Notebook / Founder's Journal
What the work is teaching me.
Unfinished thinking, technical discoveries and questions from the process. Browse the notes and open one to read.
Notebook index
Find a subject, then follow the thinking.
Showing all 15 notes
15 FOUNDER / ENGINEERING · NOV 15, 2026The invisible work is still the product.What Hustle and CareSlot are teaching me about trust, correctness, and the work behind the interface.
This week reminded me that the most important parts of a product aren't always the ones people see.
While working on Hustle, I've been exploring how to make research, onboarding, and operations more dependable. That has meant spending time on things like access control, consent, deletion, isolated testing, and making sure interface actions actually do something meaningful.
I've also been working on CareSlot, where appointment booking pushed me to think more carefully about concurrent requests, persistent records, and the difference between checking for a conflict and preventing one.
I'm beginning to see a pattern in the way I build.
It's easy to measure progress by how many screens or features I've added. It's harder to measure whether a system behaves correctly when something unexpected happens.
A good interface creates confidence. The engineering behind it has to deserve that confidence.
I'm learning to pay more attention to the boundaries: who can perform an action, what happens when two requests arrive together, whether a user can recover from a mistake, and what evidence proves the result.
Some of this week's work is still in development branches. Some has passed local tests. Not all of it has reached production.
That distinction matters.
The invisible work isn't separate from the product. It's part of what makes the visible experience worth trusting.
14 FOUNDER / ENGINEERING · NOV 8, 2026Building it isn't the same as proving it.Learning to separate passing tests from proving the full experience works.
This week, I've been thinking about the distance between writing software and knowing that it works.
While building Hustle, I've been working through creator onboarding, permissions, private media, and the infrastructure needed to support them.
Some changes passed automated tests. Others reached preview environments. But I kept encountering another question: can someone actually complete the intended journey, safely and reliably?
A feature can be implemented without being ready. A deployment can succeed without proving the user experience. A system can pass isolated tests while still needing verification across its real dependencies.
I'm learning to treat those distinctions seriously.
I'm also beginning to appreciate how much of product engineering happens outside the visible interface. Authorization, data isolation, recovery, configuration, and deployment reliability all influence whether people can trust a product.
For Hustle, that means resisting the temptation to keep adding features before the existing experience is dependable.
I want to build products that work beyond demonstrations.
My next challenge is turning the engineering progress into verified experiences, one complete journey at a time.
What I'm taking forward: A feature isn't finished because I can show it. It's finished when I can demonstrate that it works under the conditions it was designed for.
13 FOUNDER / PRODUCT · NOV 1, 2026A better booking is worth more than another signup.Growth starts with making discovery, consultation and booking dependable.
I've been thinking about what growth actually means for Hustle.
It's tempting to keep adding features and finding more people to try the platform. But what happens after someone discovers a professional? Can they understand the service, ask the right questions, and complete a booking without unnecessary friction?
This week, the engineering work moved forward across onboarding, authentication, booking operations, and test-only payment journeys. Some improvements are merged into preview branches, but the complete experience still needs more verification.
I'm learning that having a feature isn't the same as making it useful.
My next priority is improving the journey people already need before chasing bigger numbers.
Learn. Apply. Exist.
12 ENGINEERING / FOUNDER · OCT 25, 2026Three different meanings of shipped.What I am learning about the difference between merged, deployed and verified.
This week, I kept returning to a simple question: when can I really say something is ready?
I made progress across Hustle and MONIFlow. Consultation and booking journeys became more connected, sandbox financial experiences became easier to test, and the iPhone previews moved closer to the way I want people to experience these products.
But the unfinished parts taught me just as much. A deployment can succeed while a final verification fails. A feature can be merged while the environment it depends on still needs work.
I'm learning to separate writing the code, deploying the system, and proving that the complete experience works. Each deserves its own evidence.
I still believe in moving quickly. I just want speed to produce something dependable—not something that only looks finished.
My next step is to close the remaining testing gaps and keep documenting what actually works, what doesn't, and what changes because of it.
How do I build faster without losing confidence in what I'm shipping?
Learn. Apply. Exist.
11FOUNDER EDUCATION · OCT 23, 2026Money has a price, and its value can change.Day 18 — inflation, interest rates, exchange rates and founder decisions.
Day 18 challenged how I think about money. Having the same naira balance does not mean I can afford the same things. Inflation can reduce purchasing power even when the amount in my account has not changed.
I learned to separate the interest charged on a loan from the total repayment. On a simple one-year ₦2 million loan at 15%, the interest is ₦300,000, while the total repayment is ₦2.3 million. I confused those numbers at first and corrected it through practice.
