AI — Page 32
Artificial intelligence, machine learning, LLMs, and AI tools transforming development.
Karpathy's Obsidian Setup Challenges RAG Orthodoxy
Andrej Karpathy's markdown-based knowledge system questions whether most developers actually need traditional RAG systems at all.
Claude Code 2.1.91: Three Updates That Actually Matter
Claude Code 2.1.91: Three Updates That Actually Matter
Claude Code's latest update brings shell execution controls, 500K character handling, and session reliability fixes. Here's what changed and why it matters.
AI Agents That Optimize Themselves While You Sleep
AI Agents That Optimize Themselves While You Sleep
Kevin Guo's AutoAgent extends Karpathy's auto-research loop to let AI agents rewrite their own operational code overnight. What happens when agents program agents?
Google's Gemma 4: Running Frontier AI on Your Phone
Google's Gemma 4: Running Frontier AI on Your Phone
Google's Gemma 4 brings frontier-level AI to consumer devices. Free, open-source, and offline-capable—but does it deliver on the promise?
Google's Agent Skills Update Just Fixed AI's Biggest Flaw
Google's Agent Skills Update Just Fixed AI's Biggest Flaw
Google's ADK now uses progressive disclosure to stop AI agents from loading unnecessary instructions. Here's why that matters for everyone using AI.
Inside Anthropic's Daily Claude Code Workflow
Inside Anthropic's Daily Claude Code Workflow
The tools Anthropic's team actually uses in Claude Code—from open-source plugins to internal skills reverse-engineered from leaked source code.
Anthropic's Claude Code Leak Reveals Unglamorous Truth
Anthropic's Claude Code Leak Reveals Unglamorous Truth
The Claude Code leak shows what actually makes AI agents work at scale: boring infrastructure, not flashy features. Two leaks in one week raise questions.
When Trusted Tools Betray: The Axios Hack and Trust Debt
When Trusted Tools Betray: The Axios Hack and Trust Debt
The Axios npm breach exposed how fragile our software trust chain is. Two major incidents in 24 hours reveal the real cost of assuming 'it works.'
Google's Gemma 4 Brings Powerful AI to Consumer Hardware
Google's Gemma 4 Brings Powerful AI to Consumer Hardware
Google released Gemma 4 under Apache 2.0 license. The open model runs on standard GPUs, challenging the assumption you need enterprise hardware for capable AI.
AI Agent Skills: The Markdown Files That Teach Once
AI Agent Skills: The Markdown Files That Teach Once
Skills are markdown files that give AI agents context on demand—solving the problem of repeating instructions without overloading context windows.
AI Agents Are Getting Persistent—And That Changes Everything
AI Agents Are Getting Persistent—And That Changes Everything
Anthropic's Conway, Z.ai's GLM-5V-Turbo, and Alibaba's Qwen 3.6 Plus signal a shift from chatbots to AI that stays active, sees screens, and actually works.
GitHub's Copilot SDK Turns Apps Into AI Planners
GitHub's Copilot SDK Turns Apps Into AI Planners
GitHub demonstrates how its Copilot SDK transforms static planning apps into dynamic AI assistants with minimal code. But what's the implementation cost?
Google's Gemma 4: Small Models, Big Performance Claims
Google's Gemma 4: Small Models, Big Performance Claims
Google releases Gemma 4, claiming frontier-level AI performance in models small enough for consumer hardware. The numbers look impressive. The questions remain.
Two Hidden Claude Code Commands That Actually Matter
Two Hidden Claude Code Commands That Actually Matter
Most Claude Code users ignore /power-up and /insights. Here's why these slash commands might be the productivity hack you didn't know you needed.
How to Build AI Agent Skills That Actually Work
How to Build AI Agent Skills That Actually Work
Nufar Gaspar's masterclass reveals the frameworks for creating effective AI agent skills—from trigger design to the critical 'gotcha' section most skip.
Anthropic's DMCA Mess: What Happens When 8,100 Repos Go Down
Anthropic's DMCA Mess: What Happens When 8,100 Repos Go Down
Developer Theo got DMCA'd by Anthropic for changing one word in a markdown file. The story reveals how DMCA enforcement can go catastrophically wrong.
Claude Code's Hidden Features That Actually Matter
Claude Code's Hidden Features That Actually Matter
Claude Code ships features faster than users can discover them. Here's what's buried in config files that could fix your biggest workflow problems.
Why AI Might Create More Jobs Than It Kills
Why AI Might Create More Jobs Than It Kills
A 160-year-old economic principle suggests AI efficiency won't eliminate jobs—it'll create demand for more (and different) human work.
Anthropic Accidentally Leaked Claude Code's Secret Agent
Anthropic Accidentally Leaked Claude Code's Secret Agent
A source map mishap revealed Kairos, Claude Code's unreleased background AI agent with memory consolidation, push notifications, and proactive coding help.
Replit Agent 4 Turns Napkin Sketches Into Real Apps
Replit Agent 4 Turns Napkin Sketches Into Real Apps
Replit's Agent 4 lets you build full-stack apps from text prompts. We tested the new design canvas and parallel agents to see what actually works.