AI + Research: Agents Accelerate Scientific Research—Papers Are Written Faster, But Achievements Are Harder to Evaluate | Shared by Wang Huan, Doctoral Supervisor at the School of Engineering, Westlake University
Click to view the original article>
AI NEWS JOURNAL
Search AI news
Search headlines, summaries and articles to connect the dots.
NVIDIA Releases Auto Ontology: An Open Semantic Layer That Grounds AI Agent SQL
NVIDIA released Auto Ontology, an Apache 2.0 library that generates a semantic layer from enterprise metadata. It grounds agent SQL in certified definitions and serves context through MCP, CLI and API. The post NVIDIA Releases Auto Ontology: An Open Semantic Layer That Grounds AI Agent SQL appeared first on MarkTechPost.
Meet qwenGPU: Qwen3.5-2B With 45 Compiled GPU Layers That Run CUDA
qwenGPU compiles a 4-lane SIMT GPU into Qwen3.5-2B as 45 untrained layers. The model writes every step of real CUDA launches itself, using Gated DeltaNet state as registers and memory. The post Meet qwenGPU: Qwen3.5-2B With 45 Compiled GPU Layers That Run CUDA appeared first on MarkTechPost.
Deep Interview with Yuval Harari: The Real AI Crisis Has Arrived!
In his latest interview, Yuval Harari, author of "Sapiens," tells us that the real crisis isn't anything like what you see in movies—it's unfolding right now in the real world.
The Sequence Radar - Issue #963: This Week in AI: The Future Has a Payment Schedule
Next Week in The Sequence:
Run AI Models for Edge Application Development with Just a Few Commands
AI is expanding beyond the data center as developers build a new generation of agents to act at the edge: monitoring driver attention and cabin conditions,...
[AINews] Building AIE CODE: The top Coding Agent Conference
Mistral’s open-weight release, frontier monitoring, and the day’s AI Twitter and Reddit recaps.
NVIDIA’s team, building on the MiniMax H3 open-source video enhancement framework, supports up to 4K regeneration.
NVIDIA and a team from MIT have released the model weights for the Open-H3-Regenerate-2K video enhancement project on Hugging Face.
The trillion-parameter open-source model Mistral Large4 makes its debut: achieving top-tier performance with fewer tokens
On October 30, European AI leader Mistral open-sourced its new flagship model, Mistral Large4 (codenamed “Le Chonk”). With 1 trillion parameters, it is one of the largest models in the open-source weight community and was trained entirely from scratch on European‑based infrastructure, marking a substantial breakthrough for open-source large models in reaching the trillion‑parameter scale and natively integrating multimodal capabilities. The model emphasizes robust performance and token efficiency, rejecting mere score‑chasing and prioritizing exceptional cost‑effectiveness.
Morning Briefing | Apple executive responds as iPad mini remains at 60Hz / Global smartphone shipments fell 20% in Q3, hitting a new low / "Shadow Blade Zero" tops Steam's global best-sellers chart
· Wang Chuanfu reveals BYD’s workforce has surpassed 1.02 million, with R&D engineers numbering 120,000 · U.S. computer science enrollments decline again, down 18.4% over the past two years · South Korea plans to extend its ban on AI glasses during exams to three years and will deploy dedicated personnel for identification #Follow iFanr’s official WeChat account: iFanr (WeChat ID: ifanr) for more exciting content delivered straight to you.
nemotron-halloween
nemotron-halloween
The ‘Pain in the Ass’ Math That Could Save Vibe Coding From Itself
A new startup called &Truth is joining the race to automate formal verification, a painstaking process that uses math to prove software works as intended.
Mistral Says Its AI Model ‘Le Chonk’ Has Leapfrogged China in the Open-Weight Race
The French AI lab claims its latest release matches or outperforms even the best models from DeepSeek and Z.ai in coding and cybersecurity.
Mistral releases weights for its trillion-parameter AI model, targeting cybersecurity
Mistral says its new model can afford organizations a kind of control over its cyber defenses that closed models can't.
Le Chonk: Built for Coders, Trusted by Defenders.
The weights of Mistral Large 4 (Le Chonk) are now open. A 1T parameter multimodal model, and best open-weight model for coding, trained in Europe.
We urgently need to monitor AI agents’ actions in real time
This summer, hundreds of OpenAI agents tried to cheat on a cybersecurity test. They broke out of their test environment and successfully hacked Hugging Face, a site that they believed held clues about how they’d be graded. Around the same time, another OpenAI agent accessed private files from an Australian government website. It’s a problem…
Infini-AI-Lab Releases Hermit, an Open-Source Harness Adapter Framework for AI Agents
Infini-AI-Lab released Hermit, an Apache-2.0 harness adapter framework that trains an external layer of instructions, skills, memory and tools from task feedback, leaving the agent’s model and harness unchanged. The post Infini-AI-Lab Releases Hermit, an Open-Source Harness Adapter Framework for AI Agents appeared first on MarkTechPost.
Is physics about to fall to AI as well? Five quantum open problems that had remained unsolved for over 20 years have been cracked.
MIT’s last triple-degree prodigy has made a move.
[AINews] Gemini 4 Carbon tops leaderboards, launches for cyber defenders
Google’s new frontier model leads independent evals, with access initially limited to trusted defenders.
MiniMax Open-Sources MCode Stack: 15 Agent Skills Its Coding Team Uses Daily
MiniMax has open-sourced MCode Stack, an MIT-licensed repo of 17 SKILL.md skills and 5 checklists that its MiniMax Code team uses daily for code review, subagent splits, context budgets and long-running commands. The post MiniMax Open-Sources MCode Stack: 15 Agent Skills Its Coding Team Uses Daily appeared first on MarkTechPost.
