The Garage Is Open Again
We are told to fear what AI will take away. But I’ve seen this moment before — and last time, it liberated everything.
By Susan Lammers
I was recently a guest on the podcast embedded.fm with Elecia and Chris White to talk about the upcoming anniversary edition of Programmers at Work. We talked about the origins of the book, and the stories these programmers had to tell. Listening back to those stories forty years later, I was struck by how much the moment they describe feels like the one we are living through now.
Listen to the conversation on Embedded.fm here
In 1978, Dan Bricklin was at Harvard Business School doing his homework when an idea took hold. The product idea had germinated from his time working as an engineer at DEC on a word processor and from a job he had with a real small company that had successfully competed in the cash register space against NCR. Bricklin soon shared his software concept with his production professor who remarked how corporate teams doing materials inventory planning flew in weekly to fill blackboards the size of two large rooms — and how, if any single number changed, every calculation on every board had to be redone by hand. He encouraged Bricklin. That fall, Bricklin sat down to design the concept with Bob Frankston, who coded through the nights in his attic. They had sixteen kilobytes of memory, an idea, and a fearless conviction it could be done. The result was VisiCalc, the world’s first commercial spreadsheet.
I interviewed Bricklin for Programmers at Work in 1985. He was one of nineteen — a small group who fueled the personal computer revolution, most of whom knew each other. Listening back to those conversations forty years later, in a moment when artificial intelligence is rewriting what it means to build software, what strikes me most is not how different things were. It is how similar the feeling is to right now.
The feeling of a blank page. The feeling of new tools powerful enough to make almost anything, in the hands of an individual willing to try. The feeling, terrifying and liberating at once, of a future that has not yet decided what it will be.
Andy Hertzfeld was in his twenties at Apple when Steve Jobs pulled him from his desk one day for a side project that many at Apple considered farfetched. “I thought there would be guards with machine guns around the secret project,’” he told me, “but instead it was just sitting there. I was amazed there weren’t crowds of people around it all the time.” That project was the Macintosh. The handful of maverick engineers including Hertzfeld who worked on it were driven not by greed or fame but by a genuine passion and deep understanding of how these new personal machines could be putty in people’s hands for building something new and exciting. They wanted to make users happy. Constraints were not insurmountable obstacles. They were the mother of invention.
A personal computer back then was essentially, as Jaron Lanier declared to me back then, “an empty mind; it depended entirely on the person involved” and was under their control. He talked of the day coming when individuals could “speak and breathe programs.” For those creating software back then, it meant deeply understanding your vision and having well-honed skills to overcome the limits the machine presented. John Page said it well, “the real challenge is to design software that is simple on the outside but complex on the inside.”
The cost for those pioneers was enormous effort. The benefit was sovereignty and freedom. As Ray Ozzie told me recently, “in those days, perhaps half of the person*hours invested in developing an idea were focused on using well-honed skills to overcome constraints. Jonathan Sachs and Bob Frankston were masters of size and performance. I spent more than half of five years squeezing Lotus Notes into the constraints at the time.”
Regarding the computing environments today, Ozzie put it plainly when I caught up with him: “I’ve come to believe that constraints are deeply empowering. Whether the number of bytes or the length of the groove or the size of the canvas, limitations bring creativity to the fore. The walls of the software garden are now nowhere in sight; the tools limitless; boundaries almost non-existent, AI slop knows no bounds. What constraints will emerge within which creativity will best blossom? The medium needs them. We all need them.”
Here is what I think is happening now, beneath the noise of the tech layoffs and the hype cycles and the hand-wringing about what AI will replace: the tools have gotten powerful enough to give creative control back to the individual. Leverage, in the history of technology, has always been the precursor to disruption — not the corporate, venture-backed kind of leverage, but the genuine kind, where some young person in a room somewhere builds something no one expected and it changes the way a hundred million people live. A person with good AI coding tools today can build, alone or with a small team, what would have required twenty engineers and several years a decade ago. The fulcrum has shifted.
Are folks dropping out of college and the corporate world to work in garages again? No doubt. The barriers the cloud era erected — infrastructure costs, distribution dependencies, large engineering teams — are coming down. AI compresses the distance between a well-considered idea and a working product to a degree that the PC pioneers, who had to do all coding by hand, never experienced. The transcendent power of creating something from nothing is landing in the hands of a broader range of people than ever before.
What might the developers of today build? I find myself drawn to the mavericks looking to reshape our relationship with the cloud — to return or turn the tables so more personal power, privacy and control and value of content is by default in our hands. The cloud, for all its convenience, has been a slow reversal of ownership toward an emphasis on the needs of the profit often ad driven corporation at the expense of the individual customer. With every click of “I agree” dressed up as a new feature, we cede a little more of what we create to the collective cloud and LLMs. Perhaps these AI tools can help developers invent a way to put the individual back in the driver seat and the truly personal back in personal computing.
The risk-taking PC programmers of forty years ago were not trying to defeat the mainframe giants. They were liberated from them and working to solve problems that fascinated them, with new tools that had finally freed them from practicing their craft only on costly, time-sharing machines. The dominance of the mainframe companies faded as a consequence. That is usually how it works. The person who builds the thing that changes everything is rarely the person who set out to change everything. They set out to build the thing.
What ideas will the next generation of software creators pursue? I don’t know. But I am fairly confident of a few things: it will be something the current incumbents have decided is not worth their attention or have made deliberately difficult because their business model depends on the friction. It will look deceptively simple at first — a toy, a curiosity, a niche product for obsessives. And then, fairly quickly, it will not. The sparks will catch fire.
The programmers who fueled the PC revolution were working with almost nothing. They built almost everything. The software creators starting now have better tools than any generation before them — and they are inheriting a set of problems -- privacy, identity, creative sovereignty, freedom from extractive platforms--- that the original revolution never had to solve, because it hadn’t yet created them.
The garage is open again. The question is who walks in.
Susan Lammers is the author of Programmers at Work, first published in 1986. The 40th Anniversary Edition will be published Fall 2026 by 8080 Books / Simon & Schuster. You can pre-order it today on Amazon or wherever you buy your books. Please do!
