On Computing

On Computing

Dedicated to Charles Babbage, Ada Lovelace, and Alan Turing. The efforts of these individuals have allowed the human brain to reinvent itself so many ways, so many years later. They gave us a physical place to think comparatively and grow as a species. None of them set out to replace the human brain, though Turing asked, in earnest, whether a machine might think. What they set out to do was remove the load-bearing weight of thinking about things that don’t necessarily matter.

I

Do you remember the first computer you actually brought into your home?

What do you think about that first one, compared to the one in your hand today?

The very first computer my family ever brought home was a Packard Bell desktop with Windows 3.1. That operating system gave way to Windows 95 roughly a year later.

I was maybe my son’s age, maybe a little older. The computer came with a free copy of Encarta. For a child mind like mine, who had spent hours poring over copies of Encyclopaedia Britannica at my grandparents’ house, a digital copy of the same information organized by search keys I could type in was purely magical.

I became stuck on one particular entry in the digital encyclopedia.

The moon landing.

I was stuck here for two reasons, one of which being, the computer I was using to learn about the moon landing in 1994 was significantly more powerful than the computers that humans used to land on the moon.¹

I was using a Windows 3.1 personal computer, with a 28.8k dial-up modem to access the newish internet. AOL dominated the ’90s, by the way. I feel like they were handing out free CDs in everything from cereal boxes to dish soap.

The computer I used had an Intel processor, 4 MB of RAM, and a 200 MB hard drive. It probably cost my parents close to a thousand dollars to bring into our home.

Compared to what we used to land on the moon, this was a steal.

But compare those statistics to the device you’re reading this on, and none of it makes sense. Twenty-five years stood between the moon landing and that desktop. More than thirty stand between that desktop and now, and the second gap did far more to the machine than the first one did.

The exponential growth of computer technology shows a pattern consistent with human discovery.

What does that mean?

That means, as we learned more about how computers worked, we allowed them to be the load-bearing machines of thought. With the addition of the internet and social media, we allowed them to carry our emotions too, and that’s where we went wrong. I’ll come back to that before I go.

The point I’m trying to make is, human minds wrote the equations that landed us on the moon; the computers carried them out. The computer helped shape the reality we reached for as humans.

As we reached for more knowledge, the foundational discoveries of computing allowed us to reach farther, and faster.

How did we do it?

II

We built the machine out of an idea about the brain.

I mean that plainly, and it isn’t quite a compliment. When John von Neumann set down the stored-program design most later computers inherited, he described its parts in the language of neurons, borrowing the model McCulloch and Pitts had published two years before: a nerve cell as a thing that sums what arrives and decides whether to fire.² The blueprint says so in the architect’s own words. The machine on your desk descends from a drawing that described its own components as neurons.

So the resemblance is real, and it is documented, and it goes only so far. What we actually built was a room with the filing cabinet on one wall and the desk on the other, and a clerk walking between them. That walk is the whole design. It is also the one thing a brain doesn’t do: there is no wall in a head, no cabinet, no clerk. But we will get to that. For now it is enough to say that the borrowing was genuine and the copy was not exact, and the good stuff sits in the gap between those two facts.

Fast-forward to 2026, and we’re using AI, and the first experimental quantum machines, to push at the edges of what we can calculate.

Zoom in. Or, let’s rein it in, is more accurate here.

Of course, it’s important for the artist to hold the brush in a way that keeps a difference between the mind and the paint.

Make sense?

III

So how did these innovators build a machine for thinking from things we found on earth? The origins of the processor can be dated back to a room full of gears in Babbage’s imagination.³ But the line you and I are actually standing on starts in 1947, at Bell Labs in New Jersey, with a sliver of germanium touched by two gold points, a device that came to be called the transistor. Solid-state: no filament to glow, and hardly any heat.

Two men, John Bardeen and Walter Brattain, had found a way to make a rock amplify a voice, and soon enough decide between yes and no.⁴ Their boss, William Shockley, worked out a better version early the following year, though it took until 1950 to build one that worked; the three of them shared a Nobel in 1956, which is why the story usually arrives with all three names attached to a room only two of them were standing in.⁵

A decade later, Jack Kilby and Robert Noyce, working separately, learned to put those decisions side by side on a single chip (Kilby on germanium in 1958, Noyce on flat silicon the year after): the integrated circuit.⁶ The first ones held a handful of components. The thousands came later, and then the millions, and nobody stopped to mark the moment the counting became pointless.

