It was 100x more productive at doing this than I would have been.

Maybe the 100th or 1000th attempt will get somewhere new!

This looks like AI is finally approaching human intelligence levels!

you are vibing at the wrong frequency

wub wub wub <- correct vibe frequencyyou're welcome

womp womp

ChatGPT does this over and over. I point out a mistake, it apologises, promises to correct, then spits out the same mistake.

I've also seen several cases on various models where they get stuck in a tradeoff, where they will fix a problem by creating another one... and then fix that problem when prompted by putting the first problem back

@Andrew Beyer Yes, and I used to do this as well for a long time. If you have n bugs to avoid in a system (Christopher Alexander calls them misfits) but your design process can only handle n-2, ping-ponging can result, often over a time of months as a bug gets reported in production, gets fixed, the other bug gets created, gets reported in production..

A lot of the value of tests for me is shortcutting this sort of ping-ponging between bugs. But if you tell AI to write the tests, and give AI carte blanche to modify tests at any time.. 🤷‍♂️

I found it an even bigger issue when it was more "soft"/nonfunctional issues that maybe couldn't be easily verified w/ automatic testing

Seems like we need two AIs, each watching the other. Maybe one can write the tests and the other the code

Indeed, and I am somewhat surprised that this isn't done yet, given how important the idea of adversarial training has been in the short history of deep learning.

Ivan Reese are you telling me llms vibe in dubstep?

my new system prompt: you are at a skrillex show waiting for the drop...

Konrad Hinsen isn't this what reasoning models do, effectively?

Ivan Reese According to my understanding, no. They are trained on reasoning stories written by humans. That's a form of supervised learning, whereas adversarial training is unsupervised: two AI models confronting each other. Or even a single model switching sides, as was done with AlphaGo.

General reasoning models need supervision because there is no obvious arbiter for deciding if a reasoning is correct. For code, "compiles, runs, passes tests" provides three consecutive automatable arbiters.

this one? github.com/ivanreese/visual-programming-codex

if you mean literate review as in paper, I don't know 🙂

ooh, yeah, that wasn't the one I was thinking, which was more of a paper style presentation...but that one's actually maybe even closer to what I need

If you find the one you were thinking of, open an issue so I can add it to the codex!

I think I know the one you mean, maybe. It was a massively heavy page packed with screenshots? [update: ah, no, not a "paper" style at all!]

blog.interfacevision.com/design/design-visual-progarmming-languages-snapshots

this one - is that it? [update: no, but it's a fun page!]

found this quite interesting article along the way:

drossbucket.com/2021/06/30/hacker-news-folk-wisdom-on-visual-programming

My personal favourite (couldn't find it on the wiki) is the Monty Python's one: grotesque retro cutouts.

I have been thinking recently that it’s probably time for a more elaborate, decorative style of UI to have its turn.

I’m working on an app atm, so I might try something there.

what would a baroque user interface look like?

I’ll need to experiment. I’m thinking subtle animation in the background, on borders and such. Maybe some more elaborate filigree? I’m planning on looking at video games for inspiration.

Also subtle background sounds. I think sound is greatly under-used in apps. If there were subtle background sounds indicating the state of an app, the user would come to learn them and be able to somewhat tell what an app was doing without looking.

The intent is to increase the delight in using the app, and to try to discover ways to improve usability. For example, I’ve a notion that you might use background animation behind the controls to indicate what just affected what when the user changes something.

Partly vibes, at the moment, I admit.

Vibes are essential. Your description reminded me of an older Windows theme: youtu.be/-Xl08y3GoRg

I think trying to map some of the less tech-related aesthetics to tech might require a lot of vibe-based thinking.

you reminded me of one of my favourite game UIs

Material gets out of the way but it’s so dull .

This is nostalgia speaking, but I really do miss Y2K aesthetics. Oh how I wish I knew how to make Aqua-style lickable buttons...

Looking at the wiki, I see that I also quite enjoy what this wiki calls Frutiger Aurora... Swimming in the digital sea is quite nice - and I actually wonder how this could work in AR/VR.

not mine, but had this in my bookmarks: codepen.io/simeydotme/pen/MWXxKrP

1-bit pixel art, but: the pixels are in free space, and they can be any size.

I've been browsing this all day in free moments, and I don't think I'll be stopping. That is all.

I’m not getting it from that wiki, but I’d love to make an Art Deco interface someday. Something like this.

(I started thinking about aesthetics of programming systems after making a swords-and-sorcery adventure-game programming system: vezwork.github.io/drostes-lair-post)

You might enjoy moldable development then.

For sure, I am a rabid fan of your work

I am glad it resonates. Please feel free to use that term then 🙂

Did/does the project succeed?

Did you reach the goal you wanted?

that was not my work.

my work is this which does use the term moldable but is also a WIP.

ah, I see. I apologize 🙂

I should have posted in #linking-together now I realise

Nice!