The central challenge of working with AI
The central challenge of working with AI is speed vs. quality. Work too slow and your...
Tutorials
The Cost of Self-Hosting Open-Source LLMs: GPUs, Staffing, and Operations
Self-hosting an open-source LLM in production costs roughly $150,000 to $2 million or more per year...
Human Preference Tuning for Small VLMs: SmolVLM2 + DPO in Transformers 5.12.1
A modern guide to preference-tuning SmolVLM2 with Direct Preference Optimization, TRL, PEFT LoRA adapters, and the current Transformers image-text API.
The OpenAI Agents API Makes the Harness a Managed Product
TL;DR OpenAI launched the Agents API in public beta on September 10, 2026: the same...
1,137 stars, 127 forks, 2,000+ tests. What 8 months of open-source GEO tooling taught me about what AI actually reads
I shipped the first version of GEO Optimizer on February 18, 2026. Eight months later it is at...
I tried "weight-space programming" on 6 kinds of models. Here's where it works and where it breaks.
A few days ago, MIT CSAIL researchers published Sparks of Weight-Space Programming: a frontier model...
I built a month-end close room where AI agents investigate and only a human can approve (with Switch and Discord)
Month-end close review is slow. Variances, missing commentary and bad data get chased across...
EU Finds TikTok’s Addictive Design May Breach the DSA, Signaling Marketing Changes
The European Commission has preliminarily found that TikTok’s service design may breach the Digital...
Your AI told the customer they were cancelled. The subscription renewed 11 days later.
A dental practice software vendor — 9 people, about 4,200 subscribers on monthly plans — found this...
Flutter app with a Laravel backend (2026 guide)
Short answer: install Laravel Sanctum, give the app a handful of JSON routes under /api, and have...
AI Isn't Killing Junior Developers. It's Breaking the Learning Path.
AI is automating tasks juniors used to learn from. New hiring and QA research shows why the real challenge is rebuilding the path from beginner to experienced engineer.
Stop hand-timing captions: word-by-word Remotion captions from ElevenLabs timestamps
A clip from our launch film. Same script, same voice. Left: what hand-timed captions tend to look...
I built a private, offline AI chat you can run with one command: npx unmuzzled
A local LLM chat app built on llama.cpp, Node and a hand-built web UI, plus what it took to make localhost secure and ship the same model on Android.
I benchmarked 8 models on our own code reviews. The first bug was ours.
Claude Opus 5.5 failed the same review three runs out of three. Same pull request, same reviewer...
Cybersecurity in the AI Era
Scams are becoming more convincing. AI tools are gaining more access. A practical guide to trust,...
The Only Coding-Agent Benchmark That Matters Is Your Own Revert Rate
Leaderboards score a model plus a harness on someone else's tasks. Tag landed commits with Agent: trailers and count reverts, fixes and red CI per agent.
AutoMode: a Claude Code mod that picks the model and effort for every prompt
If you use Claude Code every day, you know the routine. You ask a quick question on Opus because...
Breached during App Review: what a solo developer checked when the hosting provider was hit
I want to share what happened during one week in October, because I think it can help other solo...
Floor plans that compile: byte-identical output, errors as data, and an access graph
Floor plans that compile I compiled the same floor plan twice and diffed the...
We Built a CV Scanner That Compares Skills Against Real Jobs. Then Production Exposed a Bottleneck.
Building an AI-powered application is one thing. Getting it to behave reliably when someone actually...
I built a security-report desk that keeps vulnerabilities under embargo, and the chat platform enforces it
Open-source maintainers are drowning. One well-funded project, OpenSSL, went from about nine...
Building a 2,135-Project Static Showcase (No Database, No Framework, No Regrets)
Building a 2,135-Project Static Showcase (No Database, No Framework, No Regrets) How...
tmux vs Zellij vs Hermes IDE: Five CLI Agents, and Which One Is Waiting on You
Five Claude Code and Codex sessions, one blocked on you. Real tmux and Zellij configs, what Hermes IDE adds, and where a multiplexer still wins.
How AI in software development changed my views on the value and dangers of automation
When I began learning programming, some 20 years ago, I was excited about all the possibilities of...