Surprise! My book arrived two weeks before the official release date. 🎉
"Spec-Driven Development: From Specs to Code with AI Agents" (Apress Pocket Guides) is now available, both as softcover and eBook.
It is a compact guide on how to make specs the contract between you and your AI agents, based on the AI Unified Process (AIUP). Requirements, use cases, entity models, and a complete case study, all in about 150 pages.
I hope it helps you bring more structure into AI-assisted development.
👉 https://t.co/Pb7hkPXmTa
#SpecDrivenDevelopment #AIUnifiedProcess #GenerativeAI #KeepITSimple
If you want to have original ideas, learn to do nothing. Literally. Sit with boredom. No phone. No music. No podcast. No endless scrolling. Just you and your thoughts. At first, your brain will hate it. It will look for stimulation. Distractions. It will tell you that you’re wasting time. Sit with the discomfort anyway. When you stop consuming other people’s ideas, your mind finally gets the space to create its own. If you never turn off the noise, you'll live and die an echo.
@ajibadde There are three ways to beat the odds:
1. Get the government to back the product and drive adoption.
2. Build an extremely affordable product that solves the problem well.
3. Build for the wealthy and build it exceptionally well.
This task changed how I approach difficult projects.
Today, I spend less time writing code first and more time understanding the system I'm about to change.
The biggest productivity upgrade I've ever had wasn't a framework.
It was learning to reduce uncertainty before writing code.
What's one project that permanently changed how you think as an engineer?
@blackie_360 What if it requires long running workflows ? I think you're trying to say MCP now supports and encourages stateless server implementations.
I know a company that has been trying to rewrite a SaaS for over 4 years. The big deal is the name has changed over 4 times and they still haven’t gone live and are still struggling. Now I am wondering why the entire company's technical team would decide to be inefficient.
The worst strategy is often a complete software rewrite from scratch. Thoroughly tested software that has been in production for a long time has value on its own. Don’t discard it too quickly.
That’s why we build our software stack in layers, with older, proven components at the centre. We build upon the past rather than starting anew: doing so is both risky and inefficient.
Study what makes you anxious. The topic you avoid. The person who makes you water yourself down. The question that feels insulting. The conversation you keep postponing. The silence you can't sit with. Pay attention. Every trigger points to something unfinished. An old wound. An insecurity. A belief you're still protecting. Sit with it. Study it. What triggers you reveals the parts that quietly control your life. Remember that.
Multiple disparate services
Multiple instances of the same service
Multiple versions of the same service
What's the best approach to implement BFF guys?
I missed my flight in Nairobi once because of traffic on Mombasa Road.
Like fully missed it. I got to the airport sweaty, exhausted, dragging my suitcase like it personally betrayed me.
I already knew there was no chance, but I still ran to the counter hoping for mercy.
The woman checked her screen, looked at me, then said, “You were the last passenger.”
I laughed a little because what else do you do at that point?
Then she lowered her voice and said, “The plane is still on the ground.”
Next thing I know this airport employee is SPEED WALKING me through the terminal like we’re in an action movie.
Security waved me through, another worker grabbed my carry-on to help me run, and I’m apologizing to literally everyone while fighting for my life.
I got to the gate completely out of breath.
The guy scanning boarding passes looked at me and said, “Eh, Nairobi traffic. We understand.”
People on the plane actually clapped when I walked in looking half dead.
I have never respected airport workers more in my life. Big up to them.
They must pick a multi-tenancy pattern that balances scalability, availability, latency, consistency, observability, and fault tolerance while keeping operational complexity manageable. Which multi-tenancy approach should Liam choose?
Liam is a backend architect building a SaaS analytics service. He expects 10,000+ small customers (light storage and compute) and a handful of enterprise customers who require strict isolation, custom backup/restore, and per-tenant scaling during large jobs.