Quits: a group tab that remembers who doesn't drink
A shared expense tab that learns a group's unwritten rules into Walrus Memory and settles debts in USDC on Sui through Tatum, with x402 payments and an agent bound by a Sui Allowan...
Tutorials
Why Full Transparency Freezes Multi-Agent Markets
A Stanford and UMD study reveals that handing frontier LLMs complete information triggers aggressive rent extraction in double auctions, cutting realized market welfare by up to th...
I can't afford a smartwatch, so I built a 75-second check with just my phone camera
This is a submission for the Presage SmartSpectra SDK Challenge What I Built I can't...
Stop Swapping Models: Optimizing the Agent Environment Cuts 70% More Cost
When an agent workflow runs up an unsustainable token bill, developers usually pull two levers. The...
Why 99% 'Verified' AI Code Fails: The 40x Vacuity Gap in Formal LLM Pipelines
Every engineering team deploying autonomous coding agents eventually runs into the same economic...
AI coding goes from generation to verification
ICYMI AI: The latest AI trends, delivered to Dev.to Tech Talk: Can (and should) AI verify...
I wrote a 1,669-line kernel that boots a mouse. The ISO goes public next Friday.
1,669 lines, 563 KB, mouse and clock included. ISO next Friday.
AI Visibility Metrics Need Repeat Sampling Before They Guide Decisions
AI visibility is becoming a regular reporting line for brands and agencies, but a single prompt run...
Free Tokens Have a Break-Even Failure Rate
Free Tokens Have a Break-Even Failure Rate Free capacity is cheaper right up until it...
I Ran DeepSeek's dsh-superagent End to End: A Hands-On Deep Dive
Originally published at GeekCoding101. Why Everyone Is Talking About It On October 31 the...
One line of JSON stops the pump
How I gave a Claude Code agent ears, a seat in the plant's Discord, and a supervisor to answer to, with Switch.
An MCP server in pure Python, under 100 lines, where tools can only propose
Many MCP tutorials start with a framework install and end with a tool that adds two numbers. This one...
FORCE_HYPERLINK=false Turned Claude Code's Links On. Inside OSC 8 Terminal Hyperlinks
Before Claude Code 2.1.314, FORCE_HYPERLINK=false enabled links. Why a string that parses to NaN meant yes, and how OSC 8 links work byte by byte.
Floating Versions Are an Open Door: Pin Your Actions and Dependencies
The problem Look at almost any CI workflow and you'll find a line like this: - uses:...
Portability Before Spill: A Two-Path Conformance Harness for Local-First Agents
A small team ships an agent that reads incident notes and emits a structured triage record. On a...
Four Hidden Config Settings That Stop Token Leaks and Unapproved Writes in OpenAI Codex and Claude Code
When you run OpenAI Codex CLI and Claude Code side by side across a real codebase, the biggest token...
AI has revealed the pride divide
So, I'm someone who takes great pride in quality work. Always have, always will. I use AI to help...
Containing the Autonomous Mind: Sub-Second Zero-Trust Execution with GKE Agent Sandbox
Why standard Kubernetes pods fail when AI agents start writing code and how to build a sub-second,...
I gave Claude Code 10 translation tools. One run in three, it ignored them.
I asked Claude Code to translate the same 92 missing strings nine times. Three times it had only the...
Claude Code email in User-Agent headers: how it happened and what the fix covers
Claude Code put the account email in the model's context, and the model used it as a User-Agent contact. What the privacy mod checks and what it misses.
10 AI Terms Every Developer Should Know to Build Better Systems
In my last speech, I wanted the audience to leave with something they could use beyond a particular...
What I got wrong estimating LLM costs before the request
Ever since AI has taken over the internet, I've been wondering how we can optimize the full ability...
Building a multi-pane workspace for coding agents: PTYs, replay, and the backpressure we have not solved
SpaceApp is a self-hosted workspace that runs AI coding CLIs side by side in the browser, each in its...
Your agent can rewrite its own instructions. I built a pre-run lock that proves it didn't.
Your agent reads its skills from plain markdown files. It can also edit them. Nothing in the format...