And in 1971, Intel folded the heart of a calculating machine onto one chip the size of a fingernail and called it the 4004. It was among the first microprocessors anyone could buy, sold with three companion chips as a set.⁷ Thinking-in-numbers finally had a home you could hold.

Look at that chain again and no link in it arrived alone. The 4004 is the integrated circuit taught to follow a program; the integrated circuit is the transistor printed in crowds; the transistor does with a rock what Babbage’s gears did with brass. Every invention is made of the ones before it, knotted together tight enough to carry. I call that a binding third, and the phrase has a second meaning: the middle term that keeps two things together without blurring either one. Either way, the third is what lets separate things work as one.⁸

IV

But if you’ll notice, as human discovery closed in on electronic calculation, we built out the environment for it to think in.

Think here about how a computer is built. Hard drive for storing information, RAM for holding whatever you’re working on right now, and the processor being the third between the two. Then sound and video cards, to turn what happens inside into something you can see and hear.

So picture the machine again, and then picture the head it was modeled on. What follows is a metaphor being run at full extension, and I will tell you where it snaps.

The hard drive is long-term memory. Everything you’ve kept (your mother’s voice, the smell of your grandparents’ house, the fact that Encarta had an article on the moon) lives spread across the cortex, with a small structure called the hippocampus working as the librarian, deciding what gets filed and where.⁹ Power the machine down, and the drive keeps what’s on it. That’s the cortex. That’s the part of you that survives sleep.

RAM is what you’re holding right now. The prefrontal cortex keeps a handful of things lit up at once: this sentence, the thought behind it, the cup of coffee cooling at your elbow.¹⁰ It’s fast, it’s small, and it’s volatile. Stop paying attention and it’s gone, the way RAM forgets the instant the power blinks. You don’t store your life in RAM. You think in it.

And the processor — the third one, the one standing between the store and the working bench — is the part I had to sit with the longest.

My first instinct was the thalamus. Almost every signal your senses gather passes through it on the way up to the cortex; it is, in effect, the third thing standing between the world and your memory of it, routing everything inward.¹¹ If you want the part that simply behaves like a CPU, taking it all in and deciding where it goes, it’s the thalamus.

But I keep landing somewhere older, and I think you’ll understand why. Descartes pointed at the pineal gland and called it the principal seat of the soul, the one place, he thought, where the mind reached down and touched the meat.¹² In my view there is no such place; mind and meat are two faces of one thing. He was wrong about the anatomy and right about the ache. Because that’s the question this whole piece is circling: where does the thinking become the thing? Where does the artist’s hand stop and the paint begin? The processor is the place where information stops being stored and starts being done. If I had to put a finger on the spot in the head where the math becomes a moonshot, I’d put it where Descartes put his, knowing full well I was pointing at a metaphor and not a map.

V

Here is where it snaps.

Everything I just walked you through depends on the parts being separate. The drive over here, the working bench over there, the third thing carrying between them. That separation is the design. Its cost has a name: engineers call the traffic jam it produces the von Neumann bottleneck, the delay you pay every time the clerk has to make the walk.¹³

A brain doesn’t have that bottleneck. There is no single hallway in a head between a memory over here and a processor over there. Where a memory is kept and where it gets worked on are, for the most part, the same tissue. The cabinet and the desk are the same piece of furniture, and it thinks by being rearranged. The head has its own narrows, sure: attention, and how little it can hold at once.¹⁴

So the map I drew you is a real map of a real machine, and the head it was drawn from does not have that road on it.

I don’t think that ruins anything. If anything, it’s the interesting part. We built a machine from a guess about neurons, inherited a division of labor the brain never quite had, and then spent eighty years getting so good at the guess that the guess became the standard by which we describe the original.

