this post was submitted on 15 Jan 2024
24 points (85.3% liked)

Futurology

1812 readers
198 users here now

founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
[–] CanadaPlus 3 points 10 months ago* (last edited 10 months ago)

A link to the report itself.

I'm going to skim through, because I'm really wondering how they decided which things are exposed and how. The abstract makes it sound like they just went by sector, and while that's a good simple option, I think it would miss a lot of important nuance. LLMs are set to totally wreck advertising copy writers, but only feed electrical engineering jobs. Both just look like a college education at the very coarse level.

Edit: So, they've broken down jobs into low-exposure, high-exposure low-complementarity and high-exposure high-complementarity, which includes all the prestige-y occupations. Then they actually do look at the fine-grained occupations level.

The one thing I find suspect is the complementarity, which I'm still looking for the sources on. It seems to me a lawyer is just as likely to be laid off as their assistant, if not more, since most of them just sit behind paperwork all day anyway, while the assistant has a name and a personality. They offer some alternate scenarios in the appendix because I guess I wasn't the only skeptic.

In the model with no complementarity, inequality actually collapses because capital income doesn't increase as fast as it decreases for the fancy occupations. It has problems too though - overall income goes down, which doesn't really make sense.

Edit 2: Here's their source. Basically, they consider which occupations are likely to be protected by legal and social factors, and then just literally adjust for prestige ("job zone") because they figure more prestigious people will have better opportunity to adapt. Personally, I suspect they underestimate the market's tendency to cut all corners, and overestimate their own adaptability. Thank you to anyone that actually read this far, haha.