WRITING
ALL POSTS
17 TOTAL- 01
Claude Desktop vs Cowork vs Code: Which One to Use
Three Claude surfaces run one engine, and Dispatch now routes work between them. What separates them is whether the setup shaping Claude's behavior is a file you own or state the app holds for you.
[READ] - 02
Agents Need an Identity Layer. Adtech Shows Why.
Machines have been transacting without a working identity layer for twenty years. Adtech patched the gap with vendors who grade the transaction afterward, and that patch cannot grade a fake person. The standard now being written for agents is that layer.
[READ] - 03
Trigger.dev vs Inngest: Which Is Better Depends on Where Your Code Runs
Both platforms give TypeScript teams durable background execution without hand-rolling queues. Inngest orchestrates and calls functions on infrastructure you already pay for. Trigger.dev runs the code itself, in its own container, billed by the second. Almost every other difference between them falls out of that one split, including the pricing, which is where most comparisons get it wrong.
[READ] - 04
Tooling That Shows Its Work
I've had 91 Trigger.dev tasks running my content and social pipelines. Most AI products won't show you what happened inside them, and the ones that do turn out to be the ones an AI can actually help you fix.
[READ] - 05
What Circle Actually Bought From IBM
Circle just bought nearly a thousand blockchain patents from IBM. The framing is USDC protection; the real story is trust infrastructure for an economy where machines transact and nobody's checking IDs.
[READ] - 06
Write Down the Decision, Not Just the Fact
A bug in a Claude Code Stop hook got fixed four times, diagnosed correctly every time, and survived all four. The mechanism, the source-level regression test that catches it, and why a bug that keeps coming back is usually a missing decision rather than a missing note.
[READ] - 07
Building a Commit Guard for AI Agents That an Adversary Can't Slip Past
I gave a coding agent a guard that checks every commit lands in the repo and branch I intend. My 417 tests were all green. Then I hired a second agent to break it, and it walked through two holes the green suite never thought to check.
[READ] - 08
Graph Engineering Makes Agent Failure Legible, Not Agents Reliable
A video making the rounds argues graph engineering supersedes the loop-based agent pattern. After building multi-agent systems, I think the loop-versus-graph framing asks the wrong question. Topology decides whether you can SEE a failure, not whether one happens. Reliability lives in the deterministic checks at the edges, not in how you wire the agents.
[READ] - 09
Remote-Controlling AI Coding Agents From Your Phone: What's Real, What It Costs, and Where It's Going
For fifty years, controlling a computer meant sitting at it. AI coding agents broke that. Here's how I drive my agents from my phone, what it actually costs, the local-vs-cloud distinction, and why this turns you from an operator into a conductor.
[READ] - 10
Running Multiple AI Coding Agents in Parallel: A System for Safe, Self-Merging PRs
Running several AI coding agents on the same repo turns into a coordination problem fast. Here's the system I built to keep them isolated, review each other's work, and merge their own pull requests safely.
[READ] - 11
Why I'm Building AI Education for Kids
I didn't think I'd hear back from Runway. Then this morning, I got accepted into their Builders program. Reflecting on why this mission is personal, and how we can ensure kids don't get left behind in the AI era.
[READ] - 12
Using AI Without Losing Your Humanity
Everyone has the same AI tools now, so volume stopped being a differentiator. What's left is whether any of it makes someone feel something, and whether the parts worth keeping survive being automated.
[READ] - 13
How I Built a Publishing Stack for AI Search
People are finding things through ChatGPT and Perplexity instead of Google. I built a pipeline that handles the optimization for that automatically. Here's how it works.
[READ] - 14
Building Systems, Then 10x the Output
At Zeta Global, my team gave me room to figure things out. I tried to do the same for the people who came after me, from responsive templates to training engineers across time zones.
[READ] - 15
45 Rule Files Were Making My AI Worse
I had 45 rule files telling my AI how to behave. The output was getting worse, not better. Then I measured. 170KB of context loaded per session. Here's what happened when I deleted a third of them.
[READ] - 16
Why I Let AI Fight With Itself Before I Ship Anything
When I build with one AI for too long, the output starts drifting, confident but incomplete. A second model catches what the first one misses. Here's the two-pass process I use now.
[READ] - 17
How I Turned My AI's Mistakes Into Guardrails
I added Husky because of an anti-pattern. CI because of an anti-pattern. PR requirements because of an anti-pattern. None of it came from reading best practices. It came from breaking things and writing it down.
[READ]