1/ Your AI strategy is actually designing your organization.
2/ Most AI strategy still treats AI like an instrumental capability: acquire models, stand up a some kind of platform, define use cases that it seems the tool is suited for, add governance, and policy, and rules-of-use..and then just measure your ROI as the goose lays a bunch of golden eggs for you.
3/ But, that frame is already way too small, and likely to trip you up before it even gets going.
4/ Way too nearsighted.
5/ AI is not a productivity layer.
6/ I know it feels like it is, and it says so on the bill-of-sale.
7/ It is more like a medium one wonders into — a fog of confusion and possibility that is actively reshaping the meaning and value of the terrain this fog settles over.
8/ When you deploy AI at scale, you are not just adopting some new fangled tool. You are setting new defaults for attention, authority, learning, power, coordination, maintenance, and even what counts as legitimate knowledge.
9/ Sigh..well, bear with me here for a moment. I want to tell you a story that’ll help this all make sense.
10/ So..I found a screwdriver at a garage sale once. I liked it because it was super machinic and faceted and built for the fastening and loosening of things. It looked like something an astronaut would take to do repairs on the moon. White. Utility. Art. It had a bit of patina that made it feel like it was an artifact with a life to it. Like it had been around since the Dawn.
11/ It cost half a dollar. I felt like I was ripping the guy off. I bought it.
12/ Something about it made me want to use it more than the other screwdrivers I had. It seemed like it encouraged the work as the work was undergoing. Plus, people admired it when I used it — so that made me feel special. It made the screwdriver feel special. I kept it in a drawer in my tool chest in a felt cloth. It made me want to do more fastening of things. I found myself tightening screws all over the house. A loose cabinet door here. A wobbly chair leg there.
13/ I even found myself using it to wedge things open. It was good with that kind of thing, too. Prying open tins of paint and such. it was especially good at loosening stuck things, like drawer slides, or stuck battery compartment doors. Once, in a pinch, it got me into an old filing cabinet that the salt air had corroded shut.
14/ The screwdriver made me want to do things I never thought of doing before. It changed the shape of what kind of little screwdriver-enabling projects I would take on around the house. Pretty soon, I found myself looking for things to fasten or loosen just so I could use the screwdriver. Neighbors would call me over to help with little repairs. I became the go-to guy for tightening, loosening, and prying things.
15/ One day, while I was tightening a screw on a television mounting bracket, I had a sudden thought: what if this screwdriver is not just a better screwdriver? What if in the act of using it, it is changing the shape of the work? What if it’s changing what it is to fasten things? What if it’s changing the nature of the fastener itself? Even the value of the things I felt compelled to fasten, loosen, and pry. Were they something other than something to be fastened, loosened, and pried?
16/ I soon started to realize something weird — I actually didn’t care about the furniture as furniture anymore. It was more like a set of fastening opportunities that I felt drawn to with that curious, seductive, beautiful, astronaut screwdriver.
17/ Are we sure that AI is just a better screwdriver?
18/ And not something that, in the act of using it, is changing the shape of the world itself? Changing the ontological furniture that we used to think of as stable and given?
19/ I’ll just get to the point: AI is more than changing the work, it is changing the physics of the world itself.
20/ Let me elaborate on that with another story. This one is a science-fictional.
21/ You’re tightening a bolt to fasten a wing to an airplane and as you turn the bolt, the specification for the wing changes.
22/ “Damnit..” you mutter to yourself. “Do I have to get off this gantry and talk to Edgar in engineering again?”
23/ You turn it some more, hoping that the specifications change was just a snaful. But now things become even more unstable: the physics of fastening change such that you can’t be sure if you’re actually tightening, or loosening, or if the wing is even a wing anymore. “What are you even fastening to what?”, you wonder aloud. “My KPIs are based on fastening wings to fuselages, not whatever this is becoming!”
24/ Heck, this is weird, but you keep trying to get that bolt to tighten because well, tighten you must otherwise that quarterly earnings report is going to be as much fun as a kitchen grease fire. But with another quarter turn you find that there’s no reason to keep going because airlines are no longer using airplanes; they’ve sold their brand IP to a hedge fund and are now a global network of steakhouse franchises. Another turn and, lo!, turns out that belief in the physics of aerodynamic lift is punishable by forever banishment, as it is now settled fact that aerodynamics was a tenent of a sham religion orchestrated by an evil cabal of toothy false prophets known as the “Travel Agents”.
25/ The point I’m making here is that the instability — or capability — of the thing we think of as a tool is actually changing the ground rules of the work itself. The physics of the world itself is changing. What it is doing to us and our organizations and institutions is not just a matter of efficiency or productivity. It is changing what it means to know, to decide, to coordinate, to maintain, to learn, to have authority, to have power.
26/ I’m neither exaggerating nor being mystical here.
27/ Neither am I being alarmist nor utopian. I’m actually somewhat curious about what kinds of organizations are possible when AI becomes ambient and just part of the world, like breakfast cereal or buttered toast.
28/ Can we imagine a world where we have these metaphorical screwdrivers that are not just better screwdrivers, but worldmaking mediums that reshape the very ground we walk on? Things we find in a terrain where practices, cultures, organizations are here to amplify human potential, rather than compress it into plausible answers, faster workflows, and extractive efficiencies? Can we prototype organizations that are more capable, more discerning, less captive to whatever interface happens to be in fashion this quarter? How do we design our relationship to AI before the relationship designs us? What does that even look and feel like?