Why no "societal impact" two years after ChatGPT?

This is more of a rambling through based on multiple “where’s the impact?” opinions getting thr

I’ve seen many variations of this take elsewhere, but this is the most blunt (part of why this gets twitter engagement): https://x.com/kareem_carr/status/1874883243352457723

It’s crazy how earth-shattering progress in building thinking machines has created nearly zero societal impact. No significant positive effects on society at all, after almost 2 years, and the biggest negative effect is just kids are cheating on their homework more.

I am not well-versed in the history of science and technology, probably much less familiar than I should be for someone working in tech. But these takes seem willfully ignorant, as though there’s a long list of universally-deemed-beneficial technologies that were immediately apparent an adopted by all. We are talking about two years!! That used to be how long it might take everyone in one research field to learn about an invention from their peers.

Even my best attempts to find an invention with a 2 year “societal impact” are failing¹. The closest I got was penicillin, but that had more than a decade of lag between invention and mass usage. Other older inventions, when historical diffusion of ideas was much slower, have no chance to compete. Even modern medical inventions face more regulatory barriers that make a sub-2-year impact hard.

Basically, this question smuggles in an ignorance of invention history, where the end result has a neat narrative and seems obvious in hindsight, then asks “so what’s wrong with this AI stuff? why so much hype?”. In reality, the fact that more than a billion people have chosen to use it within 2 years already makes it one of the most impactful technologies.