Field notes
Practitioner essays on building production software with AI agents.
Operator-grade reports on the systems, decisions, and economics of shipping software with AI — what worked, what failed, and what changed.
Articles
-
Picking the Right Model Won’t Fix Your Product Roadmap
Executive Deck ↗Exec summary ↗Listen ↗Download EPUB ↓Fix your value forecasts and investment decisions, then judge AI spending against labor cost and delivery. A cheaper model won’t rescue the wrong roadmap.
-
The Premature Primitive Problem: More AI Agents, Same Broken Delivery
Executive Deck ↗Exec summary ↗Download EPUB ↓Friends who now lead engineering organizations keep asking about missing AI returns. I’d look at what their agents can actually test and deploy.
-
You’ve All Gotten This Email
Your AI cost memo sounds sensible until you send it to San Francisco taxi drivers and ask them to avoid hills, skip breaks, and push the cab.
-
We Kept Delivery Flat While Cutting Our AI Spend to Zero Using These Simple Tips
Executive Deck ↗Exec summary ↗One company cut its $200-per-engineer AI bill to zero and kept delivery flat. Its playbook: 98-hour weeks, story-point accounting, employee-funded training, and attrition.
-
Elon Musk Started Cutting Twitter Before ChatGPT. Your Company Will Take Ten Years.
Executive Deck ↗Exec summary ↗Elon Musk started cutting Twitter before ChatGPT launched. Most companies will spread the same AI-assisted operating change across a decade because that is what executive incentives reward.
-
If You Make AI Scarce, Give It All to Your Best Engineers and Learn Something
Executive Deck ↗Exec summary ↗If leadership makes AI inference scarce, do not divide it equally. Give it to your best engineers, aim them at the company’s most important work, and learn what the constraint exposes.
-
AI Is Too Expensive for Your Software Organization. Turn It Off for 90 Days.
Exec summary ↗If a senior engineer's AI inference budget is less than one Starbucks drink every other working day, turn AI off for 90 days and measure what the shutdown exposes.
-
Are You Prematurely Optimizing the Machinery Around Your Vendor’s Agent?
Custom skills, agents, routers, and reviewer loops can improve a vendor's agent. Standardizing that extra machinery before it beats the supported baseline is premature optimization.
-
Every Engineer Has Four Jobs Now: Track the Models, Learn the Techniques, Know the Customer, Build the Software
Being an AI power user is no longer a specialist role. Every engineer must track the models, master the techniques, know the customer, and judge the software agents produce.
-
Your CRM Can Cost $3.5 Million a Month. Finance Panics Over a $100,000 AI Bill. Introducing the Inference Investment Theory (IIT).
AI value is not an inference metric. Start with product outcome, cycle time, total cost, and cost of delay, then overlay inference investment.
Corporate fiction
Three books. One operating problem. No clean hero.
Read 2028, Meridian, and AgentDrivenDevelopment.com’s Survive free online.
Read free online →