Just so I understand: this HN post consists of a 6-sentence plain English prompt, and a web app that passes the prompt plus presumably raw user input to a black box LLM API? Is there more?

The Mona Lisa consists of six paints, all jumbled together? Is there more?

Of course there's more to it but someone taking a picture of the Mona Lisa with their smartphone doesn't add much to it and certainly doesn't commend the same respect as Da Vinci painting it in the first place.

I see what you're getting at, but in your analogy, Leonardo da Vinci (the creative and intellectual driver of the overall work) maps to the black-box LLM behind an API, where the boilerplate web app wrapper maps to the Mona Lisa's constituent oil paints, and perhaps also the backing poplar panel and frame. In other words, wrapping an LLM is like updating the frame around the Mona Lisa.

Well, there's also the huge LLM which is probably ~70-100 billions of parameters (just estimates for 3.5 Sonnet), running on a farm of GPUs that cost tens of thousands of dollars. But in summary you're correct :)

Welcome to the LLM API wrapper future. This is basically what 99% of the "AI startups" do.

Just so I understand: you’re talking about setting up an FTP account, using curlftpfs, and SVN/CVS for Linux users? And even with all these, you’d still need USB drives for connectivity issues? Plus, you're naming it Dropbox? Is there more?

There was more to what dropbox was to FTP/CURLFTPFS/etc. then than what this webapp/page is to a Claude API now.

You got it correctly, there isn't anything more there.

Welcome to 90% of all AI startups born this year. It's both a testament of how powerful LLMs have become, and how wonderful it is that someone else is footing the bill for the massive amount of resources they consume.