"Right! Today I Want to Be Able To…" (Fifty Years After Xerox PARC, the Malleable Computer Finally Exists) – Dave5

"Right! Today I Want to Be Able To…" (Fifty Years After Xerox PARC, the Malleable Computer Finally Exists)

Jul 13, 2026

Uncategorized

ai, AI/ML, artificial-intelligence, awesome tech, beauty, culture, llm, technology

TL;DR: Open system + AI colleague + AI-aware editor = the malleable computer Xerox PARC promised fifty years ago. Smalltalk bet the missing piece was a friendlier language; it was a colleague.

Most mornings now, somewhere between the first coffee and the second, I sit down at my computer with a thought that begins: "Right! Today I want to be able to…" Recently it was: kick off agent work in my editor, wander off to another workspace, and get tapped on the shoulder when the agent needs me. By mid-morning, my computer did the thing. Not an app that approximately did the thing, found after a search and a compromise. My computer, reshaped, doing exactly the thing.

I want to convince you that this is not a productivity anecdote. It’s the arrival, fifty years late, of the most important idea in the history of personal computing, one we’ve been living in the shadow of the whole time. And it took an unlikely combination to get here: an AI, an operating system built by volunteers, and an editor that treats the two as colleagues.

The idea we left behind at PARC

In the 1970s, Alan Kay‘s group at Xerox PARC built Smalltalk and the machine it lived on, the Alto . Everyone knows what we took from PARC: the mouse, the windows, the bitmapped display, the desktop metaphor. Apple famously visited, saw the demo, and the next fifty years of interfaces descended from that afternoon.

But the interface was never the idea. The idea was that the entire system was live and malleable. In a Smalltalk environment there was no wall between "using" and "programming." Every object on screen could be inspected, modified, extended, while it ran. Kay’s Dynabook vision was personal dynamic media, a computer you shape the way you shape an essay, not personal consumption media.

We didn’t get that future. We got its screenshot. The industry took the Alto’s surface and shipped it on top of sealed systems where the user’s role shrank back to operator. Fifty years of windows and icons, running on machines philosophically opposed to what PARC was actually about.

Why didn’t the real idea survive? Because malleability had a brutal price: expertise. Smalltalk bet that a sufficiently humane language would let ordinary people (children, even) reshape their systems. It was a beautiful bet, and it didn’t pay off at scale. Reshaping a computer remained the province of people willing to spend years learning how. So the industry, reasonably, sealed the machines and sold us polish instead.

That price just collapsed. That’s the whole story. The missing piece was never a friendlier language. It was a colleague: one who has read every man page, never tires, and sits inside your editor. Pair an AI like Claude with an environment that is actually open to being reshaped, and the Alto’s bargain finally clears: malleability without the years of apprenticeship. The complexity doesn’t disappear; it becomes conversational.

The rest of this post is what that looks like in practice, and how I stumbled into it.

Twenty-two years in the walled garden

I spent twenty-two years inside Apple’s ecosystem, and if I’m honest about the total (the Macs, the iPhones, the upgrades, the peripherals, a music library built album by album, and the machines I bought for family along the way) it’s many tens of thousands of dollars. And that’s just my own wallet; count the purchases I steered as the tribes designated computer person, and my economic footprint in Cupertino runs well beyond myself. I don’t say that with bitterness. For most of those years it was money well spent: beautiful machines, software that stayed out of my way, a promise of "it just works" that was largely kept.

But somewhere along the line the deal changed. I came of age inspired by Steve Jobs. I know exactly how problematic he was, and the hagiography deserves every one of its asterisks. But the thing he understood, intuitively and completely, was that shareholder value is a trailing indicator of amazing products: delight the customer beyond reason, and the numbers follow. Tim Cook is a near-wizard in his own right, probably the finest operator of his generation. And under him the causality quietly inverted. The numbers became the product; delight became a line item. Services revenue, upgrade cadence, ecosystem retention: I could feel myself being optimised, and it killed the love. The premium I was paying stopped buying capability and started buying containment. I was renting a very nice room in a house whose walls I wasn’t allowed to touch, with a music library that made moving out feel unthinkable. Apple, ironically, took the Alto’s demo and built the most beautiful sealed...