Blog

Why Pylonworks defaults to Next.js: ISR, edge routing, the Vercel deploy story, TypeScript fit, and the one project type where it is the wrong call.

How a two-page brief, one design checkpoint before any code ships, and a signed spec cut my average revision rounds from 3 to 1.4 per project.

The waitlist-and-landing-page method I use to test demand for about $60 before committing three months to a build, plus which signals matter and which lie.

The four-piece stack I use to keep Claude agents running around the clock: systemd timers, the Agent SDK, Postgres for state, and OAuth rotation. Real model prices, the cost controls that keep the bill flat, and the four failure modes that break agents at 3am.

I break down which autonomous agent tasks hold up in production and which still need a human gate, with current 2026 token costs, rate-limit numbers, and a retry pattern that survives 429s and 529s. The dividing line is whether a test catches the mistake first.

The pipeline that drafts and queues about 20 personalized outreach touches a day for roughly $0.40 in Claude inference, why I never let the model press send, the Haiku scoring pass that cut drafting cost 60%, and the send limits that keep you out of spam.

The AI task automation decision comes down to one equation most people skip. I break down the real per-run token cost, the verification trap that kills ROI, and a five-factor rubric for choosing what to automate first.

I shipped ten AI side projects since 2023 and kept two. Here is what the other eight cost me in real dollars, the failure modes that killed them, and the spend caps, model choices, and caching that made the survivors cheap enough to leave running.

I ran 1,000 invoices through an agent at 94% accuracy. That missing 6% is why a human still approves every run. Here is the real line between AI automation and AI assistance, what closing the loop costs, and how to tell which one you actually bought.

Five projects, one control plane, about $180 a month in Claude usage. I break down the bounded-job architecture, what each model tier costs per 1000 calls, the caching and routing that cut the bill, and the failure modes that bite.

Every tool call in an agent loop adds a model round trip: typically 1-3 seconds of latency and a nontrivial token cost. Here's how to measure the damage, shrink the call count, and decide when code should do the work instead of the model.

Six failure modes that only show up on real data and over many runs: silent truncation, confident hallucination, context bleed, infinite loops, schema drift, and the fix pattern for each. A catalog from production agent logs.

Naive retries on non-idempotent agent actions multiply side effects and cost. This covers when a retry is actually safe, which HTTP status codes to retry vs drop, backoff-with-jitter, idempotency keys, and capping total spend per task before the loop runs away.

Running Claude in production gets expensive fast. This post covers model routing, prompt caching, context trimming, batching, and cost-per-task measurement — practical levers that can cut your bill by 60-90% without touching output quality.

My Claude agents bill about $214 a month across 1,900 runs. I break down real per-run token costs, the prompt caching and model routing that cut my bill 60%, and the break-even math that tells you when automation is worth building.

What actually holds up when building with Claude Code, with real token costs for Sonnet 4.6 and Opus 4.8, why a fat CLAUDE.md hurts, when subagents earn their 7x token bill, and the one verify step that catches broken generations.

MCP servers let you write one integration that any AI agent can call. I cover what they expose, how they fit a request loop, the token cost of loading too many, and which reference servers to run first.