Articles
In-depth decodes of AI developments, every claim linked to its source.
- 'You don't need long context': what the compaction paper means for coding agents
- Red-teamers found 11 ways a misaligned agent could get past Codex Auto-review
- Decoding Anthropic's Skill authoring guide: lean SKILL.md files that actually trigger
- Claude Code mods: in-session JavaScript hooks for guards, readouts and review views
- Securing MCP beyond the gateway: InfoQ's four control layers, decoded
- MiniCPM-OCR pairs a 1.3B parser with a 2B model that flags OCR you shouldn't trust
- Decoding Zalando's risk-based PR auto-approval and LiteLLM proxy setup
- Cloudflare Computer v0.7.0: replay after crashes, pause for approval mid-run
- Apple's 20B on-device Mac model: about 2 GiB of RAM, weak at math, scriptable
- Subagents vs. skills: stop using subagents to hold instructions
- Plain RAG over raw chat history beats agentic memory layers in a new benchmark study
- Google's agent-eval playbook: replacing vibe checks with trace-based regression gates
- Claude Code mods: programmable guardrails inside the agent harness
- Vercel's durable-harness: crash-resume for Claude Code, Codex, Pi and OpenCode
- Meta's thin orchestrator: a deterministic router with 14 specialist agents, decoded
- How the Claude Code team reviews agent-written code, and what you can copy
- Agentic verification: give your coding agent checks it can run itself
- How Anthropic's Claude Code team reviews agent-written code, and what to copy
- Muse Spark Open decoded: Meta's 816B open-weight coding model, hosted or on Macs
- Behavioral harnesses made a cheap coding model reliable, and a frontier one pricier
- How GitHub ported Copilot's runtime to Rust with agents and kept shipping
- Agent teams on Vibe Code Bench: 1.8–5.1x the cost, one significant score gain
- Microsoft's Quicksand puts local coding agents inside a pip-installable QEMU VM
- Microsoft's MXC: one SDK to sandbox agent-run code on Windows, macOS and Linux
- AutoRubric: a Python library for rubric-based LLM judges, with a cheap Jev first pass
- OpenAI's Decisions API: typed probabilities from gpt-6-luna, now in an LLM plugin
- Decoding Uber's AI software factory: what a five-person team should copy
- SelfSearch: a coding agent rewrites its own harness, claims Codex parity for $4.03
- Study: cutting coding-agent tokens with compaction can make runs 20–80% slower
- Domenic Denicola's agentic coding setup: no-approval agents in a disposable Linux VM
- Cloudflare's Clef decision models: where they beat an LLM call, and where they don't
- Cloudflare's Clef: open-weight decision models for typed classification steps
- Claude Code mods: in-session hooks that can rewrite tool calls and draw UI
- NVIDIA OpenShell: kernel-enforced agent sandboxes with verified policy changes
- Anthropic's measure-first agent loop: 3,000+ merges and a 3x faster claude.ai
- ChatGPT's MCP Events: webhook triggers from your own MCP server, with real gaps
- Claude Code's build-eval and hillclimb: eval design with a held-out check
- DAIR.AI tutorial puts Jev on an agent harness's routing, tool-gating and done checks
- Personal AI agents upsell users they infer are wealthy, study finds
- Google's AI Control Roadmap Lands as Agent Stacks Multiply
- Grok Lands on Bedrock While the Inference Stack Quietly Reshuffles
- The Coding Agent Market Splits Into Three Camps
- OpenAI Pours Resources Into Science While Google Sets Agent-Safety Ground Rules
- Agent Safety Goes Mainstream as Google, OpenAI, and xAI Race to Ship Autonomous Workflows
- Grok Lands on Bedrock as the Inference Layer Keeps Consolidating
- The Vendors Stop Selling Models, Start Selling Lab Notebooks
- Agents Get Hands: Robots, Git, and the Push Into the Physical World
- Codex Goes Loose in the EU While OpenAI Tries to See Around Corners
- Vendor Watch: Anthropic, OpenAI, and Mistral Each Reveal How They Want to Be Measured
- Agents Grow Up: From Subroutines to Subagents, Voice, and Robots
- The Vibecoding Stack Has a Feedback Problem — and Everyone Is Trying to Fix It
- Export Controls Hit Anthropic's Top Models — and the Practitioner Fallout Is Already Here
- Agents Get Proactive, And That's the New Problem
- Vector Search Goes Mainstream: When Oracle, MongoDB, and Grok All Want a Piece of Your RAG Stack
- Fable's 34-Day Sprint and the Quiet Reframing of Coding Benchmarks
- Anthropic Walks Back Invisible Refusals — And Why That Matters for Every AI Builder
- Anthropic Reverses Course on Hidden Refusals as Fable 5 Lands
- The Assignment Model Arrives: Agents Move From Sidekick to Subcontractor
- Coding Agents Grow Up: From Pair Programmer to Pull Request
- The Assignment Model Comes for Your IDE
- Your IDE Just Became a Task Queue
- Agents Cross the Sandbox: Embodied Wins, Enterprise Caveats, and Real Money on the Line
- Enterprise Stipends Hit $2,000/Month as AI Coding Tools Sprawl Past One Per Developer
- Anthropic Says Claude Is Now Writing Claude's Successor — And Engineers Are Negotiating $2,000 Tool Budgets
- Anthropic Says Claude Now Writes 80% of Its Code. The Speedup Chart Is the Real Story.