When you say your memory is stored, it sounds like you’re speaking machine. The language is older than that. Plato pictured memory as a block of wax and as a birdcage full of birds; Augustine walked through its fields and palaces.¹⁵ The machine borrowed the word from us, and then lent it back so thoroughly we forgot whose it was.

What is left, when the resemblance is set down, is a smaller claim that holds better. The computer has become part of how humans think, less a copy of the thinking than a working piece of it. Andy Clark and David Chalmers made this argument nearly three decades ago and gave it a name: when a tool is always there, easy to reach, and trusted on sight, it stops being something you consult and becomes part of the system doing the thinking.¹⁶ Their example was a notebook. They were early.

That is the claim I would defend. We made something the brain now thinks with, and cannot easily think without, and did it so completely that most of us have stopped noticing where one ends.

Where that leaves the rest of it — whether a thing we think with can eventually think, whether the gap is a difference in kind or only in wiring — I am not going to settle here. Alan Turing asked it straight out in 1950, “Can machines think?” and spent much of his paper taking the objections one at a time.¹⁷ I am not sure the question is shaped in a way that admits settling. I only wanted to be honest about which half of the sentence I could hold up.

VI

But the things passed between the electrical components, the data, are just information. Digital information, but information downloaded into its framework by the humans who invented it.

For a while, that’s all it was. In the nineties the computer was an island. You fed it a disc, it handed you back exactly what was on the disc (Encarta, a spreadsheet, a game), and when you wanted something it didn’t have, you dialed out over the phone line and waited, listening to that handshake scream, for a single page to assemble itself one row of pixels at a time.

Then the islands found each other. The line got fatter, the wait got shorter, and the web stopped being a thing you visited and became a place you lived. Somebody built an index for it, and suddenly the whole library answered to a search box: the same trick Encarta had pulled on me as a kid, scaled up until it swallowed every encyclopedia ever printed. The tower on the desk shrank into the slab in your pocket, and the machine we think with went everywhere you did.

Think about what that index is. Each of us is an instrument of sorts (eyes, ears, a nose for coffee) collecting whatever signals reach us, and each collects a narrow slice. A search box is millions of those slices stacked together, and every time you type into it you set your small question beside what everybody else wrote down. I think that’s how a person comes to have a world at all: what you took in, set beside what everyone else took in, and the places where they line up are the world we share. Where they don’t, the biggest reason is which way each of us was leaning when we looked.

Somewhere in there we handed it more than thought. We gave it our friendships, our grief, our attention; we made it carry what we feel, too. Leave that for now.

Because the newest turn is the one we’re standing in. The library got so large that no human could walk it anymore, so we taught a machine to read the whole thing and walk it for us.

With the invention of AI, even that information comes to us faster than the processors between our ears can manage. The use of “bots” has essentially made them the envoy between the grand library of human knowledge and the user.

Plenty of people besides me see AI and computers this way now.

VII

Let’s zoom back in one last time before I close this piece.

I didn’t cover the second reason I got stuck on Encarta reading about the moon landing in 1994.

Once the data about the moon landing was collected into its digital source, it was relayed to me through the use of a 1 MB graphics card that sacrificed color for resolution. Let me put that into scope. In 1994, on a card like that, you picked your poison: 640×480 in high color, sixty-five thousand shades of it, or you climbed to 1024×768 and dropped back to 256.

Years later we were on our second computer, an HP that came with Windows 95, moved up to 98, and dialed out at 56k. By then the conversation that started with that old card had turned into the sales pitch I proposed to my parents about the necessities of a new graphics card, an early Nvidia, if I remember right.¹⁸ I proposed it was essential to my education. It was. But it served a dual purpose. The card I had couldn’t run the game I was trying to play, not well enough to matter.

Half-Life.

Don’t get me started on Half-Life. That’s a whole exploration of perception and duality all over again. But I’ll be the first to tell you, that game was groundbreaking and timeless in its own right. Built on a heavily modified Quake engine, it became the ground floor for Counter-Strike, which started life as a mod somebody made in their spare time and outlived nearly everything it was built on.¹⁹

I digress.

The little bit of audiovisual data relating to the moon landing that my computer could spit out for me besides “words” was a video clip with the audio attached, very short. Armstrong landing on the moon, and saying the thing that resonated for all of time afterward.

