EDIT: these have been updated in my subsequent post HERE
I think it would be awesome in 5 or 10 years if we lived in a substantially more automated world, brought about by large language models (LLMs). In the long run, more automation means higher productivity per worker hour, which translates into higher wages and a higher standard of living. This is good.
However, I am a skeptic that LLMs will power a productivity revolution. My skepticism is rooted in the recent past: the Internet was hailed as revolutionary. And it has changed a lot of things. But it’s not like the last 20 years have been stellar in terms of productivity (see chart at BLS).
I am, however, rooting for productivity gains. They are great. In Q2 of 2023, we actually had stellar productivity growth of 3.7%.
To keep myself honest, I want to register some Quick Takes (QTs) here. As a reminder, QTs come with a statement that can be clearly deemed right or wrong, a time frame, and a percentage I think I will be right.
EDIT: these have been updated in my subsequent post HERE
Productivity growth in the US will be, on average, less than 2% over the 2024-2029 period (60%)Whatever the actual US productivity growth number over the next 5 years, LLMs or similar technologies will not be a consensus major factor in productivity growth, as measured by the consensus in the economics literature (70%)
Anyway, I hope I’m wrong but I just feel skeptical, so I’m registering my vibes here.