@vgr
Finished my online book from my twitter archives. With a preface chapter chronicling my twitter journey.
Heaviest lift vibe-coding project I’ve done so far. Started out cutting and pasting scripts from chatgpt, and at some point it started choking and I had to switch to claude code, which was worlds better to the point of magical (but props to chatgpt for getting me going a couple of months ago). Though the final book is pretty conventional looking, the heavy lift was the extreme bespoke processing required to extract and build thread-based chapters, find the most popular tweets for the compendium of singles, fix links and images etc.
Next steps: turn this into epub and pdf versions.
For those interested, here is the repo of my processing code. It’s not really drop-in reusable with arbitrary archives, but might serve as a good reference implementation and starting point for those who want to do something similar, especially if you have a lot of threads.
Repo: https://github.com/vgururao/twitterarchive
Book: https://venkateshrao.com/twitter-book
3 replies
3 recasts
30 reactions