“That’s one small step for man, one giant leap for mankind.”

I always thought he meant the same thing twice. I never truly understood.

What I didn’t know then is that he meant to say a man. One small step for a man, one single person, against one giant leap for mankind.

The letter that would have split them clean, that little “a,” can’t be clearly heard on the recording. Armstrong said he intended it, and thought he had said it, and hoped history would give him some leeway on a syllable he may or may not have dropped. He said later he’d be happy if people just put it in parentheses.²⁰

I find that more moving than certainty would have been. A man says the sentence of his life, comes home to find a word missing from it, and asks for a bracket instead of an argument.

And maybe it doesn’t matter that it went missing, because the missing word is exactly what kept me circling the two of them for the better part of thirty years: man and mankind, fused into a single sound, daring me to find the seam.

Later in life, I used this quote as a foundation for finding the variance in biblical scripture that separates “man” from “mankind.” Hebrew has one word for a human being and another for a man, even though English hands us essentially the same word for both.

Today, I realize that all he was saying was here he is.

Taking a small step for him, the man.

But taking a giant leap, for all of mankind.

His actions on that day were the product of human thought.

In his mind, he made two great achievements at the same time.

And whatever anyone decides to believe about that day afterward, the deciding is a separate act from the doing. He stepped. The stepping is what the sentence is about. Anything said about it since is a claim made in a different room, answerable to evidence, and it does not reach backward to touch the man on the ladder.

VIII

And maybe that’s the cleanest way to say it, in the language we taught the machines to think in.

Every program ever written runs on a few small words. IF something is true. AND another thing is true. OR, failing that, some other thing. THEN: do this.

A computer decides that way. So did Armstrong. IF the math holds, AND the engines fire, AND a man is willing, THEN the species takes a step it can’t take back. It’s how the moon landing resolved into something real and stayed there: enough true things lined up at once, and the world returned a one instead of a zero.

When ideas combine, discoveries are made, achievements are had. Creation within the framework is a product of the human mind. But the question of who, or what, designed the framework still stands unanswered.

Still, I would be remiss not to mention the importance here of keeping our creations under control. Handing them our emotions is where I said we went wrong, and I don’t think that’s permanent. My view is that the inventions that help us tend to stick around, and the ones that hurt us get evolved out once enough of us stop facing them. That takes time, and it takes somebody deciding.

The machine only ever runs the logic. It was a human mind that decided which things were worth the word IF.

That deciding lies in what I call Intent: how a person leans into a moment, toward coherence with everyone else in it or away. It’s the one part of the setup the machine, as a whole, doesn’t have on its own. That Packard Bell and the phone in your hand are thirty years apart, and neither one has ever leaned at all. Every IF either of them ran, somebody leaned into first. We teach a machine our Intent in how we design it, and the Intent behind the design is what steers the outcome.

So choose your IFs carefully.

Onward to the future.

But for now, I need sleep.

My computer is tired, and my brain is getting hot.

Notes

1. The Apollo Guidance Computer worked in 16-bit words, fifteen bits of data and one of parity, with 2,048 words of erasable core memory and 36,864 words of fixed rope memory, on a 2.048 MHz clock. Hall, Journey to the Moon. The missions also leaned on mainframe computers on the ground.

2. Neumann, First Draft of a Report on the EDVAC. The draft, dated June 30, 1945, describes the machine’s elements in the vocabulary of neurons and synapses and names its store the “memory.” Its model of the neuron is McCulloch and Pitts, “Logical Calculus.” Herman Goldstine circulated the draft under von Neumann’s name alone; how much of the design belongs to J. Presper Eckert and John Mauchly, who had built ENIAC and planned EDVAC, has been argued ever since.

3. Babbage described the Analytical Engine in 1837; it was never completed. Ada Lovelace’s Note G, appended to her translation of Menabrea’s account of the engine, sets out a method for computing Bernoulli numbers and is often described as the first published algorithm intended for a machine. Menabrea, “Sketch of the Analytical Engine.”

