“Woo” was the sound a 1984 Yamaha electric keyboard makes when a young girl procrastinates practicing piano. Our baby grand with all its keys lived in the living room on display, its repertoire governed by my piano teacher, my mother, recitals, and the expectations of daily practicing. The Yamaha, by contrast, was in my bedroom with buttons and pre-set voices rendering its keys to be possibility.
I would stand over it with the particular frozen-ness of a child confronting freedom thinking, in whatever syntax I had at the time, “Now that choice is infinite, how do I decide what to play?”
It was 20 years before when Robert Moog demonstrated a music-making device, a homemade synthesizer of sorts, at the Audio Engineering Society convention in New York, and opened curious maker minds. Where a pianist has 88 keys and three pedals to work with, a Moog player has access to knobs and cables that can produce any sound a human has the imagination to construct. The question was no longer what note? It was: what exactly are you trying to say?

Bob Moog’s badge at the 1964 convention via Bob Moog foundation
The current disruption with generative toolmaking, agentic coding, and rapidly emerging AI-assisted processes create the opportunity to decouple certain kinds of making from the tools. The knobs and cables are invisible, replaced by text fields, but the framework is the same. The questions about what to make and why have just moved upstream.
See also:
Kaitlyn Aurelia Smith,
modern synth wizard
and the process behind
her otherworldly music
The first generation of synthesizer musicians, largely responded to this vertigo starting by imitating the thing they already knew. Wendy Carlos’s Switched-On Bach (a favorite, apparently, when I was three years old) had success by showing machines could replicate Bach. Laurie Spiegel used touch-sensitive plates and sequencers for a more experimental approach and moved the mental model forward. The revolution was not that music machines could sound like a harpsichord. The revolution was that they didn’t have to.
Generative tools for design leave us with a similar question: when we can make anything, what is worth making? The mistake would be to repeat the harpsichord replicator error at scale: to use a tool capable of anything primarily to replicate what already exists, to ask the machine to write in the style of Didion or paint in the manner of Hopper, and call that success. One of the lesson of the synthesizer is that the first generation often imitates. This is not a failure of imagination so much as a necessary stage of grief: mourning the loss of the old constraint, which, it turns out, we confuse with craft itself.
Coda
This past holiday season, we spent time as a family gathered at my parents’ house. Each holiday, we have a tradition where everyone present must play an instrument or, if not, sing. Because of recent downsizing, the house is without its baby grand, so for the holiday, we rented a digital piano. At the end of a long evening, someone found and pressed the digital piano’s “Demo” button. We sat, while the piano self-played a music recital long into the evening.
Somehow, for that moment in time, it was exactly right.








