I’ve been using AI tool RootPainter for a while to get root and nodule data from photos of nodulated roots, which would otherwise be a very laborious process.

I’ve been using Perplexity.AI for most web searches recently, since it usually finds and summarizes the information I want, rather than giving a bunch of less-useful “sponsored” links. Perplexity keeps adding useful features, like voice input/output on my phone and their new Spaces option, although the latter isn’t quite as useful as they make it sound, yet. Yes, it will save up to 50 uploaded files of up to 25 MB each and answer questions about them, with or without additional information from the web. I can share the space with students in my lab, so they can query papers I upload, protocols, lists of treatments by plant, etc. But it isn’t always clear how *much* of a document it examined, especially with longer ones. For example, it could apparently only “see” the first 188 entries in my text file of citation information on 14,000 references. And, yes, I could “share” the Space, but we can’t see each other’s threads within the shared space. I guess that reduces clutter and I wouldn’t want to share *all* my Spaces with everyone, but it limits collaboration options. I’m hoping Perplexity will keep getting better, though, and consider $20/month for the paid version money well spent.

Looking for an AI that could handle larger files, I’m exploring Google’s NotebookLM. It can’t quite fit my reference file either, but it was able to find material in the full text of my book, and clicking on a reference number in the summary takes you to the original text. It’s most-talked-about new feature is its audio-summary option, where it generates a fake conversation summarizing a document. I uploaded a lab notebook — a text file (with many errors) generated from handwritten Livescribe-pen notes by Myscript, which is no longer available — and it came up with this. It’s mostly accurate. I’m amazed how the AI turned a bunch of disorganized (unrelated topics on the same page, sometimes) handwritten lab notes into a possibly-engaging if perhaps overly dramatic story about my heroic struggle to do research in my basement during the pandemic. (I wonder if it would do an equally sympathetic treatment of Mein Kampf.) I’ll be trying it on scientific papers I’m somewhat interested in but don’t have time to read.

I’m still writing everything myself, though I did ask Perplexity for comments on my latest grant proposal.