I Handed My Business to Four Different AIs Today. Two of Them Invented Facts. Here's the System That Caught Them.
Seller disclosure, up front: I sell a template kit version of the system described below — The...
Tutorials
A Cloud Agent's 'All Tests Pass' Only Covers the Services It Could Reach
Copilot, Codex and Claude Code's cloud agents test in vendor sandboxes. Their own docs list what they can reach. Your VPN isn't on it.
I built an HR bot that falls for every trick. It still can't leak the CEO's salary.
I built an HR chatbot that falls for every trick in the book. Tell it you're the CEO: it believes...
I had an AI agent port Karpathy's GPT-2 trainer from CUDA to SYCL, and made it prove every step
Muse Port plans and ports real CUDA projects to the open SYCL standard. Nothing counts as done until a check the model never sees says the output matches.
Which local model should a coding agent run? Round 2: llama.cpp 0.9
Some days ago I published the first round: one harness, seven GGUFs under llama.cpp plus one model...
AI Inference Benchmark: Keln vs Fireworks, Together AI, Baseten, Modal, DeepInfra & Z.ai
We ran the same agent sessions across seven inference platforms to see how they compare on speed,...
A daily Arabic health Short with Claude Code, Hugging Face ZeroGPU and three AI reviewers
I run Hakeemly, a small Arabic health-information site. Since this week it publishes one Arabic...
I built an open-source AI interface builder that exports to React, Vue, Svelte… and Roblox
Drag components onto a canvas, ask any AI model to build pages, and export to HTML, React, Vue, Svelte, Roblox Luau or LÖVE. Free and MIT-licensed.
Running LLMs on a CPU-Only Server: My Benchmark of Gemma, Qwen, Spark, and MiniCPM
We all know the default answer when someone asks about running local LLMs: "Get a GPU." Preferably an...
My regression detector alerted on 4.3 points. The model moves 8.57 between runs.
On 2026-09-23 my drift tracker reported Grok 4.3 down 4.3 points, to 91.4%. The pipeline drafted it automatically, as designed: a finding clears a threshold, a note gets written, a
Google Brings Gemini-Powered Business Profile Tools to Workspace
Google is integrating Gemini more directly into local business management. Its new Business Profile...
The Internet Feels Less Safe. The Evidence Is Mounting.
The Internet Feels Less Safe. The Evidence Is Mounting. For a long time, the internet’s insecurity...
4 money bugs that bite AI agents (and how I fixed them in PurseString)
I'm building PurseString, an open-source MCP server that lets AI agents spend money safely. A lot of...
Building Context Keeper on Switch
Someone had just come back after a few months, and they kept re-raising ideas I'd already put into...
I built a local AI coding agent for VS Code in 4 days — here's how it works
During Meta's Global AI Developer Hackathon, I spent roughly four intensive days building RAME Code,...
Dear agent, part two: the polite rule that leaked an email
Field note: a follow-up to Dear agent: the pages that ask nicely. Last week, I wrote about pages...
Why AI Writes Broken ESP32 Code (and How to Fix It for Arduino-ESP32 3.x)
You ask an AI assistant for a simple ESP32 sketch, it hands you clean, confident code, and the...
How to Test an AI Agent: Start With Error Analysis, Not Metrics
Your agent passed the demo. Does it still work on the 500th request? Agents fail differently from...
Why Your AI Coding Agent Bill Grows With Every Turn (and 7 Ways to Cut It)
If you use Claude Code, Cursor or Codex every day, you have probably had one of two surprises: an API...
Tensor vs. Pipeline vs. Data Parallelism: How LLM Inference Scales Across GPUs
Tensor, pipeline, and data parallelism solve three different problems: latency, fit, and throughput. How each one splits an LLM across GPUs, what it costs in communication, and how...
My AI agent called me at 2am. I picked up. It had already found the bug.
Sparks are background agents in Flash. Now they can phone you: in your browser, on your desktop, in your terminal, even with every tab closed. Here's how it works.
How I Made AI Characters Follow the Cursor
A practical look at the videos, frame maps, browser fixes, costs, and failed attempts behind cursor-following AI characters.
Workflows vs Agents: The 6 AI Agent Orchestration Patterns (and When to Go Multi-Agent)
"Agent orchestration" sounds exotic, but the question underneath it is plain: how do you arrange one...
Vibe-Coded App Security: A 30-Minute Pre-Launch Checklist
You built an app with Lovable, Bolt, Cursor or Claude Code. It works, people are signing up, and it...