Exchange rates matter just as much to businesses. If a $400 monthly cost moves from ₦1,250/$ to ₦1,500/$, the naira cost increases from ₦500,000 to ₦600,000 without any change to the dollar price.
My takeaway is practical: before borrowing, investing in equipment, or raising prices, I need to investigate cash flow, costs, customer demand, contribution margins and alternatives. I finished the Day 18 assessment with 4/4. Learn. Apply. Exist.
10FOUNDER EDUCATION · OCT 22, 2026Growth is more than working harder.Day 17 — productivity, economic institutions and the conditions for development.
Day 17 was about why some countries and businesses become more productive and prosperous. I learned that output can grow through better skills, tools and systems, not just more working hours.
I practised the numbers too: increasing production from 200 to 300 units with the same inputs is a 50% increase, not 150%. Understanding the difference between the new total and the increase matters.
South Korea and Vietnam gave me examples of how education, productive industries, investment and access to international markets can support development. Strong institutions—such as clear property rights and reliable contract enforcement—help create an environment in which people are willing to invest.
I also took a founder lesson from the exercises: investigate the actual bottleneck. If electricity outages stop production, another machine may not fix the problem. I completed the final assessment with 3.5/4. Learn. Apply. Exist.
09FOUNDER EDUCATION · OCT 20, 2026Owning a company is not owning its wallet.Day 16 — separate legal identity, shares, dilution, vesting and control.
Day 16 corrected an assumption I had about ownership. Even if I own 100% of a company, its bank balance and assets do not automatically become my personal property. A company has a separate legal identity.
I learned how issuing new shares can dilute ownership. If I own all 1,000 existing shares and the company issues 250 new ones to another investor, I still own 1,000 shares—but that is now 80% of the total.
We also studied founder vesting. Under a simple four-year straight-line schedule, a cofounder allocated 40% might vest 20 percentage points after two years, subject to the actual agreement.
My takeaway: share ownership, day-to-day management and board authority are related, but different. They need clear rules before a team starts making major decisions. I passed the final assessment with 3/3. Learn. Apply. Exist.
08ENGINEERING / FOUNDER · OCT 18, 2026Building the feature is only half the work.What this week taught me about shipping, reliability and the last mile.
This week reminded me that writing code and delivering a working experience are two different achievements.
I made progress on Hustle's booking and mobile experiences, continued building MONIFlow's financial workspace, and launched a portfolio that can evolve alongside my work.
But the most useful lessons came from the things that weren't finished. A recovery test failed. Some deployment paths still need verification. Features that pass automated checks still need to prove themselves in real user journeys.
I'm learning to measure progress differently. A merged pull request matters, but so does knowing what remains untested, what could fail, and what must happen before a product is ready for people.
I still want to move quickly. I'm becoming more deliberate about knowing when something is genuinely ready.
01 FOUNDER'S JOURNAL · 4 MIN READA marketplace is more than a list of services.What building Hustle is teaching me about trust, discovery and real opportunity.
With Hustle, I started from a simple observation: skilled people often have the ability to do great work but not an equally good way to be discovered, trusted and hired. A directory can show names. It cannot, by itself, show how someone works, how reliable they are or whether an opportunity is a good fit.
That is why I keep returning to content-first discovery. Seeing the process behind someone's skill can be more useful than reading a generic service description. But attention is only the beginning. If discovery doesn't lead naturally to conversation, booking, clear expectations and payment, a marketplace has not solved the hard part.
I am also learning to think about distribution earlier. People already spend time on platforms like X, Instagram and TikTok. The challenge isn't to pretend those networks do not exist. It is to make the path from discovery elsewhere to a useful action inside Hustle feel natural.
This is still an active product hypothesis, not a finished victory. I want to validate it with real people, real constraints and behavior—not just a beautiful demo.
How can a skilled person's work become the strongest proof of their value, without adding friction for the client?
02 ENGINEERING NOTE · 3 MIN READA good AI agent should know when to ask.From AEON to SABI and MONIFlow, autonomy means little without clear boundaries.
Working on AEON, SABI and MONIFlow keeps bringing me back to one engineering decision: an assistant can do plenty of useful preparation without being allowed to make every final decision.
An agent can interpret a goal, compare options, prepare a shopping plan, contact a provider through an approved workflow or map a financial intention into steps. But the consequences of a purchase, payment or real-world contact deserve explicit permission and a visible record of what will happen.
I am exploring systems where an AI model proposes, while deterministic code enforces the rules: budgets, state transitions, approvals, evidence and safe execution boundaries. That separation is less dramatic than saying an agent can do everything on its own—but much more useful for products people might actually trust.
The lesson for me is to design the approval experience as a product feature, not an inconvenient pop-up. Users should understand what an agent knows, what it has done and what it is waiting for.
