~/writing
Writing
Essays and explorations on AI, Internet Services, and Customer Experience Design. Filter by tag, or scroll the timeline.
Dumb Won
AI is moving to the edges, so the middle needs to get smarter and cost more. Not really.
Credits Due
I recently posted arguing that telecom's periodic panic about new technology is orbital, and that the outcome was settled by network architecture decades ago. Some was kind enough to call it insightful. It is, but the insights aren't mine. I'm a settler on land that was mapped a generation ago. Here are the maps and these are the ideas that make our Internet work
One day, a dinosaur took a taxi to the observatory...
Once you strip away everything already routed around at the edge, a telco's balance sheet is government paper and civil works. So why does the industry keep funding strategy engagements about capturing value from technologies it didn't build? Telcos don't hire consultants to learn where the value went; they hire them to be told slowly.
Gibson on AI
William Gibson saw it all way before we did - from the trivial use cases we've embraced, to the need for a kill switch.
The Future is for Everyone
Zuckerberg's superintelligence manifesto is a stirring read about a company that never existed. I rewrote it with adjustments that match Meta's actual track record, plus 57 supporting footnotes. Beacon, Cambridge Analytica, hospital tracking pixels: the pattern isn't subtle.
Exploring Semantic MCP
I built an MCP server with zero API keys, zero databases, and zero state. A static file and five functions. It works better than it should.
Agile was built for a slower machine
AI tooling is making software teams faster than ever - but the processes they're running were designed for when speed was the problem. What happens when building is cheap and the bottleneck moves somewhere else?
Triggering Claude Chrome from the Command Line
PSA for anyone dealing with SaaS tools that have good reporting UIs but garbage APIs: Claude Chrome + Claude Code might be your workaround.
Claude Code Skills vs. Spawned Subagents
Claude Code skills and spawned subagents solve the same problem—selective context loading, but for different architectures. One works inside a session, the other across processes.
Yes, AI Sucks. That Makes It The Best Time to Learn
Today’s models fail loudly and visibly. That makes their limits easy to study and their mistakes easy to catch. Future systems will not be so forgiving.
Multi Model Deep Research: Not Quite What I Expected
I tried to see if I could get better results from combining the effort of three AI assistants, but the results weren't all that satisfying. But I learned a few things.
Day 1
A personal site feels right again, so I rebuilt one with a simple GitHub and Netlify workflow and a bit of nostalgia for how the web used to work. If you’re curious how it came together or why I bothered, read on...