The Technology Should Eventually Disappear

Nobody thanks the electrical grid. You flip a switch, the light comes on, and your mind is already on whatever you walked into the room to do. That is not a failure of appreciation. It is the highest compliment infrastructure can receive. The technology worked so well, for so long, that it stopped being technology and became a condition of life.

I have spent more than twenty-five years building and running systems at Apple, AWS, PayPal, AppDynamics, and Fannie Mae, and I have come to believe this is the truest test of maturity for any technology: people stop noticing it. Not because it matters less, but because it has receded into the background where it belongs.

AI is nowhere near that point yet. You can tell because we will not stop talking about it.

When I sit with executive teams, I listen for what they talk about. If the conversation is about the model, the vendor, the benchmark, the prompt, then the work is still in its early, visible, self-conscious phase. That phase is necessary. It is also temporary, or it should be.

We went through the same arc with cloud. In the early years, companies held cloud steering committees and hired cloud czars and put the word cloud in the names of internal projects. A decade later, nobody at a well-run company talks about the cloud much at all. They talk about shipping products. The infrastructure question got answered, absorbed, and forgotten. That forgetting was the finish line.

The same was true of electricity a century earlier. Factories once advertised that they were electrified. Then every factory was, and the word disappeared from the signage. The value stayed. The novelty left.

So what should be left behind when AI fades from view? Only the outcome. A decision that used to take three weeks now takes a day, and nobody remembers to marvel at why. A report that consumed an analyst's Friday now assembles itself, and the analyst spends Friday on the judgment the report was always supposed to inform. A customer gets an answer at two in the morning and never wonders what produced it.

This is the standard I hold enterprise AI work to. Not how impressive the system is, but how quickly people can afford to stop thinking about it. If your teams need to understand the technology to benefit from it, you have shipped a demo, not a capability.

I learned this lesson in a harder way than any enterprise program taught me. I lost my father during COVID. Afterward, going through photographs, I realized what they actually preserve: how someone looked. They do not preserve how his presence felt, the sound of him, the way he told a story only he could tell.

That gap is why I founded The Chronicles, an AI product that helps families preserve stories, voices, and memories. And here is the thing: the families who use it do not care about the AI. They should not have to. What they care about is hearing a voice again, keeping a story that would otherwise have gone quiet. The technology is doing its best work precisely when it goes unnoticed, when all that remains is presence.

That is not a different principle from the enterprise one. It is the same principle with the stakes made plain.

If you lead an organization through an AI transformation, your job is not to make AI prominent. Your job is to make it disappear into the workflow, the way electricity disappeared into the wall and cloud disappeared into the platform.

You will know you are done when the meetings change. When nobody asks which model did the work, only whether the work was good. When the technology stops being the subject of the sentence and becomes the grammar underneath it. The transformation is finished when nobody calls it one.