Wow, well done! It's quite rare to see such a well-cited technical piece, and the review paper you linked on creativity is a goldmine of citations.<p>Some critiques and compliments:<p>1. Not sure I love the emoji choices for the buttons. This is super arbitrary, but I just don't see the connections, especially for the spiral and the sun. Perhaps there are icons from free online sets more directly related to the ideas General, Concrete, Abstract? Say, arrows-out, arrow-down, and arrow-up?<p>2. IDK what "SOS" does and the tooltip is mistaken, should probably clarify that.<p>3. The graphic is surprisingly durable, as someone who loves making such things I'm quite impressed. Handles quick entries and such well. That said, it desperately needs a loading indicator of some kind to confirm that the query went through successfully, a way to delete/clear nodes you don't need, and a more manageable initial layout when you're typing words in. As is it seems to spread them across a large area pretty much randomly, which is understandable from a technical standpoint but obviously isn't ideal. Maybe start with a circle or something?<p>4. There's some tomfoolery going on with capitalization -- I know this is built into LLM tokens from the jump, but honestly I'd endorse a blanket `.lower()` on all input. I stopped using the tool and came to write this comment after I saw `Syntax` was connected to `semantics`, but my other input, `syntax`, wasn't.<p>5. IMHO webpages without headers are suspicious. Plus I bet basically every user of this would be interested in reading your blog posts about it!<p>6. I know there's basically infinite cool things you could add to this (that's what makes a good tool, probably!), but the one that I have a need for right now is a "connection" button for two terms that are connected in my mind/project, that aren't yet connected on the wordscape. I could also see this working based on areas/quadrants/buckets, where connections would be sought when the user drags+drops a new term in with some existing ones.<p>7. Should probably just let the graph be empty after reset. I love the cuteness, but it's a bit confusing IMHO, and ties into the "delete nodes" feature request above.<p>8. I would love to see a blog post on how you would compare and contrast this tool with a thesaurus, which is my usual goto for this task. Might also be a good place to draw UX/I inspo.<p>Sorry for the rant, hopefully you have an opportunity to work on this fulltime because I love it. Despite all my complaints, this couldn't be more perfect for the kind of work I'm doing these days (philosophy-based architectonics) -- instant bookmark! Please do hit me up if you're ever looking for beta testers, collaborators, blood donors, fanatical followers, you name it ;)