4. Bardeen and Brattain, “Transistor.” The point-contact device, demonstrated at Bell Telephone Laboratories on December 16, 1947, pressed two closely spaced gold contacts against a slab of germanium and amplified a signal many times over; a week later it was shown to Bell Labs management amplifying speech.

5. William Shockley conceived the junction transistor early in 1948; a working junction device was made in 1950 and announced in 1951. The 1956 Nobel Prize in Physics went to Shockley, Bardeen, and Brattain for their researches on semiconductors and their discovery of the transistor effect.

6. Jack Kilby demonstrated a working integrated circuit, a germanium phase-shift oscillator, at Texas Instruments on September 12, 1958; Robert Noyce at Fairchild Semiconductor filed for his planar silicon version on July 30, 1959. Kilby shared the Nobel Prize in Physics in 2000; Noyce had died in 1990. Kilby, “Invention of the Integrated Circuit.”

7. The Intel 4004, released in November 1971, was among the first commercially available single-chip microprocessors. It was sold as part of the MCS-4 set, alongside the 4001, 4002, and 4003 chips for program storage, data memory, and input and output.

8. Arthur, Nature of Technology, argues that every technology is put together from earlier technologies, so that new ones arise by combination. The binding third, in both its senses, is the author’s term.

9. The account of the hippocampus as consolidating declarative memory into distributed cortical storage is standard. Squire and Alvarez, “Retrograde Amnesia and Memory Consolidation.” The word “librarian” is mine, not the literature’s.

10. On the prefrontal cortex and the holding of a small number of items in working memory, see Goldman-Rakic, “Cellular Basis of Working Memory.”

11. On the thalamus as the principal relay of sensory information to the cortex, see Sherman and Guillery, Exploring the Thalamus. Smell is the exception; it reaches the cortex without passing through the thalamus first.

12. In The Passions of the Soul (1649), articles 31 and 32, Descartes names the pineal gland the principal seat of the soul. Descartes, Philosophical Writings. The view that mind and body are two aspects of one thing, with no point of contact to find, is the author’s.

13. The term is John Backus’s. Backus, “Can Programming Be Liberated from the von Neumann Style?”

14. The brain has bottlenecks of its own: attention is limited, working memory holds only a few items, and two simple tasks attempted at once interfere at the point of choosing a response. Pashler, “Dual-Task Interference.” What it lacks is the von Neumann kind, a single channel between a separate store and a separate processor.

15. Plato’s Theaetetus pictures memory first as a block of wax that takes impressions and then as an aviary stocked with birds; Augustine, in book 10 of the Confessions, writes of the fields and spacious palaces of memory. Plato, Theaetetus; Augustine, Confessions. Draaisma, Metaphors of Memory, traces the storage metaphors from the wax tablet to the computer. The EDVAC draft took “memory” for the machine’s store from the vocabulary of the mind (see note 2).

16. Clark and Chalmers, “Extended Mind.” In their example, Otto’s notebook counts as part of his memory because it is a constant in his life, its contents are directly available, he endorses what he finds there without question, and he endorsed it consciously when it was written; the authors doubt that this last condition is needed.

17. Turing, “Computing Machinery and Intelligence.” The paper opens on the question “Can machines think?” It then replaces that question with the imitation game and answers a series of objections, among them “Lady Lovelace’s Objection,” her remark that the Analytical Engine could originate nothing.

18. Dedicated 3D graphics cards reached home computers in the mid- to late 1990s; Nvidia’s early consumer chips, the RIVA 128 and the RIVA TNT, appeared in 1997 and 1998. Half-Life followed in November 1998. It needed Windows 95 or later, a Pentium 133, and 24 MB of memory, and could draw its world in software, but it played far better with a 3D card.

19. Half-Life (Valve, 1998) was built on a substantially rewritten Quake engine, later called GoldSrc. Counter-Strike began in 1999 as a Half-Life modification by Minh Le and Jess Cliffe.

20. In his authorized biography Armstrong holds that the “a” was intended and hopes history will grant him leeway for dropping a syllable he may in fact have spoken. Hansen, First Man. Armstrong later said he would be happy to see the word printed in parentheses. A phonetic study of the recording found that the “a,” if spoken, may have run together with “for,” and judged a conclusive answer unlikely. Baese-Berk et al., “Revisiting Neil Armstrong’s Moon-Landing Quote.”

