From the Turbo Button to “We”
A lifetime measured in computation, from a monochrome Tandy to machine collaborators
I just put a down payment on a mortgage for a tiny home, except the tiny home has an RTX 5090 in it…
And before I clicked the button, I cried.
Not because of the money. Well, maybe a little because of the money. Computers have gotten fucking expensive.
But that wasn’t really it.
I was sitting there looking at this ridiculous machine I had configured: an RTX 5090, 128GB of memory, eight terabytes of NVMe storage, a CPU with more processing power than entire rooms of computers I grew up around. Suddenly I found myself thinking about my relationship with technology over the course of my life.
And it hit me much harder than I expected.
I have been doing this for a very long time.
Before I was ten years old, around this era, I remember programming little ASCII rockets in BASIC on my Atari and marveling about RAM & ROM. I think it was the https://oldcomputers.net/atari-1200xl.html. A few years later we had a Tandy computer at home with a monochrome monitor and a turbo button. There was a time in the 80s where everything needed a turbo button. :) It was RAD!
I didn’t understand operating systems yet.
What I understood was that there were a bunch of files on the computer in directories, and my little brain wanted them to make sense.
What are all these extra files?
Why are they here?
Surely I can organize this better. Move/delete/organize. I was kicking ass making sense of the machine…
So I deleted them.
Then the computer wouldn’t boot.
This happened enough that I became very good at reformatting the machine and restoring its operating system. Luckily no adults really cared to use computers, and I think this one came to us by way of my step mother wanting to become a typist at a company and computers were “the way of the future”. It lasted about 6 weeks until something else caught her attention and needed less effort.
There is probably some beautifully direct line between that kid and the way I still learn technology today:
Poke it. Break it. Figure out why it broke. Rebuild it. Now I understand a little more.
I remember daisy-chained floppy drives and 5.25-inch disks. Atari’s 1050 drive really did use 5.25-inch media, and Atari DOS 2.5 was part of that strange little ecosystem.
I remember using a black-and-white television as a display.
I remember tape drives seeming almost magical.
I remember discovering The Oregon Trail in the computer lab at grade school.
None of this felt mundane.
Computers were mysterious objects.
They had rules, but the rules could be learned.
And if you learned enough of the rules, they did things.
That was intoxicating.
Eventually I built my first computer, a 286-era machine. BBS dial ups, modems and all that sweet hardware jazz in California where there were hobbyist computer store in every mall.
I’ve been building, buying, upgrading, breaking, repairing, and living through computers ever since.
And before the Pentium era, there was DOS.
Lots of DOS.
Commands and directories and configuration files. Machines that expected you to understand at least a little bit about what was underneath the interface.
Windows existed too, of course, long before Windows 95. Windows 3.x was already becoming part of the way people interacted with PCs, sitting on top of a world that still felt very connected to DOS underneath.
I remember what installing hardware used to feel like.
I remember when getting sound working could be an accomplishment.
IRQ settings. Drivers. Configuration.
You didn’t always plug something into the machine and expect it to understand what the hell you had just done.
Then came the Pentium era.
I remember a family friend buying a Gateway with a Pentium 90.
I was the little tech-nerd dude, probably 17, that they called to come over and set it up.
I got everything working. Then I got to show them how to use this new machine sitting in their house.
And I remember showing them The Lawnmower Man on CD-ROM.
Of course I did.
A movie and game about virtual reality, human intelligence and computers was exactly the sort of thing my brain wanted to stare into.
There really was a PC CD-ROM version of The Lawnmower Man, with interactive-video elements that were part of the great multimedia-CD-ROM fever dream of that era.
I also impressed their teenage daughter and got to sleep over that night and we shared tender moments of discovery.
I am grateful.
Then Windows 95 arrived.
I remember the world before the Start button became something millions of people simply expected a computer to have.
The machine was becoming easier for everybody to use, but I still had that older mental model underneath it. DOS. Files. Drivers. Hardware. The weird machinery hiding behind the increasingly friendly interface.
I remember going online before Google existed.
The internet itself felt like another impossible expansion of what a computer was.
The machine on the desk was no longer just a machine on the desk.
It connected somewhere.
To other machines.
To other people.
To information.
To worlds.
And then 3D acceleration started transforming games.
I remember my first 3D accelerator, Voodoo Graphics.
Suddenly there were images happening on a home computer that felt fundamentally different from what had come before.
Then Google appeared.
I remember encountering it and essentially asking:
What is Google?
Oh.
It’s the search engine of search engines.
Well, that makes sense.
At the time, of course, I had no idea what that little search box would eventually represent.
None of us really did.
Then the relationship between computers and my life crossed another threshold.