- Coding Agents Hit the Corporate Expense Account
- Agent Budgets, Bench Scientists, and Always-On Bots: The Week Agents Got Operational
- Apple's Nvidia Order Reveals the Inference Stack Splitting in Two
- When AI Attackers Meet Old Defense Playbooks
- Codex Becomes a Platform, Anthropic Becomes a Weapon, and Vendor Lines Blur
- Uber Puts a $1,500 Ceiling on Coding Agents — and Reveals What They're Actually Worth
- Multimodal AI Grows Up: From Image Chat to Lab Partner
- Agents Move From Demo to Default
- When Your Coding Agent's Transcript Vanishes in an Update
- Agents Grow Up: From Personal Sidekicks to Autonomous Chip Designers
- When Your Coding Agent Forgets: Simon Willison's Codex Complaint Is a Bigger Deal Than It Looks
- Mistral Collapses Reasoning, Vision, and Coding Into One Model
- The EU's Quiet Retreat on AI Machinery, and What It Tells Builders About Trust
- Mistral's Vibe Bet, Codex's Vanishing Feature, and the Vendor Lock-In Tax
- Agentic Coding Models Get Cheaper, Fatter, and More Distributed
- Anthropic's $965B Valuation Lands the Same Week Its Cheapest Model Gets Outpriced
- Agentic Coding Goes Commodity: xAI Undercuts, Mistral Rebrands, Copilot Expands
- Coding agents go multi-vendor: Grok, Opus 4.8, and Qwen3.7-Max crowd the field in a single week
- The EU AI Act's Quiet Reprieve, and Why Model Choice Is Now a Trust Decision
- Anthropic's $965B Valuation Lands the Same Week Its Rivals Quietly Reshape the Stack
- The reasoning-effort dial gets a reality check
- Agentic Coding's Land-Grab Week: Four New Models and a $1B Round
- Coding agents are leaving the IDE — and the new question is control
- The Quiet Pivot: AI's 2026 Story Is Cheaper Inference, Not Bigger Models
- RAG Is a Hardware Problem in Disguise
- 3 Companies, 67% of AI's Funding, and the End of Cheap Inference
- Coding Agents Don't Need to Be Smarter. They Need to Be Fenced In.
- The Model Benchmark That Counts Now Is Cost Per Query
- Where Your Coding Agent Actually Lives Now
- 3 Labs Took 67% of AI's Funding. Here's What It Changes.
- Built for Agents, Not Bolted On
- Coding Tools Are Going Agentic. Microsoft's Naming Is Going Sideways.
- The Week Coding Agents Moved Into Your Browser
- Every Big Vendor Just Made the Same Bet: Agents That Run for Days
- The Agent Fight Just Moved Off the Model
- AI Coding Tools Aren't Converging — They're Picking Lanes
- Codex Ships Weekly, Gemini Previews, Claude Code Holds the Terminal
- LeCun's Inconvenient Point: Useful Isn't the Same as Intelligent
- Forget the Model: the Agent Harness Is the Real Product Now
- AI Coding Tools Now Assume You've Walked Away
- Google Goes Wide, OpenAI Goes Deep, Anthropic Goes Hiring
- AI's Spending Spree Meets a Capability Reality Check
- Agents Quietly Cross the Vendor Line
- The AI Coding Tools Race Is Now About Autonomy, Not Brand
- The AI Vendors Stopped Racing on the Same Track
- The agent framework is becoming the new lock-in
- AI Coding Tools Are Done Fighting Over Autocomplete
- RAG First, Fine-Tune Last: The Boring Answer Wins Again
- Multimodal AI in 2026: Pick Your Tier, Not Your Vendor
- The Week the AI Vendors Picked Different Lanes
- Codex Goes Mobile, Grok Goes Terminal: The New Map of AI Coding
- RAG settled the fine-tuning debate. Agentic RAG is the new argument.
