Author here, thank you so much! I really tried to make it nice to use, beyond the (quite original) modelling.
A prototype I did tried to detect some grammatical constructions, eg "it's not ..., it's ...", but I am not sure how to systematize that.
Also just a disclaimer: I am NOT tracking Claude tics, I am merely finding that a particular cluster of vocabulary increases. Tracking Claude requires labelled data IMO. I tried using model release dates in a structural model to constraint the clusters but the result was not compelling, so I ended up simplifying the model a lot!
I love this. I think you should be clearer about that - it's obvious to people who have experience doing this kind of data analysis, but a lot of people I've shown this were confused, as in "huh, load-bearing and genuinely makes sense but how'd he pick all these other words!?"
I think it should be made clear that the sorting is sort of Texas Sharpshooter-ish - the ones on top are on top because they sort that way.
The fact that load-bearing ends up on top is the proof that this works because we all know a priori that load-bearing is a Claudism.
Tangentially related. I built a little app to help me find art for my home art frame (Meural) -- it searches through the Cleveland Museum of Art, Art Institute of Chicago, and Museum of Modern Art APIs.
Google Reader was killed by Vic because it was the only "social" app at Google at the time. He wanted to grow their FB competitor (G+) and insisted on migrating users over from Reader to G+, which never happened.
Agreed. Logseq changed the game for me. Because SRS is closely coupled with note-taking and is as simple as adding "#card", it eliminates any friction and excuses I had for not using SRS. I look forward to adding new cards as much as I do reviewing old ones.
Another use I'm discovering is using logseq in this way for both cards, and annotating explanations of things that can be put back into other learning content.
Nitin discuss from a fresh perspective what it will take to prevent climate change. He distills and synthesizes books and articles into short essays that push my thinking forward.
Thanks for sharing. This was a compelling argument that nudged my understanding in a new direction.
I recently started paying creators on Patreon and subscribing to news outlets I wanted to support. But within a year, I realized that this was inefficient, I didn't like/read/watch most of the content that my subscriptions were supporting. As much as I wanted to support the content creators and quality journalism, I questioned the value of my subscriptions and cancelled - I felt like I signed up for a gym membership on Jan 1st that I wasn't using anymore.
Ads allow me "to pay" with my attention for only the content that I value. I don't like ads, I generally use an ad-blocker, but I appreciate the post and the perspective.
Using Google Reader felt like hanging out and having a beer with my favorite friends. It was teeming with conversation threads waiting to be pulled.
Seeing a new share was like getting a tiny intellectual present, filled with all the right connotations - "Here's this thing I think you would find interesting," "I care about your opinion, what are your thoughts on this," "It's not important, but it is entertaining," "not trying to share this with the world, just you, my friend."
Google Reader was our company's post-lunch ritual. A way to balance a mornings worth of engineering effort and soporific pad thai with a jolt of fresh news and virtual office banter.
Is it possible to expand this analysis beyond words to other Claude ticks? Contrastive framings, sentence length, caveating, for instance.
reply