Used properly, LLMs will enable historians to focus more than ever on the truly important stuff: finding the archival nuggets and placing them in historiographical context
Prompting the AI is an art, or at least a craft. I find driving the AI where I want it to go is a good brainstorming technique (in my area of locational political economy).
But I like the way I write, and enjoy the act, so I produce the actual output.
The prompts themselves can be shared, supporting team brainstorming, an activity of network economics. Involving, potentially, pedagogy.
The AI output is, of course, pure deep. Gemini agrees, opining that @Brad DeLong is famous as an expert on Derp…
This is excellent. I've approached LLMs from a similar angle, though focused on how we can use them to improve institutional transparency.
While I agree that those motivated to use AI well will benefit from it massively, I don't think we should be dismissive of the downsides of the ease it provides. It has the ability to supercharge and magnify every critique of modernity we've seen, particularly MacIntyre's observations.
"First, the real, irreplaceable value of the historian will increasingly be found in the time spent trolling in the physical archive." -- this is absolutely correct. I took Richard Hofstadter off the shelf last week for one thing, not remembering his chapter on bad education policy, which I have been writing about! AI would not have given me this link. I agree 100% that digitizing books and papers has been revolutionary. And luckily for us, each text is assigned to a date and an author and a publisher, so we can check and double check, as one does. Great piece!
I really don't understand the reluctance to embrace this perspective - do both. I think in as little as 2-3 years, this debate will seem quaint.
Prompting the AI is an art, or at least a craft. I find driving the AI where I want it to go is a good brainstorming technique (in my area of locational political economy).
But I like the way I write, and enjoy the act, so I produce the actual output.
The prompts themselves can be shared, supporting team brainstorming, an activity of network economics. Involving, potentially, pedagogy.
The AI output is, of course, pure deep. Gemini agrees, opining that @Brad DeLong is famous as an expert on Derp…
This is excellent. I've approached LLMs from a similar angle, though focused on how we can use them to improve institutional transparency.
While I agree that those motivated to use AI well will benefit from it massively, I don't think we should be dismissive of the downsides of the ease it provides. It has the ability to supercharge and magnify every critique of modernity we've seen, particularly MacIntyre's observations.
If you'd like to read more of my thoughts on AI as an accountability tool, here's a link: https://seekingsignal.substack.com/p/the-accountability-machine
Thanks, again, for the great piece.
"First, the real, irreplaceable value of the historian will increasingly be found in the time spent trolling in the physical archive." -- this is absolutely correct. I took Richard Hofstadter off the shelf last week for one thing, not remembering his chapter on bad education policy, which I have been writing about! AI would not have given me this link. I agree 100% that digitizing books and papers has been revolutionary. And luckily for us, each text is assigned to a date and an author and a publisher, so we can check and double check, as one does. Great piece!