The solution is clear: yes we use AI, and we're okay with it.
There's no way Claude/AI recreated another persons app exactly, with the same name and branding, without being told to do so in some way. I just don't believe it.
And him not using it and calling out AI is just him covering his ass
> There's no way Claude/AI recreated another persons app exactly, with the same name and branding, without being told to do so in some way.
If you give a description of the app you want, and it matches another persons Open Source app. Then why wouldn't Claude do what it does every other time and pull from it's corpus of knowledge?
Now I would expect the human-in-the-chain to verify what is going on and ultimately they are responsible for any output they lay claim to.
In this case, AI assisted or not, it is a dick move. TFA claims they are backing out of this space completely - which is the correct thing to do.
But as with all "my punch simply slipped!" apologies their reputation is tarnished and they will need to demonstrate they are better in future - not simply say so.
> Then why wouldn't Claude do what it does every other time and pull from it's corpus of knowledge?
I dont know why, but I do know know that it doesn't. I know because I've tried similar things when playing around with AI to see what it could do. "Make an app similar to X app [link to repo]"
Depending on how its feeling, it may just read the readme or it may go through the codebase. But I've never gotten line by line code copying.
> I would expect the human-in-the-chain to verify what is going on
I agree with this, to the extent of your second point
> they are responsible for any output they lay claim to.
How far does the rabbit hole go for verifying what is going on? If the LLM did indeed copy an entire codebase at its own behest, what reason would the author have to verify that it wasn't copied, given that every other time they've used an LLM to code it never copied an entire codebase?
A bit outdated, but IME never copies things verbatim because of limitations in context. Even when you want it to be a verbatim copy, like in a migration, it struggles a bit to remember everything exactly from memory and always has to cross-reference.
Might be better in newer models though.
But for a real "copy" it would default to calling `cp` in bash.
> The solution is clear: yes we use AI, and we're okay with it.
Except that some portion of D'addario's customers--presumably those that actually buy their guitar strings to create music via real guitars--aren't okay with it.
Yeah, I think that case is absolutely wild, because ostensibly the point of the recorded music was to showcase the strings.
Its telling that instead of showcasing the quality of their strings, they used an LLM to generate the sounds their strings should make.
Shows how much trust they have in their own product.
Wouldn't this be basically fraudulent advertising?
Maybe they prompted "make it sound like the guitars have D'addario strings"
AFAIK, a lot of commercials, particularly for cars, haven’t really shot the physical product in decades and instead used CGI. I think Ikea also makes heavy use of renderings. It’s all kind of insane. OTOH, a hundred years ago ads all used illustrations… I guess you’d have to show that the real strings don’t sound or behave as shown in the video, however it was created.
In this case there's a difference between CGI and generative AI. However they came up with arrangement of pixels on the screen, the product is still portrayed accurately.
Now, unless the company in question got their new strings into Sona training data, I wager that the generated output has little casual connection with the product being advertised. If I found out IKEA's newest product looks nothing like on the commercial, because the ad was using generative AI and hallucinated something approximating the product, I'd be looking at it as fraudulent too.
IMO in D'addario's case, the analogy would be like using CGI and illustrations to illustrate a painting.
Also, I think when judged by the court of public opinion, the overwhelming answer was "yes, it's fraudulent", or at least "misleading and useless" as most people even complained that the song sounds like an overproduced mess where any string would sound the same.
I don't understand why you're getting downvoted, but you're correct. That's the reason for the complaints and the company has already taken an anti-AI stance precisely for this reason.
It’s not worth getting upset at one or two people downvoting. This is a large forum. There will always be some of that. Plus we have no clue why they did it, it could be valid
I'm not upset, that's not the important part of my message.
> There's no way Claude/AI recreated another persons app exactly
Entropy doesn't need to be set to zero for legally significant duplication