I’m reading the report AI as Normal Technology. I haven’t finished it yet (it’s long!) but its thesis is that, despite all the noise and hype, AI is progressing on the path of every technology before it, even the highly transformative ones such as electricity and the internet, and doesn’t need to be thought of as a special case.
This is the most important chart from it:

There has been (and still is) a lot of invention, a lot of innovation, and a lot of adoption; 58% of employees intentionally use AI at work, according to this survey. But we’re still in the early stages of diffusion: it hasn’t become so integral to the ways we work that we’ve adapted our workflows to make the most of it.
And that’s apparently confirmed by a new study, Large Language Models, Small Labor Market Effects, that shows that while a lot of people are using AI tools, they’re mostly only using them for small tasks, and there are currently little-to-no economic gains shown.
Cynics have jumped on this as an Aha! moment, proof that AI is useless and that they were right all along. But that’s not right; the chart above indicates that we’re still in the early adoption phase and that adaptation will come — just not as fast as the boosters would have it.
The productivity benefits [of electrification] took decades to fully materialize. Electric dynamos were “everywhere but in the productivity statistics” for nearly 40 years after Edison’s first central generating station. This was not just technological inertia; factory owners found that electrification did not bring substantial efficiency gains. What eventually allowed gains to be realized was redesigning the entire layout of factories around the logic of production lines. In addition to changes to factory architecture, diffusion also required changes to workplace organization and process control, which could only be developed through experimentation across industries. Workers had more autonomy and flexibility as a result of the changes, which also necessitated different hiring and training practices.
In my presentations I talk a lot about capability overhang: the technology is capable of more than we can yet imagine. We have these powerful new tools but we’re mostly just applying them to the ways we already work instead of imagining different or completely new ways to work. We need to redesign the layouts of our factories.
One reply on “AI as Normal Technology”
@blog
AI is till basically a fancy research project trying to produce some useful prototypes. Its application , except for robotic tasks(pattern matching), is non-existent. Science NOT understood.