The museum edition asks for a passcode on any UID given, but no passcodes are provided? Other than that, it's very interesting, as long as you skip a lot of the uninteresting "stats" that LLMs think have to be part of every webpage, and skim the 5 or so pages of chatgpt-written history about it that has very evidently been prompted from information given by the sysop's son, but tries to do a weird third-person gimmick that is very distracting.
I'd honestly just prefer it if you'd written a 2 paragraph history yourself. It comes off as very insincere otherwise.
As a whole, the archive and webpage are very difficult to navigate, too. I've seen a lot of these sorts of live-archival projects over the years, and as best as I can tell, a lot of the context clues that normal archivist projects (or even, you know, humans that design web pages) give you to help prompt you for what to expect are missing. It's also very distracting how the entire thing is written and formatted like a VC sales pitch.
edit: I just attempted another go at the "museum edition" version to see if I had missed something, but pressing "skip" to get past the intro caused the entire thing to seemingly halt completely. Any further input prompted the "if you make a mistake, press pound" sample to play, and then it prompted me for a password. I tried yet again, and it happened again even after a cache-clearing refresh. I tried again, and just waited it out, but pressing the star key caused it to enter the "connection closed" state.
This... As someone who has broadly avoided LLM software, this entire site is one of the buggiest things I've ever used. It needs a third pass just to make it bug-free and intelligible to someone who isn't involved in the project, I think. And maybe throw $20 at a programmer you know to hand code stuff instead.
The museum edition asks for a passcode on any UID given, but no passcodes are provided? Other than that, it's very interesting, as long as you skip a lot of the uninteresting "stats" that LLMs think have to be part of every webpage, and skim the 5 or so pages of chatgpt-written history about it that has very evidently been prompted from information given by the sysop's son, but tries to do a weird third-person gimmick that is very distracting.
I'd honestly just prefer it if you'd written a 2 paragraph history yourself. It comes off as very insincere otherwise.
As a whole, the archive and webpage are very difficult to navigate, too. I've seen a lot of these sorts of live-archival projects over the years, and as best as I can tell, a lot of the context clues that normal archivist projects (or even, you know, humans that design web pages) give you to help prompt you for what to expect are missing. It's also very distracting how the entire thing is written and formatted like a VC sales pitch.
edit: I just attempted another go at the "museum edition" version to see if I had missed something, but pressing "skip" to get past the intro caused the entire thing to seemingly halt completely. Any further input prompted the "if you make a mistake, press pound" sample to play, and then it prompted me for a password. I tried yet again, and it happened again even after a cache-clearing refresh. I tried again, and just waited it out, but pressing the star key caused it to enter the "connection closed" state.
This... As someone who has broadly avoided LLM software, this entire site is one of the buggiest things I've ever used. It needs a third pass just to make it bug-free and intelligible to someone who isn't involved in the project, I think. And maybe throw $20 at a programmer you know to hand code stuff instead.