What should an agent be allowed to prepare automatically—and what should always remain a human decision?
03 LEARNING LOG · 3 MIN READBuilding something isn't the same as shipping it.Hackathons, CI failures and deployment fixes are teaching me to care about the last mile.
I like the early stage of a project: taking an ambitious idea, finding a shape for it and creating the first working experience. Hackathons have made that loop faster. They have also exposed a different kind of work: getting the thing to run reliably outside my own machine.
A working screen is not enough. Deployment paths, build environments, mobile edge cases, integration contracts, repeatable tests and clear setup instructions determine whether someone else can use what I made.
I am learning to treat the last mile as part of the design. In recent builds, I've had to revisit CI failures, mobile controls and the boundary between what a demo can show and what has actually been verified. Those aren't side quests. They are evidence about the quality of the product.
My goal is not to stop experimenting quickly. It is to make fast experiments more trustworthy: smaller checkpoints, clearer claims, better verification and less magic hidden behind the interface.
What would have to be true for a stranger to use this without me standing beside them?
04 ENGINEERING NOTE · 3 MIN READWhat if physical spaces had version history?REWIND and SENTINEL are experiments in remembering, comparing and verifying the real world.
Software gives us logs, commits and databases. Physical environments usually leave us with scattered photos, notes and a human's memory of what changed.
Projects like REWIND and SENTINEL are my attempts to explore a different model: take an observation, preserve a meaningful representation, compare it to a later one, and make the difference explainable. The interesting part isn't only identifying objects. It's the relationship between a previous state, the current state and the action someone can take.
I keep coming back to evidence. If an AI system stops mentioning a problem, that doesn't prove the problem was fixed. A later observation and a comparison are better grounds for a conclusion.
These are experiments, and environmental perception has real limitations. But the broader idea continues to interest me: intelligence becomes more useful when it remembers enough context to explain its own claims.
How do we make AI observations useful over time without pretending perception is certainty?
05 FOUNDER'S JOURNAL · 3 MIN READIn fintech, reliability is part of the interface.SWIFNA and MONIFlow are reminding me that the invisible systems matter most.
Designing a payment experience makes it tempting to focus on the fast path: choose a service, confirm, done. But financial software is defined just as much by the moments when something does not go as planned.
Working through SWIFNA's utility-payment flows and MONIFlow's human-approved money plans has pushed me to think about transaction states, provider responses, verification, refunds, retry behavior and clear user feedback. A great button cannot make up for an uncertain outcome.
I'm learning that trustworthy fintech is the meeting point of understandable interfaces and careful backend rules. The system should never confidently claim money moved when it only knows a request was sent.
Some parts of these products remain in development or prototype form. That distinction matters to me: a financial demo should explain what is simulated, what is integrated and what has actually been validated.
How can the interface be honest and reassuring even when the network, partner or transaction is uncertain?
06 PERSONAL GROWTH · 3 MIN READI'm trying to become better than my next build.A note on learning, communication, consistency and growing outside engineering.
I spend a lot of time thinking about software, but I don't want every measure of growth to be a repository or a launch. Learning to explain a complex idea simply, listen carefully, ask better questions and keep promises matters just as much when you're building with people.
I've been making space for small learning sessions outside my usual technical work: psychology, history, communication, creative thinking and personal reflection. Some days, progress is an experiment that works. Other days, it is understanding why an assumption was wrong.
My work in sales, travel services and volunteering has helped shape this perspective. Technology doesn't exist in isolation; it meets people who have competing priorities, limited time and understandable doubts.
I want to become the kind of founder who can build, communicate, learn from evidence, stay curious and keep showing up. There is no finish line on that one.
What did I learn this week that would still matter if I wasn't writing code?
07 LEARNING LOG · 4 MIN READWhy I'm developing my own way of building.Adaptive Ecosystem Development is my evolving attempt to connect experience, systems and feedback.
As projects get more complex, it becomes easy to rush from a promising idea into implementation. I've been developing an approach I call Adaptive Ecosystem Development, or AED, to remind myself to look at the whole system.
The twelve cycles move from vision and ecosystem mapping through simulation, architecture and experience design, into building, activation, observation and evolution. It is a method I am refining through actual products, not a universal rule I expect everyone to adopt.
A core lesson is that experience should inform implementation, and a running system should inform the next decision. That means considering the people, incentives, interfaces, failure paths and feedback loops around a feature—not just its code.
I don't know whether AED will remain in exactly this form. I do know why I started it: I want a more deliberate way to move quickly without losing the bigger picture.
Which assumptions should I test before I commit to an architecture?
No notes found.
Try another topic or search term.
Choose an entry from the index to read it here.