LLMs and the Handwriting Recognition Backlog
[This post assembled from a BlueSky thread and LinkedIn posts from December 2025.]
I'm not a big AI hype guy, but this feels like one of those moments where a really thorny and previously time-intensive/expensive problem just became much, much more straightforward: generativehistory.substack.com/p/gemini-3-s...
And I'm of course not the first to draw attention to this, I've been following @benwbrum.bsky.social and Sara Brumfield write about this related to From the Page, and likewise following @dancohen.org write about this in a big-picture way: newsletter.dancohen.org/archive/the-...
But it really feels like libraries & archives as a field suddenly just went from "we aren't generally attempting to do automated handwriting recognition because it's at the edge of what's possible" to "oh boy now we have another doable but labor-intensive collections enhancement task on the backlog"
For about one minute, it felt like I couldn't be wowed by new releases, but now I'm wowed again!