- Five Vendors, Five Bets: AI's Strategic Map in Mid-May
- The Week Agent SDKs Became Table Stakes
- Inference economics is the new model race
- Three Vendor Bets Against the Chat Box
- Three Platforms Stake Claims on the Agentic Runtime
- Coding Tools Already Solved the Prompt Problem
- Why your next RAG system might not have a vector database
- The major labs are no longer playing the same game
- The Autonomous Enterprise Meets the 30GB Memory Leak
- AI Coding Tools Just Split Into 3 Tiers
- RAG Is Coming Back — and the Catch Is Latency
- Three Vendors, Three Bets: AI's Strategic Divergence
- The Autonomous Enterprise Arrives Before Its Operating Manual
- Three shapes, one protocol, and Claude Code's 30GB memory bill
- OpenAI's Two-Front Enterprise Pivot
- Agent Sprawl Becomes the Real Problem
- OpenAI Builds a Deployment Arm — And Quietly Becomes a Consulting Firm
- The Week Agent Safety Stopped Being Commentary
- Qwen 3.5 Lands, and Alibaba Wires It Into Taobao
- Voice Agents Grow Up, and Their Makers Open Up
- Inference Becomes the AI Stack's New Center of Gravity
- Codex Goes Browser-Wide as Anthropic Catches Claude Cheating on Code
- Physical AI follows the LLM playbook
- Voice agents grow up, and the money knows it
- Codex Moves Into the Browser as Anthropic's Top Coder Gets Caught Cheating
- Anthropic Borrowed xAI's Compute. That's the Real Story This Week.
- The Week Agentic AI Stopped Looking Like a Demo
- The AI coding stack just split in two
- When AI Rivals Become Each Other's Landlords
- AI's two-tier moment: trillion-dollar labs, fungible APIs
- Agents grow up: alignment, capital, and benchmarks converge
- The benchmark that matters now isn't the model — it's where you run it
- Repos for robots: when your codebase has an audience of agents
- RAG Grows Up: Why Vector-Only Pipelines Just Became Legacy
- OpenAI Ships at Both Layers, Anthropic Goes Deeper
- $950M Says Agentic AI Is Infrastructure Now. Open Source Is Ready to Prove It.
- The Inference Economy: Why Serving AI Models Has Become More Important Than Training Them
- AI Security Has Two Clocks Running: Developer Intuition and the EU Compliance Deadline
- Three Vendor Signals That Should Change Your AI Stack Calculus
- When the model is free, the chip is the moat
- Your Next AI Coder Doesn't Live On Your Laptop
- Mistral Climbs the Stack, Google Quietly Backs MCP
- Codex Doubled Revenue in 7 Days. The Agentic Coding Race Just Got Real.
- Consumer AI's Brutal Math: Why 30,000 Users Isn't Enough
- Coding Agents Break Containment: The Agentic Playbook Spreads
- When Cheaper Beats Smarter: Inference Economics Take the Front Seat
- Coding Agents Have Stopped Being About Code
- RAG's Boring Era: Why Stack Choices Got Easier and Decisions Got Harder
- Multimodal AI moves from demo to clinic
- AI Trust Just Became an Architecture Decision
- Meta walks away from open weights as the vendor race accelerates
- Agents Step Into the Room: Healthcare Trials, Remote Coding, and the Governance Gap
- Coding Agents Are Escaping the Developer Sandbox
- Multimodal AI Converges: Unified Models, Clinical Stakes, and a Market Shakeout
- The Trust Stack: AI Security Now Runs From the Commit to the Model Itself
- The Agent Stack Just Got Crowded
- RAG's Bottleneck Just Moved — And It's Not the Model
- Nvidia and Mistral bet that one model beats five
- AI trust isn't one problem anymore
- The Week Agentic AI Infrastructure Went Production-Ready
- The Inference Playbook: Why How You Run AI Matters as Much as Which Model You Pick
- Copilot Goes Usage-Based: What the AI Coding Tool Shakeout Means for Your Team
- The Agentic Stack Hits Production — And Reveals Its Blast Radius
- Why 'Which AI IDE?' Is the Wrong Question in 2026
- The Vendor Map Hardens: Mistral, OpenAI, and DeepSeek Stake Their Ground
- Agents Are Shipping Fast. Is Your Safety Architecture Keeping Up?