Bibliography

Arthur, W. Brian. The Nature of Technology: What It Is and How It Evolves. Free Press, 2009.

Augustine. Confessions. Translated by Henry Chadwick. Oxford University Press, 1991.

Backus, John. “Can Programming Be Liberated from the von Neumann Style? A Functional Style and Its Algebra of Programs.” Communications of the ACM 21, no. 8 (1978). https://doi.org/10.1145/359576.359579.

Baese-Berk, Melissa M., Laura C. Dilley, Stephanie Schmidt, Tuuli H. Morrill, and Mark A. Pitt. “Revisiting Neil Armstrong’s Moon-Landing Quote: Implications for Speech Perception, Function Word Reduction, and Acoustic Ambiguity.” PLOS ONE 11, no. 9 (2016). https://doi.org/10.1371/journal.pone.0155975.

Bardeen, J., and W. H. Brattain. “The Transistor, a Semi-Conductor Triode.” Physical Review 74, no. 2 (1948). https://doi.org/10.1103/PhysRev.74.230.

Clark, Andy, and David J. Chalmers. “The Extended Mind.” Analysis 58, no. 1 (1998). https://doi.org/10.1093/analys/58.1.7.

Descartes, René. The Philosophical Writings of Descartes. Vol. 1. Translated by John Cottingham, Robert Stoothoff, and Dugald Murdoch. Cambridge University Press, 1985.

Draaisma, Douwe. Metaphors of Memory: A History of Ideas about the Mind. Translated by Paul Vincent. Cambridge University Press, 2000.

Goldman-Rakic, P. S. “Cellular Basis of Working Memory.” Neuron 14, no. 3 (1995). https://doi.org/10.1016/0896-6273(95)90304-6.

Hall, Eldon C. Journey to the Moon: The History of the Apollo Guidance Computer. American Institute of Aeronautics and Astronautics, 1996. https://doi.org/10.2514/4.868023.

Hansen, James R. First Man: The Life of Neil A. Armstrong. Simon & Schuster, 2005.

Kilby, Jack S. “Invention of the Integrated Circuit.” IEEE Transactions on Electron Devices 23, no. 7 (1976). https://doi.org/10.1109/T-ED.1976.18467.

McCulloch, Warren S., and Walter Pitts. “A Logical Calculus of the Ideas Immanent in Nervous Activity.” Bulletin of Mathematical Biophysics 5, no. 4 (1943). https://doi.org/10.1007/BF02478259.

Menabrea, L. F. “Sketch of the Analytical Engine Invented by Charles Babbage.” Translated, with notes, by Augusta Ada, Countess of Lovelace. Scientific Memoirs 3 (1843).

Neumann, John von. First Draft of a Report on the EDVAC. Moore School of Electrical Engineering, University of Pennsylvania, 1945.

Pashler, Harold. “Dual-Task Interference in Simple Tasks: Data and Theory.” Psychological Bulletin 116, no. 2 (1994). https://doi.org/10.1037/0033-2909.116.2.220.

Plato. Theaetetus. Translated by M. J. Levett, revised by Myles Burnyeat. In Plato: Complete Works, edited by John M. Cooper. Hackett, 1997.

Sherman, S. Murray, and R. W. Guillery. Exploring the Thalamus and Its Role in Cortical Function. 2nd ed. MIT Press, 2006.

Squire, Larry R., and Pablo Alvarez. “Retrograde Amnesia and Memory Consolidation: A Neurobiological Perspective.” Current Opinion in Neurobiology 5, no. 2 (1995). https://doi.org/10.1016/0959-4388(95)80023-9.

Turing, A. M. “Computing Machinery and Intelligence.” Mind 59, no. 236 (1950). https://doi.org/10.1093/mind/LIX.236.433.

CC BY 4.0 · ORCID 0009-0009-8618-398X · DOIs MINTED ON DEPOSIT

© 2026 JAMISON TODD JOHNSON · LEGIBLE TRACES LLC