There is a story that gets told about artificial intelligence that presents it as something genuinely unprecedented. A technology so fundamentally different that the old frameworks do not apply, the old regulations cannot reach it, and the old human responses — wonder, fear, opportunism, scepticism — are somehow insufficient.
I do not believe this. Not because AI is unimportant, but because I have read enough history of technology to recognise the shape of the narrative.
The pattern
Each generation of computing technology goes through the same sequence. First, a new capability arrives that appears to cross a qualitative threshold. It does not merely do the old thing faster; it does something that was not possible before, or was not practically possible, or was not possible at this cost.
Second, a period of genuine excitement. The people who build with it understand that it changes things. They are right. They are also often wrong about which things it changes, in which direction, and how quickly.
Third — and this is the part that gets forgotten — the technology becomes infrastructure. Not special. Not magical. Just another layer in the stack that people depend on without thinking about.
This happened with the personal computer. With the internet. With the smartphone. With cloud computing. Each time, the discourse moved through the same stages: wonder, overestimation of short-term impact, underestimation of long-term structural change, normalisation.
What is different this time
There are some genuine differences with AI, and it is worth being honest about them.
The speed of capability improvement has been unusually fast. The breadth of the affected domains is unusually wide. The degree to which the technology works by producing outputs that look like human outputs — text, images, code, speech — creates specific challenges around authenticity and attribution that previous computing technologies did not create in the same way.
These differences are real. They deserve serious attention.
But the social dynamics around the technology — the hype cycles, the moral panics, the investment stampedes, the breathless conference talks, the think-pieces warning of civilisational collapse followed six months later by think-pieces explaining that actually nothing has changed — these are not new. These are the entirely predictable responses of a society encountering a new technological capability that it does not yet know how to place.
The builder's position
From where I sit — as someone who builds things with this technology rather than writing about it professionally — the practical question is different from the philosophical one.
The philosophical question is: what does this mean? The practical question is: what should I build, for whom, and to what standard?
And the practical question does not change very much depending on whether the current moment is historically unprecedented or merely a variation on a familiar pattern. Either way, the work is the same. Understand what the technology can actually do, at what cost, with what reliability. Build things that work. Do not build things that do not deserve to exist.
The same shit
I have been building software long enough to have watched several "everything will change" moments arrive, partially fulfil their promises, and settle into the background.
Each time, the things that changed were real. Each time, the people who predicted that nothing would change were wrong. Each time, the people who predicted that everything would change in the specific way they specified were also wrong.
What I have learned, imperfectly, is to take the technology seriously while maintaining a certain scepticism about the narratives constructed around it. The technology is often more capable than the sceptics believe. The narrative is almost always more inflated than the technology justifies.
This is the same shit. Different day. Worth paying attention to, for exactly that reason.