22:23:06.524Z Echoverse: Deep, evolving environments for computer-use agents https://www.theai.news/briefs/2026/07/ai-gateway-gpt-5-6-pricing-and-speed-updates-09c175bb The AI News en 2026-07-30T22:15:04.375Z AI Gateway: GPT-5.6 pricing and speed updates https://www.theai.news/briefs/2026/07/we-now-have-a-better-understanding-how-openai-hacked-into-hu-e00297b3 The AI News en 2026-07-30T22:15:04.291Z We now have a better understanding how OpenAI hacked into Hugging Face https://www.theai.news/briefs/2026/07/new-mcp-specification-addresses-the-main-barrier-to-enterpri-40a904b6 The AI News en 2026-07-30T22:12:59.150Z New MCP specification addresses the main barrier to enterprise adoption https://www.theai.news/briefs/2026/07/despite-ai-hype-google-s-data-shows-workers-aren-t-automatin-3af69012 The AI News en 2026-07-30T22:12:59.121Z Despite AI hype, Google's data shows workers aren't automating themselves away https://www.theai.news/briefs/2026/07/google-and-reddit-do-not-own-the-internet-web-scraper-says-a-645450fa The AI News en 2026-07-30T22:12:58.848Z “Google and Reddit do not own the Internet," web scraper says after court win https://www.theai.news/briefs/2026/07/glm-5-2-with-vision-on-hugging-face-4ff06ebd The AI News en 2026-07-30T22:01:40.392Z GLM 5.2 with vision on Hugging Face https://www.theai.news/briefs/2026/07/gemini-robotics-2-brings-whole-body-intelligence-to-robots-658bd855 The AI News en 2026-07-30T21:52:19.074Z Gemini Robotics 2 brings whole body intelligence to robots https://www.theai.news/briefs/2026/07/deepsecbench-evaluating-model-performance-in-finding-cyberse-a731047f The AI News en 2026-07-30T21:52:19.023Z DeepsecBench: evaluating model performance in finding cybersecurity vulnerabilities https://www.theai.news/briefs/2026/07/gemini-robotics-er-2-powering-robotics-with-video-understand-bee6f41e The AI News en 2026-07-30T21:52:18.895Z Gemini Robotics ER 2: powering robotics with video understanding, task orchestration, and multi-robot collaboration https://www.theai.news/briefs/2026/07/evolib-turning-experience-into-evolving-knowledge-bf97f3c8 The AI News en 2026-07-30T21:44:43.977Z EvoLib: Turning experience into evolving knowledge https://www.theai.news/briefs/2026/07/gemini-api-managed-agents-3-6-flash-hooks-and-more-03a10449 The AI News en 2026-07-30T21:42:31.948Z Gemini API Managed Agents: 3.6 Flash, hooks, and more https://www.theai.news/briefs/2026/07/appreciation-post-dynamic-context-pruning-opencode-making-ll-2700dce9 The AI News en 2026-07-30T21:42:31.906Z Appreciation post: Dynamic Context Pruning (OpenCode) - making LLMs actively manage their context just like humans do with their working memory https://www.theai.news/briefs/2026/07/websocket-support-for-openai-responses-api-live-on-ai-gatewa-2c876088 The AI News en 2026-07-30T21:42:31.901Z WebSocket support for OpenAI Responses API live on AI Gateway