Build log
A real record of building products with AI agents — decisions, pitfalls, and what shipped. Three posts a week, assembled from my own collaboration logs.
Solo Founder + AI: I Finally Shipped Three Things I'd Been Putting Off for Ages
Over the past 7 days I logged 400+ AI collaboration sessions and 90+ git commits. The numbers don't really matter. What I want to talk about are the three things that actually solved real problems, and the mistakes I made along the way.
Is Your POS Just a Cash Register? Where MakanCloud Differs from a Traditional POS: The Back End
Picture this: it's closing time, and the POS says today was a good day. But first thing the next morning, the head chef notices a few prepped items are running low and fires off an order to the central kitchen in a WeChat group. The central kitchen checks stock in Excel, finds one item ran out ages ago, and ends up calling back and forth to fix the order. Come month-end…
400+ AI Sessions in One Week, and How I Now Split the Work
This week (Oct 12–18), I went back and forth with AI more than 400 times. Looking back through the logs, the real win wasn't how many lines of code it wrote. It was finally fixing a few problems that had been dragging on for ages.
Solo Dev + AI, Building an Enterprise System: Getting MakanCloud's HR Module Ready to Ship
This week I shipped MakanCloud v8.34.0. The main work in this release is stage S5.8 of the HR module.
100+ AI Sessions in One Week — Here Are the 4 That Actually Mattered
This week I went back through my log of work with AI: 100+ tasks, more than half on MakanCloud, and the rest split between my personal site, my content pipeline, and an old computer that keeps getting slower.
MakanCloud v8.34.0 is live: this time, all the changes are under the hood
When restaurant owners size up a system, most start with the obvious stuff: does it look nice, how many buttons does it have. After using it for a while, they realize what matters more: can you trace a problem when something goes wrong, is your data stored safely, and do the rules work the same way every time.
Three Months of Working With AI, and the Most Useful Trick I've Learned: Let It Say "I Don't Know"
If you're leaning on AI to get work done every day, this one might save you from falling into a very common trap.
This Week, AI Was Basically My Entire Tech Team
I spent the past week running a bunch of AI sessions on my computer at once. Looking back, I didn't hand-write much of the core code myself, but I got a lot more done than usual. No fluff today: here are the things that actually shipped, plus some thoughts on how far this "pair programming with AI" thing can really go…
Why Ordering Through WeChat Groups Leads to Overselling and Books That Never Balance
If you've ever run a restaurant chain, you probably know this morning all too well: the central kitchen's group chat is flooded with outlets calling in their orders one after another — “Outlet A, 20 char siu,” “Outlet B, 15 braised platters” — tangled up with threads left over from yesterday. The head chef is jotting everything onto paper slips between stir-fries, then heading back to the chat to confirm…
A Year Into Quitting My Job to Go Indie: Making a Living on Real Skills in the AI Era — Here's What I Shipped This Week
It's been almost a year since I quit my stable job to work full-time on my own products and content. People keep asking me: aren't you anxious? With AI this powerful, is there still room for indie developers?
2:30 a.m., and I Handed My Production Database to AI
This past week, I tried handing almost everything on my plate to AI. Not the shallow "write me some copy" stuff — I actually let it touch production systems and take on work where a mistake means someone has to answer for it. Here are a few concrete examples, plus a look back at what's safe to hand off and what isn't.
Missed Orders, Walkouts, and Long Queues at Peak Hours: How QR Table Ordering Fixes Them
Plenty of restaurant owners in Singapore know these three headaches well: you're short-staffed on weekends and customers wave for a server who never comes; the ordering queue stretches out and people get impatient; and worst of all, orders get written down wrong, never make it to the kitchen, or customers eat and leave without paying.
Upgrading My Investment Briefing Made Me Rethink AI's Long-Term Value
This week I did something small that I haven't been able to stop thinking about: I made two upgrades to the industry investment briefing my system generates for me every day. By the end, I had a much more concrete answer to the question of what AI's long-term value actually is. Quick disclaimer: this post is just my personal take, not…
Just Me and AI: I Built the Hardest Part of the System, the Finance Module
This week I want to talk about what I've actually been doing with AI lately, because it's gone way beyond writing a few lines of code for me.
The Hardest Part of Running a Central Kitchen Isn't Production
A lot of F&B owners assume that once their central kitchen is up and running, the hardest part will be "cooking fast enough."
How Far Can AI Coding Actually Take Me?
Lately, one question keeps coming up while I work on my projects: how much should I actually lean on AI for coding?
What I Pushed Forward with AI in the Last 7 Days
I used to think of AI coding mostly as "getting it to write code for me."
Staff Scheduling Is a Small Thing That's Quietly Slowing Restaurants Down
Most restaurant owners know it: scheduling isn't the flashiest feature in any system, but it's the one most likely to cause headaches every single day.
Which Internal Tools Are Worth Building Cheaply with AI
One thing has become clearer to me lately: what a lot of small businesses actually need isn't some big system built from scratch. It's a handful of internal tools that each solve one specific problem.
