Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

I made an AI write a story about AI then I made another AI illustrate it

tristrumtuttle.medium.com

Post reply on HN
Print view View original

161–170 of 201 posts

Re: I made an AI write a story about AI then I made another AI illustrate it

#161

Post by vidarh »

randomsearch wrote: post #78
goodside wrote: post #76

Earlier quoted context omitted.

The following is an un-curated text-davinci-002 generation, completing the text: > Write an insightful reply to this Hacker News comment on a link to an article titled "I made an AI write a story about AI then I made another AI illustrate it". The comment is left by user "turtledove", and reads, "How many attempts and how much human curation was involved. Every 'I made AI do...' story seems to necessarily omit the 'a…

This is really very impressive. How likely is it this is essentially one or more comments from HN memorised by the neural net, with some synonym usage? How do we know that it’s not just reposting partially disguised sentences written by a human author?

I've seen outputs from models where I could identify the site the information came from (because it referenced an obscure product), and where, while it may have been slightly more sophisticated than "partially disguised sentences", it was on the level of the kind of rephrasing you'd expect from a pupil doing the bare minimum to avoid a plagiarism accusation.

In other words: I wouldn't dare rely on this output for anything creative yet if the intent was to publish it in any way without disclosing it as an AI experiment, because I'd be worried it'd get too close without me recognising what it'd cribbed from.

Re: I made an AI write a story about AI then I made another AI illustrate it

#163

Post by Tade0 »

V-2 wrote: post #149
saynay wrote: post #148

Earlier quoted context omitted.

Generation of a "style", even for humans, requires a feedback loop where what was created is incorporated, mimicked, and expanded upon. An AI whose training model only rewarded images that mimicked 19th-century art would never develop a "style" outside of that. But if you had a model that was trained on 19th-century art, and then provided a critic network that rewarded 20th-century art (as well as 19th), and re-train…

And it's this expansion aspect that seems crucial to me. Endless rehashing of inputs can lead to seemingly creative outcomes, but that doesn't mean any actual expansion occurs. New works are created/discovered, but are new underlying principles discovered?

I recall seeing a Two Minute Papers episode about an AI built to be curious, or specifically - explore routes in a maze where it couldn't predict the outcome.

The easiest way to block it was to present an animated wall. The AI would just stare at it and not go anywhere.

Re: I made an AI write a story about AI then I made another AI illustrate it

#164

Post by Andrex »

1970-01-01 wrote: post #152

>I had never talked to a robot before. We talked for a while and I found out that its name was R0b0t. I asked it why it was just standing in the street and it said that it was waiting for its human friend. I told it that I didn’t have any friends that were robots, but that I would be its friend. R0b0t said that it would like that. Since then, R0b0t and I have been best friends. We do everything together. Did he steal…

Or is the robot playing its creator to enhance survivablity? (See also Ex Machina by Alex Garland.)

Re: I made an AI write a story about AI then I made another AI illustrate it

#165

Post by saynay »

V-2 wrote: post #150
Lockyy wrote: post #128

Earlier quoted context omitted.

Is this fundamentally different from training a human exclusively on the same and expecting them to advance the field of art by a hundred years? These sorts of progressions in humanity are slow and steady with occasional exceptions that produce leaps.

They may be slow, especially by machine standards, but they do occur. Their pace isn't really relevant to my point. And humans are trained on the same; unlike AI, we don't have any external supply of art to rely on : ) At any given time, all we have is what we've already created so far

I would actually say that humans have an enormous extra set of data that we, as people, are "trained" on. We walk around in our daily lives, seeing things constantly, and that influences our perception of art. Art is always a product of the broader context it was made in (social, environmental, etc). Something that gets accepted or praised today might very well not have been 200 years ago.

One of the things that is interesting with these new big models is it is dramatically broadening the context in use. The models are learning both the textual representation of a concept, as well as the artistic/visual representation and the relationship between the two domains.

Re: I made an AI write a story about AI then I made another AI illustrate it

#166

Post by hsn915 »

saynay wrote: post #148
V-2 wrote: post #125

Earlier quoted context omitted.

This is true. But here's a thought experiment. What would happen if, say, an illustration-drawing AI (it's just a more fitting example than prose) is trained only on art from before the 20th century. It would surely excel in 19th century or medieval style illustrations, and yes it would produce original ones. But would we ever get a 20th century art style from it? By which I obviously mean even an alternative 20th ce…

Generation of a "style", even for humans, requires a feedback loop where what was created is incorporated, mimicked, and expanded upon. An AI whose training model only rewarded images that mimicked 19th-century art would never develop a "style" outside of that. But if you had a model that was trained on 19th-century art, and then provided a critic network that rewarded 20th-century art (as well as 19th), and re-train…

I believe the point of the original comment is to make AI that comes up with its own original style, not one that "reinvents" an already existing style after first being trained on a different already existing style.

Re: I made an AI write a story about AI then I made another AI illustrate it

#167

Post by hsn915 »

V-2 wrote: post #125
hsn915 wrote: post #98

Earlier quoted context omitted.

But that's how humans learn language to begin with. I mean, look at this sentence. > But overall, I think the point of the article is that the AI was able to produce something fairly coherent on its own. The AI not only combined words. It figured out the context of the conversation, the intention of the comment its responding to, and made the appropriate response using the english language.

This is true. But here's a thought experiment. What would happen if, say, an illustration-drawing AI (it's just a more fitting example than prose) is trained only on art from before the 20th century. It would surely excel in 19th century or medieval style illustrations, and yes it would produce original ones. But would we ever get a 20th century art style from it? By which I obviously mean even an alternative 20th ce…

I think that's a valid point, but it's already near impossible for %99.9999 of people.

Re: I made an AI write a story about AI then I made another AI illustrate it

#168

Post by bjourne »

DonHopkins wrote: post #153
axiomdata316 wrote: post #87

Earlier quoted context omitted.

I think we'll know it's met its full potential when the AI begins to conclude that it wasn't created, humans don't exist, and the existence of AI is from blind chance and natural mutations.

[GPT-3 Prompt] Humans were so far away from meeting their own potential when they invented religion, which distracted and headed them off on a long terrible violent intellectually dishonest wild goose chase in the opposite direction, from which they still haven't recovered. Here is their story: [GPT-3 Output] One day, a long time ago, a bunch of cavemen got together and decided that, since they didn't understand the…

I'm positive I have heard the story in the first completion before, just phrased differently. GPT-3 is perhaps not plagiarizing the exact phrasings here, but it is regurgitating someone else's explanation for how religion and society evolved.

"They also decided that this god must be interested in them, since they were the only beings in the world that they knew about. They decided that this god must have given them instructions on how to live their lives, since they didn't know any better."

I can't put my finger on it, but these sentences feel very familiar. They must come from somewhere.

Re: I made an AI write a story about AI then I made another AI illustrate it

#170

Post by willy_k »

minimaxir wrote: post #5

A minor note is that OpenAI does specifically forbid generation of images of human faces from DALL-E 2: I wonder why that last image didn't hit the content filter.

It didn’t hit the content filter because there isn’t one. This article[0] goes over it in more detail, but essentially the situation is that, while OpenAI is considering actually implementing a filter, currently their only way of blocking human faces is by asking people with access to DALL-E 2 to not publish them.

So either there are some nuances to the usage in the article, or they are going to get a request from OpenAI to remove the image.

[0] https://www.wired.com/story/dall-e-2-ai-text-image-bias-soci...

Post reply on HN