Today's stories
DeepSeek launches V4.1-Flash and retires V4-Pro, its flagship model
T
S
Covered by 2 sources Mistral wants open-weight AI to compete at the frontier. It just raised $3.5 billion to do it.
T
The New Stack (AI) Salesforce introduces Enterprise AI Harness, AI Control Plane
S
SiliconANGLE OpenAI Releases GPT-6 Astra for Coding and Computer Use
I
InfoQ (AI, ML & Data) OpenAI’s Supposed Mathematical Breakthrough Devolves Into Explosive Drama as Mathematician Accuses It of Stealing His Work
F
Futurism GPT Images 2.5 promises edits that leave the rest of your image alone
T
The New Stack (AI) Arlequin AI raises €28M to build novel AI models that learn complex relationships at scale
S
SiliconANGLE NVIDIA Details BioNeMo Inference Runtime (BioIR): 2.90x Higher Boltz-2 Folding Throughput and 58.5K Residues per GPU-Hour on 8xH100
M
MarkTechPost Anthropic caught scientists using Claude to further biological weapon research
E
Engadget ToolGrad: Efficient tool-use dataset generation with textual "gradients"
G
Google Research Blog Nvidia and Palantir fine-tune a 30B Nemotron model for Nvidia’s supply chain. It beats a model 18 times its size.
T
U
T
Covered by 3 sources CoreWeave Puts Field Engineers Inside Customer Teams for Physical AI
U
S
Covered by 2 sources Weatherwatch: AI model beats standard methods at predicting cyclones
T
The Guardian OpenAI pulls out of Caltech math hackathon after mathematicians issue open letter criticizing AI math 'slop'
B
Business Insider Reduce LLM latency with prefix-aware routing on Amazon SageMaker Inference
A
AWS Blog Model-agnostic PII detection with LLMs
A
AWS Blog Meet Redis LangCache: A Managed Semantic Cache That Cuts LLM API Costs by Up to 90% and Returns Cache Hits Up to 15x Faster
M
MarkTechPost Reduce inference cold starts on Amazon SageMaker HyperPod with model caching
A
AWS Blog Abacus.AI Releases Three Open-Weight Smaug Models for Agentic Workloads
U
Unite.AI Top AI researcher Andrew Tulloch ditches Meta to work on training and inference at Anthropic
B
Business Insider Cohere Debuts Open-Weight 218B Mixture-of-Experts Machine Translation Model
U
Unite.AI Antioch raises $32m to test robots in simulation before they touch hardware
T
The Next Web Alibaba Backs Ex-Staffer’s AI Testing Lab at $2.5 Billion Value
B
Bloomberg The AI You Didn't Know You Bought
F
Forbes From Process Control To Industrial AI: What Actually Changes
F
Forbes Agent Evaluation Metric for multi-turn conversations
A
AWS Blog NYU-DRP AI Model Predicts Five-Year Breast Cancer Risk From 3D Mammograms
U
Unite.AI Is AI The Silver Bullet For Project Portfolio Management?
F
Forbes 47,000 job listings reveal the engineering roles that AI is creating
T
The New Stack (AI) Certinia launches new System of Action as the foundation of autonomous service operations
S
SiliconANGLE Feature Engineering in Scikit-Learn: A KDnuggets Cheat Sheet
K
KDnuggets What Is MLOps? How Teams Build, Deploy, and Monitor Machine Learning Systems
U
Unite.AI LandingAI Releases Agentic Document Extraction Gen2 with DPT-3 Pro and DPT-3 Verity
M
MarkTechPost In AI, the Walls Aren’t Real: Rethinking Data & Safety
U
Unite.AI This week
Last Week in AI #343 - GPT-6, OpenAI’s agents chatted on a wiki, Fable 5.1
L
T
A
+4Covered by 7 sources The Fight Over OpenAI’s Math Breakthrough Is a New Kind of Scientific Arms Race
G
T
Covered by 2 sources A clinically-oriented foundation model for intraoperative pathology
N
Nature Google and NASA JPL Unveil AI Model Mapping Global Methane Plumes
U
T
Covered by 2 sources When pathology segmentation learns to listen
N
Nature IBM releases SOTA Granite Time Series PatchTST-FM-r2 model with commercial-friendly license
H
Hugging Face IBM Releases Granite PatchTST-FM-R2 Zero-Shot Time Series Model
U
Unite.AI K2 Horizon just shipped as six new fully open models — developers aren’t fully convinced
T
The New Stack (AI) CloudNC aims to accelerate AI supply chain machining
A
U
Covered by 2 sources Lightbits set to release KV cache engine to boost GPU performance
S
SiliconANGLE Claude performed best on a new benchmark for ‘agents that build agents’. But it passed fewer than a quarter of the tests.
T
The New Stack (AI) Man Announces That He Has Synthesized a New Schizophrenia Treatment in His Garage, Based on a Formula Devised by ChatGPT
F
Futurism Qualtrics Announces XM Data & AI Platform for Experience Management
U
Unite.AI Alibaba.com Says Accio Ran E-Commerce Tasks at Over 50% Lower Cost
U
Unite.AI Where AI Adds Value: SKF Uses AI For Smart Product Defect Detection
F
Forbes Databricks adds adaptive search model to speed agent retrieval
S
SiliconANGLE Wint Raises $36 Million Series D to Scale AI-Powered Water Intelligence Globally
U
Unite.AI Modernizing complex legacy code with AI agents.
M
Mistral AI News Cleveland Clinic, RIKEN, IBM named Gordon Bell finalists
I
IBM Research AI model predicts which breast-cancer drugs work best
N
Nature Reducing cost of fluorescent microscopy through dye exclusion, distillation techniques, and deep learning
N
Nature Breaking timescales with generative sampling of conformational transitions
N
Nature Explainable data generation from observed small samples by matrix tri-factorization
N
Nature Pathway’s brain-inspired architecture development on Amazon SageMaker HyperPod
A
AWS Blog OpenAI fought dirty on career-making math problem, says NYU mathematician
T
TechCrunch S. Korea, France sign multiple pacts on cooperation, including Samsung-Mistral AI investment
Y
Yonhap News Agency That's everything we have on Machine Learning right now