Great stuff. I've been doing some digging in this period of American history recently and it has lead me to brainstorm about similar applications of modern AI for historical analysis or presentation.
I think there's a lot of newly ripened low-hanging fruit right in fields like History now that LLM tools have either digested a lot of historical sources or can be used to build systems such as this for presenting that data in new ways that engage the reader.
One of the ideas I was thinking about was a map for presenting the provenance of an object or an idea.
It would be neat to be able to see a map that had markings on it that showed where someone first had an idea -- the apocryphal 'apple falling on their head' moment and then where they wrote their first draft -- maybe they mailed that draft to a friend who offered editing suggestions, then they travelled with a new draft to a distant town because of a neighbouring country invading where they were, then they sent it off to a printing press where a small run of pamphlets was made and they were distributed across a region to a few cities.
Seeing all those key moments for the genesis of an idea, invention, or object in a map with play/pause rewind/fast forward controls + the ability to put other things on the map ot watch their spread and visualize the flow and interplay of information over time would be fascinating and could lead to some really good insights.
I love the idea of a map to study provenance. For example, consider the phrases “legislative power,” “executive power,” and “judicial power,” appearing in Articles I, II, and III of the constitution. The first is defined by reference to the subsequent article. But the second and third are not defined in the constitution. Apparently those terms were used as such in writing during the English Civil War, more than 100 years before the constitution was written. It would be very cool to have a tool that could trace these ideas back in time.
These are great thoughts. I may do some investigation into how the Federalist Papers spread, but even with AI, these may be significant research efforts. I need graduate assistants!
Lovely project, I've given it a skim and I quite like the presentation.
The only thing that I would work on would be to rewrite a lot of the website copy since it's a bit heavy with the llm-isms, which gives it an unfortunate sheen.
If theres a repo link, I'd be happy to contribute, since I realize it's no small task.
Completely a fair point. Definitely was heavily, heavily AI built (including the copy) - there is no way I could have justified putting in the time it would have taken to build this for a "pro-bono" project pre AI! Even with AI, I've spent a considerable amount of time steering, editing, and revising (including some of the most egregious AI adds).
Would absolutely appreciate edits and revisions. Repository here:
Really hugely appreciated. I'm doing an AI pass at removing some more of the AI-ism's. It came up with some good suggestions. I'm reviewing and should have those in before long. I'm sure it won't be complete, but might give a better starting point.
Thank you! AI clearly helped a ton, but the number of rounds of me having it render different fonts / formats was staggering. And, I was fortunate enough to come across justif (https://justif.lyall.co/) - I found via HN! ... It perfectly implemented the system of justification I needed.
I made this. It started as a weekly reading project with my mother: one Federalist paper a week. I couldn't find an online edition that was actually pleasant to read - most are walls of text or scanned PDFs - so I built one.
It's all 85 essays with the original wording (text from the Project Gutenberg edition, publication dates checked against the Library of Congress record and Founders Online), set the way they first appeared: as essays in the New York newspapers of 1787–88, under the pen name "Publius." There's also a clean modern reading mode if the facsimile isn't your thing.
The part this crowd might like: the justified text uses Knuth–Plass line-breaking (via a small library called justif - which I found here on HN!), with hanging punctuation and optical margin alignment, set in the IM Fell types. No ads, no accounts, no tracking; reading progress is stored locally on your own device.
It's free and the text is public domain. I'd genuinely welcome feedback on the reading experience.
(It was recently featured on kottke.org, which is the only reason I'm brave enough to post it here.)
Great stuff. I've been doing some digging in this period of American history recently and it has lead me to brainstorm about similar applications of modern AI for historical analysis or presentation.
I think there's a lot of newly ripened low-hanging fruit right in fields like History now that LLM tools have either digested a lot of historical sources or can be used to build systems such as this for presenting that data in new ways that engage the reader.
One of the ideas I was thinking about was a map for presenting the provenance of an object or an idea.
It would be neat to be able to see a map that had markings on it that showed where someone first had an idea -- the apocryphal 'apple falling on their head' moment and then where they wrote their first draft -- maybe they mailed that draft to a friend who offered editing suggestions, then they travelled with a new draft to a distant town because of a neighbouring country invading where they were, then they sent it off to a printing press where a small run of pamphlets was made and they were distributed across a region to a few cities.
Seeing all those key moments for the genesis of an idea, invention, or object in a map with play/pause rewind/fast forward controls + the ability to put other things on the map ot watch their spread and visualize the flow and interplay of information over time would be fascinating and could lead to some really good insights.
I love the idea of a map to study provenance. For example, consider the phrases “legislative power,” “executive power,” and “judicial power,” appearing in Articles I, II, and III of the constitution. The first is defined by reference to the subsequent article. But the second and third are not defined in the constitution. Apparently those terms were used as such in writing during the English Civil War, more than 100 years before the constitution was written. It would be very cool to have a tool that could trace these ideas back in time.
These are great thoughts. I may do some investigation into how the Federalist Papers spread, but even with AI, these may be significant research efforts. I need graduate assistants!
Lovely project, I've given it a skim and I quite like the presentation.
The only thing that I would work on would be to rewrite a lot of the website copy since it's a bit heavy with the llm-isms, which gives it an unfortunate sheen.
If theres a repo link, I'd be happy to contribute, since I realize it's no small task.
Completely a fair point. Definitely was heavily, heavily AI built (including the copy) - there is no way I could have justified putting in the time it would have taken to build this for a "pro-bono" project pre AI! Even with AI, I've spent a considerable amount of time steering, editing, and revising (including some of the most egregious AI adds).
Would absolutely appreciate edits and revisions. Repository here:
https://github.com/vwalke/federalist-papers-reader
Thanks for the feedback! Vann
I totally understand, thanks for sharing the repo, I'll do my best to pitch in. Always happy to support making history more accessible!
Really hugely appreciated. I'm doing an AI pass at removing some more of the AI-ism's. It came up with some good suggestions. I'm reviewing and should have those in before long. I'm sure it won't be complete, but might give a better starting point.
Thanks again!
Very well done, love it!
Thank you, this is really cool. Lovely!
Thanks for the kind words. It has been fun to put together - Hopefully a good addition to the world.
this is excellent typesetting
Thank you! AI clearly helped a ton, but the number of rounds of me having it render different fonts / formats was staggering. And, I was fortunate enough to come across justif (https://justif.lyall.co/) - I found via HN! ... It perfectly implemented the system of justification I needed.
yeah who cares what tools you used to make it: this is a beautiful reproduction
Hi HN,
I made this. It started as a weekly reading project with my mother: one Federalist paper a week. I couldn't find an online edition that was actually pleasant to read - most are walls of text or scanned PDFs - so I built one.
It's all 85 essays with the original wording (text from the Project Gutenberg edition, publication dates checked against the Library of Congress record and Founders Online), set the way they first appeared: as essays in the New York newspapers of 1787–88, under the pen name "Publius." There's also a clean modern reading mode if the facsimile isn't your thing.
The part this crowd might like: the justified text uses Knuth–Plass line-breaking (via a small library called justif - which I found here on HN!), with hanging punctuation and optical margin alignment, set in the IM Fell types. No ads, no accounts, no tracking; reading progress is stored locally on your own device.
It's free and the text is public domain. I'd genuinely welcome feedback on the reading experience.
(It was recently featured on kottke.org, which is the only reason I'm brave enough to post it here.)