13 Comments
Aug 23Liked by Vince Mancini

The best explanation as to how/why ChatGPT does things like this that I’ve found is in a paper called “ChatGPT is Bullshit;” it helps that it is written by philosophers and not tech autists:

https://link.springer.com/article/10.1007/s10676-024-09775-5

The basic gist is that the entire point of ChatGPT isn’t to do tasks or provide information, but rather to fake being human.

Expand full comment

Then it's doing it pretty well by making things up to seem smarter and better.

Expand full comment

Let me second this article. It's great.

Expand full comment

AI in the entertainment industry is so noxious that we let a whole piece about finding contrarian critics go and not one mention of Armond White? The future is dumb

Expand full comment

I was fully expecting Armond's name to appear at some later point in this sentence:

"It’s not hard to find real critics who hated on great masterpieces..."

Expand full comment

There's a broad misunderstanding about what the point of this type of generative AI is. They're Large LANGUAGE Models. They are built to sound convincingly human, not to provide information. Obviously, one of the use cases moving forward will be to marry naturalistic language with information retrieval, but that's not where we are nor is that the point of this iteration.

Expand full comment
Aug 22Liked by Vince Mancini

I think this is largely understood by developers and people paying attention. Of the people who don't seem to be internalizing this, the two groups seem to be your regular everyday person, and corporate decision makers who are using it to reduce labor costs. Only one of those really matters, and I'm not even sure they don't understand it. I can't be sure they wouldn't be acting the same way if they fully understood it.

Expand full comment
author

True, very much just an handy excuse to do the kind of cost cutting/shareholder value maxxing they always do anyway in any given situation.

Expand full comment
founding
Aug 22·edited Aug 22

Three things stick out to me:

- It would have taken MAYBE ten extra minutes to pull actual quotes from actual reviews. Almost every review Roger Ebert ever wrote is on his website. Many of the links for old reviews on Rotten Tomatoes aren't dead.

- There are currently 68 reviews for Megalopolis on RT (it's sitting at 53 percent fresh). They could have pulled positive and negative quotes and done a "decide for yourself" campaign. It's been a few years since we had one of those. Remember The Hunt?

- I think they might have gotten away with this if the trailer hadn't opened with The Godfather, which won Best Picture and was the highest grossing movie of 1972.

IGN still has the trailer up, if anyone missed it.

https://www.youtube.com/watch?v=NA5ydinNXiE

Expand full comment

"I'm getting Ayn Rand vibes from this film" - Internet Forum Commentator

Expand full comment

I was just looking for royalty free stock images on a few sites I've used successfully many times, but this time the results are just kind of nonsense, not related to the search term. Most of them were by a Google image generator.

I'm guessing if you use a search term that isn't tagged, it generates nonsense based on the AI's best guess. Useless shit, on demand! What a time to be alive.

Expand full comment

I don't mean to be a luddite but the amount of times I've heard AI* accused of "dishonoring the dead" or "corpse desecrating" seems like more than we as a society should be comfortable with.

*I'm thinking of this and Ian Holm being in the new Alien, which is more CGI, but it's related I think to how AI is being used in deepfakes?

Expand full comment
author

They used it to have Anthony Bourdain read stuff he'd written in that documentary a few years back. (Which doesn't bother me that much if you just cite it)

Expand full comment