The archive

AI Research & Insights

Daily digests of what's actually happening in AI — from research breakthroughs to new model releases, minus the hype.

Showing 1–12 of 906 articles

Page 1 of 76

AI Agent Engineering

AI Agents Research-to-Code: Build Skills Without Fine-Tuning

Use an AI agents research-to-code workflow to turn papers into testable agent skills with clear specs and evaluation, without changing model weights.

Oct 23, 20266 min read

AI Workflow Tools
OpenAI Decisions API: Faster Workflow Decisions at GA
OpenAI Decisions API is now generally available, adding voice and visual inputs. See Speak's reported speed gains and learn how to test it in your workflow.
Oct 23, 20263 min read
AI Developer Tools
Google TorchTPU: Run PyTorch on TPUs With Less Friction
Google TorchTPU aims to ease PyTorch migration to Google TPUs. See what the open-source project, migration agents, and partners mean for developers.
Oct 23, 20263 min read
AI Agents
Claude Duet: Anthropic's Proactive AI Agent for Work
Claude Duet is Anthropic's proactive AI agent for desktop and mobile. See how its app integrations, background tasks and approval safeguards work.
Oct 22, 20263 min read
AI Genomics
AI Genomics Limitations: Why DNA Still Defies Prediction
AI genomics limitations stem from context, regulation, and data gaps. Learn what genomic models can predict - and how to validate their results.
Oct 22, 20266 min read
AI Coding Models
Recursive Flint AI Coding Model: Laptop Speed and Claims
Explore the Recursive Flint AI coding model, its reported laptop speed, MoE design and AI-assisted development claims - plus what remains unverified.
Oct 22, 20263 min read
AI Agent Safety
AI Agent Rate Limiting: Prevent API Overload and Runaway Traffic
AI agent rate limiting, concurrency caps, and sandboxing can prevent API overload. Learn practical controls for safer, more reliable automation.
Oct 21, 20266 min read
AI Infrastructure
AWS TorchNeuron: Move PyTorch Apps to Trainium More Easily
AWS TorchNeuron aims to simplify moving PyTorch apps to Trainium. Learn what developers should test before choosing AWS AI chips.
Oct 21, 20263 min read
Enterprise AI Security
Microsoft Guardian Agents: How They Secure Enterprise AI
Microsoft Guardian Agents help enterprises monitor AI behavior, discover agent access, and control risky actions before they expose sensitive systems.
Oct 21, 20263 min read
Weather AI Forecasting
AI Weather Forecasting: How Bias Correction Improves Models
AI weather forecasting can correct recurring model errors. See how Microsoft's PBC improved ECMWF forecast accuracy and what the results mean.
Oct 20, 20263 min read
AI Agent Access
AI Agents Blocked by Websites: What Yelp's Data Licenses Mean
See why websites are blocking AI agents, how Yelp's data licenses work, and what partnerships mean for safer AI-powered shopping.
Oct 20, 20263 min read
Multi-Agent AI
AI Reasoning Repair Framework: How SAT Teams Solve More
See how the AI reasoning repair framework helps agents audit and fix one another's work, and what SAT's benchmark results reveal.
Oct 20, 20263 min read

AI Agents Research-to-Code: Build Skills Without Fine-Tuning

OpenAI Decisions API: Faster Workflow Decisions at GA

Google TorchTPU: Run PyTorch on TPUs With Less Friction

Claude Duet: Anthropic's Proactive AI Agent for Work

AI Genomics Limitations: Why DNA Still Defies Prediction

Recursive Flint AI Coding Model: Laptop Speed and Claims

AI Agent Rate Limiting: Prevent API Overload and Runaway Traffic

AWS TorchNeuron: Move PyTorch Apps to Trainium More Easily

Microsoft Guardian Agents: How They Secure Enterprise AI

AI Weather Forecasting: How Bias Correction Improves Models

AI Agents Blocked by Websites: What Yelp's Data Licenses Mean

AI Reasoning Repair Framework: How SAT Teams Solve More