40 Photos and videos
so basically, it's that the money never had to move in the first place.
opay → opay transfers don't touch NIP at all. it's two rows updated in one internal ledger. instant because there's nothing to wait for.
the float idea actually shows up on opay → traditional bank transfers, your balance updates instantly while the real interbank settlement (NIP) runs in the background, async.
as a backend dev, this is just optimistic UI sitting on top of a queue. update the user immediately, let the slow network call finish behind the scenes. feels instant because the user is never blocked on the part that's actually slow.
it's just good architecture.
The reason Opay or Palmpay transfers feel faster than GTB isn’t magic.
They debit you instantly from internal ledgers.
NIP settlement comes later.
They’re essentially lending you your own money for 30 seconds while they settle with the network.
It’s a float play disguised as speed.
1
1
4
614
Mr A is building a system for clinics to manage patient records. each clinic gets their own private space makes sense, you don't want clinic A seeing clinic B's patients.
he almost made one decision that would've worked fine for the first 5 clinics... and quietly destroyed him by clinic 20.
his plan: each clinic gets their own database. good. correct. patient data fully separated.
but he was also leaning toward: each clinic gets their own full copy of the entire app installed separately, hosted separately.
sounds like the same thing. it's not. and it's a trap.
1
9
one more landmine he almost stepped on: where do you store billing info. who's paid, what plan they're on?
not inside each clinic's database. that becomes 50 separate places to check who owes money.
one central database tracks all clinics their plans. each clinic's database only holds their patients.
1
6
spent way too long trying to get VS Code remote-ssh working on shared cPanel hosting.
kept failing. shared hosting just doesn't allow it.
found a simpler fix instead: connect straight from my terminal into the server. now i can sync code changes, set everything up, and apply updates directly on the live site, while still writing everything locally on my machine.
@tonyewills0 been in my ear about getting a vps. about to see if this setup holds before i make that move.
1
62
I find it very hard to believe.
a hiring pipeline that actually works would surface engineers from everywhere. not just people who fit a specific mould.
500 empty seats isn't a referendum on nigerian developers. it's a reflection of how you're looking for them.
necessity builds skill in ways a classroom never will.
May 4
The real issue is that the "500 job openings" at Moniepoint do not exist. It's a popular corporate scam where you constantly recruit without actually hiring anyone, because showing a large number of job vacancies while advertising your aggressive recruitment campaign is a growth metric you display to investors. It shows that the company is "growing fast" which means they should invest more money.
"We can't fill 500 job openings" in an absurdly deep job market like Nigeria is just a fintech bro performing for USD to locate him from whichever venture capital or private equity investor he's targeting. It's the same thing as when an NGO lawyer jumps on every trending human rights case without really achieving anything - it's a performance for the oyibo with USD to reward them.
Motion without movement. Fake activity scam.
4
151
building a product in nigeria that leans into foreign market pricing means you're already handicapped.
naira devaluation. settlement gaps. chargeback exposure. infrastructure fees that don't exist in other markets.
you can try to take the freedom fighter route and attack em, but you'll learn that your business model should be architected around it.
in my models I usually treat exchange rate as a variable, not a constant. i have a multiplier that absorbs naira movement before it touches my margin so by the time the rate shifts, my model already accounted for it.
naija tax is real tbh. but it's avoidable.
4
95
just dropped something new on mgbah.dev.
bespoke website templates built for specific niches.
hotels, law firms, dental practices, schools, real estate.
not surface level tho. each one has a curated system designed around how that business actually converts for your use-case.
mgbah.dev/templates
4
95
a little thing for those of you breaking into the tech space.
when you're on the come up, talent alone won't cut it.
you need exposure. the right kind.
i've been privileged to have the best kind of mentors pull me through. shoutout @tonyewills0 and bosskings. these guys have seen it all, done it all. and that's exactly where i wanted to be. i learned from their wins and their mistakes. everything to push me, they did.
that shortcut is worth years.
most people in tech are learning everything the hard way because they won't ask, won't reach out, won't put themselves in rooms where the right people exist.
find someone 3-5 years ahead of you in what you're building toward. study how they think. ask real questions. absorb the mistakes they'll actually share if you ask right.
you could call it luck if you want.
i call it leverage.
7
130
been on Linux for a few months now and i'm soo not going back.
specifically for backend development. the difference is real.
on windows i used laragon running in the background just to serve my apps locally. a whole external server stack just to see my own code work.
on Linux i got tired of even that. so i built my own CLI. (totally different from create-php-starter)
one tool. spins up apache virtual hosts, generates SSL certs with mkcert, tunnels with ngrok, tails logs, all from one command per action.
debugging webhooks locally used to be a pain. now i just run wayne share and i have a live HTTPS tunnel in seconds.
on windows you work around the system.
on Linux you build on top of it.
took me a while to fully switch tho. still configuring things honestly. but the productivity difference for backend dev is hard to argue with.
if you're a backend dev still on windows, at least dual boot. give it a month.
you'll understand what i mean.