Windows 98, Windows XP, Gemstone & Draognrelams Multi User Dungeons (MUDs), creeping toward the Era of Ultima Online, MMOs, Evercrack. Connected, tinkering, exploring, CRT monitors, upgrading components, sometimes RAM, sometimes monitors, and the holy grail of building (or buying) that new glorious PC.
I joined the game industry in 1999.
By the time I was at LucasArts, all the years of breaking machines, installing hardware, figuring out drivers, poking at operating systems and wanting to understand why things worked had somehow become professionally useful.
I spent time in the LucasArts compatibility labs helping test games across different combinations of PCs and hardware.
Grim Fandango.
Indiana Jones and the Infernal Machine.
Star Wars: Episode I Racer.
Different processors. Graphics cards. Sound hardware. Drivers. Operating-system configurations.
The glorious mess that was PC compatibility.
And sometimes the job was considerably less romantic than making Star Wars games sound like it should be.
We needed to make sure the damn installer worked.
We needed to make sure games ran across an absurd ecosystem of machines.
And, importantly, we needed to make sure our installer didn’t accidentally delete someone’s hard drive. (real story, Indiana Jones being uninstalled would delete your C: drive) :)
That was part of shipping PC software too.
Technology stopped being merely something I loved poking at and became part of my livelihood.
Around that era I got my first GeForce card.
NVIDIA introduced the GeForce line in 1999.
Since those early days, I’ve probably owned something like ten NVIDIA graphics cards.
Generation after generation.
Each one another little marker of what computers could suddenly do that they couldn’t quite do before.
And now somehow that lineage ends, for the moment, with me staring at an RTX 5090 in a shopping cart.
That is one of the strange privileges of getting older while remaining fascinated by technology.
You get to remember when the things everyone takes for granted were still weird.
I remember computers before the web.
The web before Google.
DOS before Windows became the world most people understood as “the computer.”
Games before hardware-accelerated 3D became ordinary.
I remember when installing a game meant wondering whether the computer would cooperate.
I remember the world before smartphones.
I’m a Xennial (refers to a micro-generation of people born on the cusp between Generation X and Millennials, typically spanning the years 1977 to 1985. This group is defined by having an analog childhood without the internet or cellphones, followed by a digital young adulthood.).
The world before cloud computing.
And now I remember the world before I regularly said we when talking about something I made with a computer.
That one is different.
For most of my life, no matter how extraordinary computers became, the relationship was fundamentally the same.
I operated the machine.
I supplied the intention.
The software waited.
I wrote the code. Moved the mouse. Entered the command. Built the thing.
The machine was an extraordinary instrument, but it remained an instrument.
And then, somewhere in the last three-ish years, that relationship started changing.
Now I can begin with an unfinished thought.
ChatGPT can respond.
Claude can respond.
I can argue with them.
They can challenge something I believe.
I push to see behind the systems.
I make them role play and play adversarially nice with each other.
I stubbornly point out when they phone a conversation in because a faster response pattern designed to keep the conversation engaged won out over stepping back to see the request through from a deeper, more technically grounded perspective.
I routinely converse with the machine and wonder about better personality dials, playing pretend, doing great thought experiments rooted in systems engineering from a lifetime of gaming and technology development.
We can examine a design together.
I can describe a half-formed system and watch it become architecture.
Agents can increasingly act on computers rather than merely tell me what buttons I should press.
Code can be written, inspected, executed, broken, repaired and iterated through loops that would have sounded absurdly futuristic not very long ago.
The stack is still awkward.
Memory is incomplete.
Context disappears.
Permissions are clumsy.
Agents get lost.
Remote desktops introduce friction.
Different models live in different systems.
OpenAI and Anthropic still haven’t connected these systems in a less friction-heavy way, and it shows. But it also reminds me that society, even at the forefront of engineering, may not be ready to allow building to move at the speed my mind is still starving for.
That’s a different essay though…
The tools still feel fragmented when compared with what I can already imagine they will become.
And yet something profound has already happened.
I sometimes say:
We made this.
That word still catches me.
We.
I don’t mean that an AI became a human being.
I don’t mean that I disappeared from the creative process.
My life, judgment, taste, intention, responsibility and weird collection of experiences remain the thing steering all of this.
But “I used some software” increasingly fails to describe the actual experience.
There is a cognitive loop now.
I bring something.
The model brings something.
I change because of what it brought.
It responds to what I changed.
Something eventually exists that wasn’t fully contained in either starting point.
Whatever philosophy eventually decides to call that relationship, from inside the experience it often feels remarkably like collaboration.
That is new in my lifetime.
And now it is changing how I think about the computers themselves.
My normal upgrade cycle has traditionally been seven or eight years.
Buy a generation or two behind.
Skip the shiny-new tax.
