I started writing code in BASIC on an IBM PC.
That sentence is short. The years behind it are not. There was no Stack Overflow. There was a manual, a blinking cursor, and the particular kind of
stubbornness you get when the machine will do exactly what you told it — including the wrong thing — and nothing else.
BASIC was the on-ramp. It was enough to prove the computer was not magic. It was a box that obeyed.
Assembly, because someone has to talk to the metal
I taught myself MASM 5.0. Microsoft Macro Assembler. Registers. Interrupts.
The kind of programming where you stop thinking in “features” and start thinking in bytes and flags.
Assembly does not make you smarter than other programmers. It makes you harder to impress. Once you have moved bits around yourself, a framework is a convenience, not a religion.
C, and the book that stayed on the desk
I worked in C. A lot of us did. C was the language that felt like a contract with the machine without forcing you to live in hex all day. On my desk I kept Kernighan and Ritchie — The C Programming Language. Not as decoration. As the thing you opened when the compiler was right and you were wrong.
K&R is a thin book. That is the point. It does not try to be a career. It tries to be clear. You could outgrow tutorials. You did not outgrow what, at the time, was a reminder.
There was a lot in between: tools, shops, languages that came and went, jobs that taught you more about people than about pointers.
Windows 3.1 to Windows 95 was not a small step
If you were not there, Windows 95 sounds like branding. If you were there, it was weather.
Windows 3.1 was real work for a lot of us — programmers, WIN.INI folklore, the sense that the PC had a face now but the face still argued with you.
Then Windows 95 arrived and the conversation changed overnight. Start menu. Plug and Play that sometimes plugged and played. A consumer machine that wanted to be a product, not a kit.
It was a big deal. People waited long hours outside computer stores. Everyone knew the world was in for a big change.
IBM had its answer — OS/2 lived in that argument for years. Blue logos, serious engineering, and a market that was already leaning in another direction. I remember the gravity of IBM in that period.
It had only been several years since knowing the IBM "Blue Book" could actually keep me from getting a job developing a compatible BIOS.
Compatibility was a legal minefield. The people building PC-compatible BIOSes had to be able to show that they had not copied IBM's implementation.
Sometimes knowing too much about how IBM did it wasn't an advantage — it was a disqualification.
PC-Shell was my version of Windows 95. We kept the boxes we had bought for it for years in case they could be used. You didn't compete against Microsoft in those days either. lol
PC-Shell — a multi-threaded OS. IBM finally did OS/2. It bombed too.
I can say that sentence casually now. Windows 95 isn't truly threaded, I thought back then, so why is this not selling? It was my first glimpse into the realities of marketing.
Along the way I knew people like Walt Prue at USC’s Information Sciences Institute — ISI guys who were not writing demos for a keynote. They were building and operating pieces of what became the foundation of the net. RFCs. Regional research networks.
The unglamorous work of keeping packets honest when the internet was still a small enough idea that you could know the humans on the other end of the wire. It was pre-e-commerce even.
Later, I created something that became known as e-commerce. It was an online store for people to sign up for my hosting service and to build their online stores. It did well. I wasn't the only one; I just did not know at the time that there were a few of us all working on the same thing. Back then, dial-up modems connected to Visa and other processors. It was funny — you really waited for a transaction to complete. That wait has never gone away; it just reduced a notch in how long.
I also still remember fighting spam. It was a thing in 2000. We still deal with it, but now we have lots of solutions that did not exist 25 years ago.
There is always lots in between.
Networks. Databases. The web. Phones.
Security that used to be optional and became the job.
Clients who wanted yesterday.
Vendors who renamed the same problem and sold it as a platform.
Stored procedures people skipped because waiting felt worse than writing SQL into the app.
Contact forms that became spam cannons.
Email that forgot it was supposed to be a letter and became a landfill.
Forty-plus years is not a highlight reel. It is a stack of receipts.
You keep the ones that taught you something and you throw away the ones that only made you tired — except you never throw all of them away.
Pattern recognition is just memory with better lighting.
Then AI showed up
I have watched a lot of “next things” arrive. Most of them were louder than they were deep. AI is loud and deep. It's a big change. Ranked up there with the net. AI couldn't have formed without the net. What's next?
That is the question: what will be invented that needs the net and AI?
It was a few things: the net, years later the phone, then AI. Then boom — lots of things, because productivity will go through the ceiling, saving millions of hours across the board.
The fun of asking an AI to write about the path is not that the AI lived it. It did not.
The fun is the mirror: a machine trained on everyone’s text, asked to hold one person’s hallway of doors. It's part of life to look backward and forward.
There is AI on the desk next to whatever the new thin book turns out to be.
We will see lots of changes and technology in scientific and other fields. I've seen how one change creates lots of new changes before.
AI will be something that starts to take years off the next breakthrough. That means months, not decades. Things are about to change fast. I think it's the one thing people did not see as a result of AI, with all the noise there has been over the years.