1
1
10
108
stayed up all night!
Ws!
demo vid launch soon!
create-php-starter has been humbling me this past week.
react vite scaffold wasn't running npm install. pest was installing but never initialising. composer commands hanging silently because nobody told them to run non-interactively.
9 bugs. all found. all fixed.
the tool worked well enough to get 400 downloads. it wasn't working well enough for me to ship the new version.
well.. v2.1.0 is on its way. same one command. actually does what it says this time.
1
4
53
Michael retweeted
Apr 26
I was watching one of our old games today. We used to have Hazard, Willian, and Pedro as wingers. Killers in front of goal.
Now, we have Neto, Garnacho, and Gittens. They can't even score 15 league goals together.
Our quality in attack is not good enough, if we're being honest.
181
264
3,985
96,934
create-php-starter has been humbling me this past week.
react vite scaffold wasn't running npm install. pest was installing but never initialising. composer commands hanging silently because nobody told them to run non-interactively.
9 bugs. all found. all fixed.
the tool worked well enough to get 400 downloads. it wasn't working well enough for me to ship the new version.
well.. v2.1.0 is on its way. same one command. actually does what it says this time.
5
96
self improvement is masturbation.
in fight club, there's a scene. the narrator and tyler are in a bus, they see a calvin klein ad.
a commercial of an "ideal body".
the narrator is looking at a "manufactured ideal", a body that was designed by a corporation to make you feel inadequate so you'd buy something. a gap opens up between who you are and what the image says you should be.
that gap is the product. calvin klein isn't selling underwear. it's selling the feeling that you are not enough.
tyler durden then proceeds to drop an absolute banger.
self improvement is masturbation. (not literally tho)
he's saying that thing you're about to do? improving yourself toward that image? you're not doing it for you.
this maps directly onto a distinction nietzsche makes that most people miss:
self-improvement ≠ self-overcoming
self-improvement keeps the self intact and polishes it. self-overcoming destroys the self as it currently exists and builds something that couldn't have existed before.
"man is something that shall be overcome."
because the concept of modern day self improvement, is actually ego preservation, not transformation. you're not becoming someone new, you're just polishing the person you already are.
it feels like you're doing something. the morning routine, the journaling, the reading list, all of it feeds the internal narrative: "I am becoming." but becoming what, exactly?
1
1
5
90
self-improvement culture can be a sophisticated avoidance mechanism.
"i'm working on myself" is one of the most socially acceptable ways to not actually live. you never ship the thing. you never have the hard conversation. you never take the real risk. but you're reading books about it, so it counts, right?
the gym, the cold showers, the productivity system all potentially masturbatory if they're feeding the narrative of becoming without contact with actual reality.
real growth has stakes. it happens in friction, in failure, in risk, in relationship, in the thing you build that either works or doesn't.
he doesnt finish the statement tho, he goes on to say
"self-improvement is masturbation. now self-destruction..."
this is the most important moment in the exchange. and i think that's not an accident, it's the point.
there's multiple ways to read this, but i feel the ellipsis IS the philosophy. the sentence trails off into open space. the implicit message: figure out what comes after destruction yourself.
because if you need someone to finish that sentence for you, you're still in the same mode!
the most dangerous version of this isn't the person chasing the calvin klein body.
i see a lotta people claiming to read all the right books, planning all the right habits, saying they'll do all the right things and calling it growth.
at least the first guy knows what image he's chasing.
the second guy? well...
apply yourself.
1
37
openai just dropped GPT-5.5 yesterday and i'm genuinely considering trying Codex.
smarter agentic coding. fewer tokens to complete the same tasks. terminal bench scores that are honestly hard to ignore. benchmarks are showing it ahead of every model currently available publicly including Opus 4.7.
i've been all in on Claude but when a tool ships results like this, you try it. that means for N31,500/month (~$20). it's the ideal move. will definitely cancel a few pro plans i have on Claude for this.
competition is good for developers. when one tool starts pricing out its users, you move. and right now the math is clear.
(but it's anthropic, i might be wrong so let's see!)
Introducing GPT-5.5
A new class of intelligence for real work and powering agents, built to understand complex goals, use tools, check its work, and carry more tasks through to completion. It marks a new way of getting computer work done.
Now available in ChatGPT and Codex.
7
123