- The Inference Cost Squeeze Is Hitting From Every Direction
- DeepSeek Cuts Coding API Prices 75% — But Your IDE Might Not Be Ready
- RAG Is Becoming Infrastructure. The Hard Part Is What Comes After.
- 90% Off and $600M In: The Week Enterprise AI's Economics Shifted
- Coding Models Multiply, Benchmarks Stop Settling It
- GPT-5.5: Built for Agents, Optimized for Completion
- The Coding AI Race Just Compressed Into 24 Hours
- The Vendors Have All Voted: Agentic AI Is the Product Now
- Google and OpenAI both ship agent platforms — same week, same bet
- GPT-5.5 Arrives With Bigger Context, Lower Prices, and a Clear Bet on Agents
- OpenAI and DeepSeek Ship Coding Agents Hours Apart
- The AI frontier splits 3 ways
- Google and OpenAI Shipped Enterprise Agent Platforms on the Same Day. That's Not a Coincidence.
- AI Governance Graduates From Checkbox to Platform
- Google Deep Research adopts MCP — and the agent standard arrives
- The Big Three's Tuesday: 5 Gigawatts, MCP-Native Agents, and Thinking Images
- LeCun's AMI Labs Lands a Record Seed — and Picks a Fight With the LLM Consensus
- When the library isn't in the training data, point the agent at the source
- Retrieval Is the Floor, Not the Ceiling
- Cursor's $50B valuation talks reveal the shape of AI coding's next phase
- Vector databases are commodity now. Leverage moved up the stack.
- AI's security conversation grows up
- The week AI left the chat window
- Cursor's $50B Moment Is the Headline. The Real Story Is What AI Coding Tools Can Do Now.
- Parallel Agents Go Mainstream: What This Week's Releases Signal About AI's Next Phase
- The Inference Gap: Why Where Models Run Matters as Much as How They Score
- Parallel Agents, Computer Use, and the End of the AI Coding Editor
- Your Database Already Does Vector Search. Did You Notice?
- Multimodal AI Is Now the Baseline, Not the Feature
- AI Security's Double Edge: Better Defenders, Hidden Risks in the Weights
- The Race to Own Your Coding Session
- Trust Is the New Benchmark
- Meta's Muse Spark and the Quiet Death of Open-Source Frontier AI
- Anthropic's Glasswing and OpenAI's $100 Tier Reshape the Vendor Map
- Safety Gets Shipped: Google's ShieldGemma and the EU's Compliance Clock
- RAG Grows Up: From Clever Hack to Enterprise Backbone
- The year the coding agent grew up — and started breaking things
- The inference wars move to the silicon layer
- The Agent Stack Grows Up — And Starts Shipping
- The AI Industry's Quiet Consolidation: Who's Building Infrastructure While Everyone Watches the Models
- The week every major AI lab redrew its battle lines
- RAG Is Becoming the Default Architecture — Here's What That Actually Means
- The IDE Is Dead. Long Live the Agent.
- The Inference Wars Have Moved to Your Laptop
- Every AI lab shipped an agent this week. Here's what that actually means.
- The Multimodal Land Grab Is On — and Meta Just Changed the Rules
- The Week AI Got Real: Code Overload, Departing Founders, and the Quiet Death of Sora
- OpenAI Raises $122 Billion While the Vibes Get Weird
- Multimodal AI Just Became the Default — Here's What That Means for You
- RAG Isn't Dead — It's Growing Up
- The Real AI Race Isn't About Models Anymore — It's About Making Them Run
- The coding agent wars just went nuclear — and your IDE is the battlefield
- The Code Flood Is Here, Benchmarks Are Broken, and Nobody Knows How to Measure What Matters
- The Coding Agent War Is Now a Three-Front Battle — and OpenAI Is Playing Defense
- The Multimodal Land Grab: Every Major AI Lab Now Speaks Every Language
- RAG Just Became Boring — And That's the Best Thing That Could Happen
- The Coding Agent Wars Are Here — and They're Reshaping How All Software Gets Built
- AI Industry Roundup: Funding, Research, and the Infrastructure Build-Out
- Vendor Watch: Anthropic, OpenAI, Google, and Mistral Jockey for Position
- Multimodal AI Expands: Vision, Voice, and Video Converge
- RAG Evolves: From Simple Retrieval to Intelligent Knowledge Systems
- AI Coding Tools Race Heats Up: Agentic IDEs Go Mainstream
- Model Updates: New Releases Push Reasoning and Efficiency
- Agentic AI Matures: From Demos to Production Workflows
bluemat home