From Mainframes to Smartphones
The Computer That Weighed 30 Tonnes and the One in Your Pocket
Every smartphone you've ever held contains more raw computing power than the entire United States had access to in 1960.
The Idea
The arc from mainframe to smartphone is not simply a story of miniaturisation — it's a story of who gets to compute. For most of computing's early history, a computer was a shared civic object: enormous, expensive, housed in its own climate-controlled room, and operated by specialists in white coats. The ENIAC, completed in 1945, filled 1,800 square feet, consumed the electricity of a small town, and could only be programmed by physically rewiring its panels. Access was rationed like a scarce resource, because it was one. What the transistor revolution of the 1950s and 60s began — and what Moore's Law famously described — was a relentless compression: of size, of cost, of expertise required. Each generation of hardware democratised access slightly further. Minicomputers in the 1970s brought computing into universities and small businesses. Personal computers in the 1980s brought it into homes. The internet in the 1990s networked those homes together. And the smartphone, arriving in earnest after 2007, did something genuinely new: it made the computer ambient. Not something you go to, but something that goes with you. What's easy to miss in this timeline is that the fundamental logic — store a program, execute instructions, produce output — hasn't changed since the 1940s. What changed is the surface area of who participates. The computer went from being a tool of institutional power to being, in some sense, a personal extension of mind.
In the World
In 1969, the Apollo Guidance Computer helped land humans on the moon. It weighed about 32 kilograms, ran at 0.043 MHz, and had roughly 4 kilobytes of RAM — numbers that today seem almost comically modest. The engineers at MIT who built it had to invent an entirely new form of integrated circuit manufacturing just to make it small enough to fit in the lunar module. It was, at the time, the most advanced portable computer ever made. Forty years later, in 2009, the first phone running Android 2.0 shipped. It fit in a shirt pocket, cost the equivalent of a few weeks' wages, and had roughly 130,000 times more processing power than the Apollo computer. By 2023, a flagship smartphone was running chips with over 15 billion transistors etched onto a piece of silicon the size of a fingernail. But the more striking contrast is social, not technical. The Apollo computer was operated by a team of experts; its outputs were interpreted by other experts; its existence was a matter of national prestige and geopolitical competition. The smartphone is operated by eight-year-olds taking photographs of their lunch. That shift — from computing as institutional ceremony to computing as invisible habit — is arguably the most significant social transformation of the past half-century, and it happened so quickly that we barely paused to notice it.
Why It Matters
Understanding this arc changes how you read the present. When people debate whether AI will be democratising or centralising, whether social media empowers individuals or institutions, or whether the next wave of technology will be liberating or surveilling — they're really debating which direction the pendulum is swinging. The history of computing is not a straight line toward liberation. Mainframes gave way to personal computers, but personal computers gave way to cloud services — meaning the data and processing that once sat on your desk now sits, again, in someone else's climate-controlled room. The smartphone feels personal, but its infrastructure is deeply centralised. Knowing this rhythm — democratise, then recentralise; distribute, then consolidate — gives you a more honest lens for evaluating what comes next. The question worth asking about any new computing technology isn't just 'what can it do?' but 'who controls it, who has access to it, and who benefits when it scales?' Those questions were worth asking in 1945, and they're worth asking now.
A Question to Ponder
If the defining shift in computing history was moving from shared, institutional machines to personal ones — what does it mean that so much of our personal computing now runs on infrastructure we don't own and can't inspect?
Get a new one of these every morning.
Start learning with Thinkable