Toby Nangle Dec 16, 2022

I've experimented with using ChatGPT as an intern. It's not awful awful, and has trained me in being much more precise in what I want it to do.

Furthermore, it's pretty easy to see its limitations eroding over time: giving it the ability to review and summarise literature on a topic rather than just its (fairly basic/generic training data) is surely within reach in the short term and would make it much more useful.

/1

Show threadToby Nangle Dec 16, 2022

With development, ChatGPT could be a useful research tool like Google search or Wikipedia. Google search has changed the world radically, but it's not obvious that this change has manifested in obvious changes to the labour market.

Rather than seeing it as a threat to people's jobs I'm more inclined to see it as a productivity enhancement that will increase consumer surplus.

Show threadTom BowkerDec 16, 2022
@tobyn as ever, though, this will require deliberate policy-making, to avoid the benefits just accruing to the rich and just exacerbating inequality.
Show threadToby Nangle 
@TomBowk In a kind of Elysium scenario? I can see that's one path, but wonder how this is so different to Google search as a tool.
Dec 16, 2022 at 9:37amWeb
Show threadTom BowkerDec 16, 2022

@tobyn You see, if it wasn't for Google I'd have had to hire someone to explain what Elysium is!

I don't know, do Luddites tend to be proven wrong in the long term?

I wouldn't oppose progress anyway I guess. And would anyway argue in favour of Pro-equality econpol...

Show threadToby Nangle Dec 16, 2022
@TomBowk
Maybe I'm too relaxed about the whole thing. This was something I wrote a while back - perhaps the biggest take-away of which is that Elysium is a pretty good action sci-fi flick about inequality, apartheid, communitarianism vs cosmopolitanism, etc. https://principlesandinterest.wordpress.com/2015/11/12/policy-making-in-an-elysium-scenario/
Policy making in an Elysium scenario

principlesandinterest