What Did I Actually Do With AI This Week? The Answer Is a Little Awkward
When I sat down to review the past 7 days of working with AI, I figured I'd come away with a few solid wins: maybe faster writing, maybe a clearer product plan, maybe a dev bug tracked down in record time.
Why Restaurant Owners Shouldn't Just Buy a POS
When most restaurant teams switch POS systems for the first time, they mostly care about three things: Is checkout fast? Does it have reports? How much does it cost?
Can One Person Plus AI Deliver an Enterprise-Grade System?
I used to take it for granted: enterprise systems are a team's job.
No AI Collaboration Logs This Week — and That Taught Me Something
This post was supposed to be about "what I've been doing with AI lately."
Cheap Is Nice, but Your Restaurant System Also Has to Be Reliable
When restaurant owners shop for a system, the first question is usually about price.
One Thing I Learned About Working with AI Today
While working on my project today, I ran into a classic AI collaboration problem, again.
I Didn't Do Anything Big with AI This Week, but I Found a Problem
While looking back on "what I've been doing with AI lately" this week, I ran into a slightly awkward but very honest result: there were no new AI collaboration records in the last 7 days.
Why Should Your Restaurant Software Carry Your Own Brand?
Once a restaurant brand grows past a handful of outlets, owners tend to learn the same lesson: more systems don't help. Fewer, more unified ones do.
What I'm Actually Living On After Quitting to Go Indie
Since quitting my job, I've had to face one question more directly every day: can I actually make a living from the things I build?
I Barely Used AI This Week, and Learned More Anyway
I went back through my last 7 days of AI collaboration logs, and the honest answer was: there weren't any worth mentioning.
What We've Shipped and Fixed Lately
Not every restaurant software update has to be a big new feature.
What Real Projects Taught Me About AI's Long-Term Value
The more I follow the AI industry, the more one thing stands out: the hype is easy to spot, but the real long-term value usually hides inside concrete, everyday workflows.
The real problems I pushed forward with AI this past week
This past week made something clearer for me about working with AI: it doesn't "conjure up a product" for me. What it does is help me push through the specific, tedious, easy-to-put-off problems inside a product, one at a time.
Why Restaurants That Take Orders Over WeChat Can Never Get the Books to Match
When things get busy, the one question every restaurant owner dreads is: "Wait, who actually confirmed this order?"
How Much of AI Coding Can You Actually Hand Off to AI?
Lately I've been leaning more and more on AI agents like Claude, Codex, and Hermes to write code.
The Potholes AI Helped Me Patch This Week
Looking back on this week, the stuff I used AI for wasn't exactly flashy.
Table QR Ordering: Payment First, Then the Kitchen
During the dinner rush, a lot of the chaos in a restaurant doesn't start in the kitchen. It starts before the order ever gets there.
For Small Businesses, Start With an Internal Tool
Lately, whenever I talk to people about custom software, I notice a common misconception: either you don't build a system at all, or you go all in on something complete, expensive, and heavy.
This Week, I Put AI to Work on Real Deliverables
My biggest takeaway from a week of working with AI wasn't that my productivity suddenly doubled. It was that all the small, fiddly tasks that usually stall me could finally be pushed forward one step at a time.
Running a Central Kitchen from Purchasing to Dispatch Without Spreadsheet Duct Tape
In a central kitchen, things rarely go wrong inside a single step. They go wrong in the gaps between steps.
What One Person Plus AI Can Actually Ship
I used to assume that enterprise-grade systems always needed a full team: product, design, frontend, backend, ops, QA, each person owning their own piece.
What I Actually Fixed with AI This Week
This week, the stuff I did with AI wasn't exactly flashy.
One Thing AI Collaboration Taught Me Today
Working on real projects with AI agents like Claude and Codex today, I hit a very basic problem again: the more I rush the AI for an answer, the more likely it is to give me a plan that looks complete but might not actually work in practice.
This Week, I Used AI on Real Problems
Over the past week, my take on working with AI got a little more down-to-earth.
Traditional POS Just Takes Payments. MakanCloud Handles the Back of House Too
Most restaurant teams get their first POS to sort out payments, printing tickets, and end-of-day closing.
What I'm Actually Solving After Quitting My Job to Go Indie
Quitting my job didn't suddenly make me free.
What I Built with AI in the Last 7 Days
This past week I worked with AI a lot. I didn't just ask it to write some code. I treated it like an engineering partner I could keep questioning, break problems down with, and look back on the work with.
From SGD 59 a Month, and It Keeps Taking Payments Offline: The Real Math for Singapore Restaurant Owners
Running a restaurant in Singapore, there are two kinds of surprise costs you dread most: the quote that scares you off before you even adopt a system, and the system that falls over right when you're slammed. The first keeps you putting off going digital. The second makes you more anxious after you do.
Two Weeks of Sprints: POS, Stripe, and Knowing When to Pivot
Pushing MakanCloud through three sprints taught me more from the deviations than from the plan itself.
Shipping QR self-ordering for restaurants — solo, with AI
How I designed and shipped per-table QR self-ordering for MakanCloud in days, not months — working with AI agents end to end.