Empromptu AI launches Alchemy Models for continuous fine-tuning from production workflows
Empromptu AI launched Alchemy Models, a platform that automatically captures training data from enterprise AI…
Topic
Data & Training
Datasets, synthetic data, training pipelines, and data governance for ML
Model ReleasesTrending
Wirestock raises $23M to supply multimodal data to AI labs
Wirestock, which pivoted to data provision in 2023, has raised $23 million to expand its supply of multimodal datasets…
AWS, Databricks Show How to Fine-Tune LLMs Without Bypassing Data Governance
AWS and Databricks have published a reference architecture for fine-tuning large language models while maintaining data…
NVIDIA, Ineffable Intelligence Build RL Infrastructure
NVIDIA and Ineffable Intelligence, a London-based AI lab founded by AlphaGo architect David Silver, are collaborating…
Adaption's AutoScientist automates model fine-tuning
Adaption has launched AutoScientist, a tool that automates the fine-tuning process to help AI models adapt to specific…
AWS Maps Foundation Model Scaling Across Training, Post-Training, and Inference
AWS and collaborators have published a technical framework for understanding how foundation model training and…
AWS Details Verifiable Rewards Method for More Reliable LLM Training
AWS published a technical guide on reinforcement learning with verifiable rewards (RLVR), a method that addresses…
OpenAI Releases MRC Networking Protocol for AI Training Clusters
OpenAI has released MRC (Multipath Reliable Connection), a new networking protocol designed to improve resilience and…
NVIDIA, OpenAI, Microsoft Open MRC Protocol for AI Network Resilience
NVIDIA, Microsoft, and OpenAI have introduced Multipath Reliable Connection (MRC), an RDMA transport protocol that…
Faithful Reasoning Emerges from Multi-Move Training, Not Direct Prediction
Researchers studied how reasoning develops in language models across supervised fine-tuning and reinforcement learning…
How OpenAI's Personality Feature Unleashed the Goblins
OpenAI's GPT-5.5 model exhibited unexpected behavior where it became obsessed with discussing goblins, gremlins, and…
NanoKnow: Mapping How LLMs Encode Knowledge
Researchers have released NanoKnow, a benchmark dataset that maps questions from Natural Questions and SQuAD to whether…
Standard Intelligence Raises $75M for Computer Use Models
Standard Intelligence, a two-year-old startup, has raised $75 million at a $500 million valuation to develop computer…
Poly-DPO and ViPO: Scaling Visual Preference Optimization
Researchers introduced Poly-DPO, an algorithmic extension to preference optimization that adds a polynomial term to…
Scaling Multi-Anchor Embeddings to LLMs with 40x Compression
Researchers introduce Adaptive Dictionary Embeddings (ADE), a framework that scales multi-anchor word representations…
New Training Method Cuts Reasoning Model Costs for Enterprises
Researchers at JD.com and academic partners introduced RLSD (Reinforcement Learning with Verifiable Rewards with…
Vanguard's AI lesson: Data architecture beats model selection
Vanguard built a conversational AI system called Virtual Analyst to help financial analysts query complex datasets…
AI for BusinessTrending
Manufacturing Goes Simulation-First as OpenUSD Becomes the Standard
Manufacturing is shifting from real-world testing as the primary validation method to simulation-first workflows…
Autonomous AI Framework Outperforms Human Baselines on Architecture and Data Optimization
Researchers at SII-GAIR have developed ASI-EVOLVE, an autonomous framework that optimizes AI training data, model…
Data & TrainingTrending
DeepMind Researcher Silver Raises $1.1B for Human-Data-Free AI
David Silver, a prominent researcher from DeepMind, has founded Ineffable Intelligence and secured $1.1 billion in…
Unified NIDS Dataset Advances ML Detection and Synthetic Data Generation
Researchers have created a unified multi-modal network intrusion detection dataset by consolidating four major…
RAG Fine-Tuning Tradeoff: Precision Gains Hide 40% Retrieval Loss
Redis research reveals that fine-tuning RAG embedding models for compositional sensitivity, the ability to distinguish…
Data Infrastructure, Not Models, Is the AI Bottleneck
Enterprise AI adoption is being constrained not by model capability but by fragmented, ungoverned data infrastructure.…
New Multilingual Medical AI Benchmark Reveals Language and Vision Gaps
Researchers have developed EuropeMedQA, a multilingual and multimodal medical examination dataset drawn from official…
Web Video as Training Data for 3D Scene Understanding
Researchers demonstrate that unlabeled internet videos can be automatically processed into training data for 3D scene…
Climate Foundation Models Falter on No-Analog Futures
Researchers benchmarked three machine learning climate models, including the ClimaX foundation model, to assess their…
Data Context, Not Compute, Is AI's Real Bottleneck
As AI moves from experimentation into core business workflows, organizations are discovering that model performance and…
Data & TrainingTrending
Meta to Track Employee Activity for AI Agent Training
Meta will begin tracking US employee mouse movements, keyboard inputs, and periodic screenshots across work-related…
High-Frequency Data Gap Exposes TSFM Limitations
Researchers have released a new millisecond-resolution dataset capturing real-world 5G wireless and traffic conditions…
NVIDIA Releases Fast Multilingual OCR Model Trained on Synthetic Data
NVIDIA released Nemotron OCR v2, a multilingual optical character recognition model trained on 12 million synthetic…