ai-agents.
49 writings found
Latest Archives
Rewriting 800K Lines with AI Agents: What We Learned
GitHub rewrote the Copilot agent runtime from TypeScript to Rust using AI agents. Here's what that massive migration reveals about the future of large-scale rewrites.
Why Your AI Agent Works Once But Fails the Next Time
Consistency gaps in LLM agents matter more than average accuracy. Introducing consistency guidelines to stabilize agent decisions.
Why Your AI Agent Works Once but Fails Twice
AI agents achieve high average accuracy but fail inconsistently on identical tasks. A new consistency measurement and guideline system reveals why, and how to fix it.
ToolGrad Flips the Script on AI Agent Training
Google researchers reveal how generating tool-use solutions first, then prompts, creates better training data for LLMs with 100% pass rates and lower costs.
ToolGrad Flips AI Agent Training on Its Head
Google researchers show that generating tool-use solutions before prompts leads to cheaper, more reliable AI agent training datasets and better model performance.
ToolGrad Flips the Script on AI Agent Training
Google researchers reverse dataset generation: build tool-use chains first, then queries. The result? Cheaper, more reliable AI agent training with smaller models outperforming proprietary ones.
ToolGrad Flips the Script on AI Agent Training
Google researchers reveal how generating tool-use answers before queries cuts costs and improves LLM performance. What this means for the future of AI agents.
Why AI Coding Agents Should Target Java
Java's stability and massive ecosystem make it the ideal target for AI coding agents. Here's why that matters for your development future.
Why AI Agents Need Sandboxes, Not Just Permissions
Explore how sandbox environments contain AI agents safely, handle credentials securely, and let developers experiment without risking their machines.
YOLO Mode and AI Agents: Speed Without Risk
Why AI agents need sandboxes, not just permission prompts. How isolation changes the risk calculus for autonomous coding tools.
How OpenClaw Maintainers Are Rethinking Trust at Scale
OpenClaw's explosive growth to 388k stars forced maintainers to rebuild code review, security, and contributor trust from first principles. What we can learn.
ChatGPT Work: When AI Gets Real Internet Access
OpenAI's new Work mode gives Claude a powerful upgrade: real internet access, browser automation, and persistent file systems. Here's what it means.
Why Lines of Code Still Matter When AI Writes Your Software
AI coding agents are changing productivity metrics. We need new frameworks to measure what agents actually deliver, and what gets lost in the speed.
Proving AI Agents Work: Why Lean Language Matters Now
Leo de Moura on using Lean to verify AI agent correctness. How automated reasoning and probabilistic models converge to build trustworthy systems.
Why Lines of Code Still Matter for AI-Assisted Development
Exploring how coding agents change productivity metrics, conceptual integrity, and why team structure still matters in the age of AI.