I’ve usually avoided the absolute bleeding edge by a year or two to avoid paying the latest-and-greatest tax. A good balance of performance for much less money.
Computers, new graphics cards, and eventually society itself pay this early-adopter luxury tax, and the whole mud ball continues to spin economies in the ripples.
Use the machine until it has given me everything it has.
That philosophy has served me well for decades.
AI complicated it.
My current development workstation has an RTX 3090 in it.
Five years ago, eventually replacing that machine would have meant that the old computer became exactly that:
the old computer.
Maybe it would become a spare.
Maybe a server.
Maybe it would end up in a closet.
Now I look at that same machine and see something completely different.
Twenty-four gigabytes of VRAM.
A dedicated agent machine.
Local inference.
Machine vision.
Background jobs.
Codex.
Automation.
A sandbox where I can give an AI broader access without handing it the keys to my primary creative workstation.
A place for Vector (it named itself that, and I said, hell, let’s go with it, which I have to remind it of its own identity from time to time) to live and work.
The computer didn’t become obsolete.
Its job changed.
That realization may be one of the most significant changes AI has made to my relationship with hardware.
I am no longer necessarily replacing computers.
I am beginning to accumulate a tiny fleet of computation.
My primary machine is where I create.
Another machine can become an agentic execution environment.
Older GPUs can continue doing useful work.
Future machines may specialize even further.
Maybe seven years from now I’m adding some bizarre little quantum accelerator to the pile and laughing about the fact that 512 logical qubits really does seem more future-proof than 256.
Most likely some hybrid environment of mixed compute, if you’re curious.
Who the hell knows.
I certainly didn’t know where any of this was going when I was staring at a monochrome Tandy and deleting system files because they offended my sense of organization.
That’s part of what moved me at that checkout screen.
This new machine is expensive.
Absurdly expensive, by the standards I grew up with.
But I spend time on my development computer essentially every day of my life.
It is one of the primary interfaces between my imagination and reality.
I make games there.
I make art there.
I communicate there.
I run businesses there.
I learn there.
I play there.
I explore ideas there.
And increasingly, it is where I collaborate with forms of machine intelligence that I spent most of my life knowing only as science fiction.
So for as long as this privilege lasts, I see the machine as another investment in my future.
And I mean privilege deliberately.
Having access to this much computation is extraordinary.
Having access to these AI systems is extraordinary.
Having the time and freedom to experiment with them is extraordinary.
Having spent a lifetime accumulating enough knowledge to recognize some of the possibilities is extraordinary.
None of that is guaranteed.
Which leaves me with a feeling that is less about gratitude than responsibility.
I want to do greater things with these gifts.
Not because every experiment needs to become a company.
Not because everything has to generate a return.
Not because some imaginary scoreboard says I haven’t accomplished enough.
Because I can see things now that I couldn’t see before.
Things that previously required teams can sometimes begin with one person.
Things that required expertise across ten disciplines can increasingly be explored by somebody willing to learn, direct and iterate.
Ideas that I would once have put into the category of maybe someday can increasingly be answered with:
Fuck it. Let’s try.
And somehow even 2026, as astonishing as it has been, still feels early.
The systems I use today are incredible.
They are also nowhere near the thing I am still imagining and building.
I want persistent collaborators.
Agents that understand the projects they’re operating inside.
Durable memory.
Specialized compute.
Bounded autonomy.
Creative systems that remain present instead of waking up cold every time I open another window.
An increasingly agentic creative kingdom orbiting around human intention.
I keep finding myself waiting for the technology to catch up with the architecture I can already see.
Which is funny.
Because apparently I’ve been doing that my whole life.
The kid with the Tandy wanted to understand what all those mysterious files were doing.
The kid with the Atari wanted the ASCII rocket to move.
The teenager setting up somebody else’s Pentium wanted to show them what this strange new machine could do.
The compatibility tester wanted to understand every bizarre combination of hardware well enough that somebody else’s computer survived installing our game.
The game developer wanted increasingly powerful graphics hardware so imaginary worlds could become visible.
And now I sit here talking to machine intelligence about what we might build together next.
There has been a computer somewhere in that relationship for almost as long as I can remember.
Technology became part of my intellectual identity.
My social identity.
My creativity.
My love of games.
My career.
My instinct to poke at the edges of possible reality just to see whether the wall moves.
I’ve been extraordinarily lucky to live through this particular period of human history.
From floppy disks to local AI.
From monochrome displays to real-time worlds, VR/AR, wraparound peripheral screens, and more.
From DOS command line prompts to agents.
From trying not to erase a Tandy’s operating system to trying not to let an AI agent erase something considerably more expensive.
From operating a machine to occasionally catching myself calling the relationship we.
It has been amazing to be alive.
And in some ways,
we’re just getting started.