I Ran a Real Build Inside a Sprite. Here's What Survives
Before you give an AI agent a remote computer, test the boring parts: builds, sleep/wake, checkpoints, Docker, credentials, and exit codes.
10 min read
Playbooks
Long-form walkthroughs of automation stacks that work, and the parts that quietly don't.
Parallel Coding Agents in 2026: Seven Tools, One Bottleneck
Seven tools for running parallel coding agents, from worktree managers to cloud swarms — and the review bottleneck that decides whether you should.
19 min read
Airtable Alternatives in 2026: What Actually Leaves With You
Bending Spoons now owns Airtable. Before you shop alternatives, run the exit test — most of what you built doesn't leave with you.
12 min read
Your Automation Said It Worked. Sixteen Records Never Existed.
A tool reporting success is telling you about a message it received, not a row that exists. Why agents make that gap dangerous, and the check that closes it.
8 min read
HIPAA-Compliant AI: Your Vendor's BAA Probably Doesn't Cover What You Think
Vendors say 'HIPAA compliant.' Compliance is a property of your whole chain, not any one vendor — and the boundary runs inside each product too.
11 min read
What "MCP" Actually Means for Your Business
You keep hearing "MCP." It's the standard that lets your AI assistant do things in your other apps instead of just talking about them. Here's whether an operator should care yet, and where to start safely.
9 min read
Front-End Your Airtable and Stop Paying Per Seat
Keep Airtable as your backend, put a no-code app in front, and pull viewers off billable editor seats. The full build behind 'don't switch — front-end it.'
7 min read
Run a CLI Agent in a Disposable Sandbox (with Slack as the Front Door)
Let an AI coding agent do real work without risking your laptop: run it inside a throwaway, checkpoint-able sandbox and drive it from Slack.
9 min read