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

An image of an archeologist adventurer who wears a hat and uses a bullwhip

theaiunderwriter.substack.com

Post reply on HN
Print view View original

441–450 of 927 posts

Re: An image of an archeologist adventurer who wears a hat and uses a bullwhip

#441

Post by MgB2 »

ramraj07 wrote: post #439
MgB2 wrote: post #30

Idk, the models generating what are basically 1:1 copies of the training data from pretty generic descriptions feels like a severe case of overfitting to me. What use is a generational model that just regurgitates the input? I feel like the less advanced generations, maybe even because of their limitations in terms of size, were better at coming up with something that at least feels new. In the end, other than for co…

So I train a model to say y=2, and then I ask the model to guess the value of y and it says 2, and you call that overfitting? Overfitting is if you didn't exactly describe Indiana Jones and then it still gave Indiana Jones.

The prompt didn't exactly describe Indiana Jones though. It left a lot of freedom for the model to make the "archeologist" e.g. female, Asian, put them in a different time period, have them wear a different kind of hat etc.

It didn't though, it just spat out what is basically a 1:1 copy of some Indiana Jones promo shoot. No where did the prompt ask for it to look like Harrison Ford.

Re: An image of an archeologist adventurer who wears a hat and uses a bullwhip

#442

Post by awesome_dude »

dcow wrote: post #436
awesome_dude wrote: post #413

Earlier quoted context omitted.

There are precisely three models for funding Private - this includes funding by selling item(s), licensing work, and private equity State Charity - this includes volunteers, patrons, donations, sponsorships. Charity relies on people willing to donate for the betterment of others. State funding fails because of the political nature of the person holding the purse strings. Licensing, copyright, physical sales are the o…

I am supportive of private and charity funding. I think we can do it without a focus restricting copying. I think this because there is precedent with any industry that relies on trade secrets. Once I can copy a Coke with a food printer we'll be having some really weird internal consistency issues with copyright.

I'm not seeing a convincing argument from you other than "Once I bought a book that was public domain"

Re: An image of an archeologist adventurer who wears a hat and uses a bullwhip

#443

Post by snowwrestler »

dcow wrote: post #423

Either, (1) LLMs are just super lossy compress/decompress machines and we humans find fascination in the loss that happens at decompression time, at times ascribing creativity and agency to it. Status quo copyright is a concern as we reduce the amount of lossiness, because at some point someone can claim that an output is close enough to the original to constitute infringement. AI companies should probably license al…

This is the opposite of how people have thought about creativity for centuries, though.

The most creative person is someone who generates original, compelling work with no prompting at all. A very creative person will give you something amazing and compelling from a very small prompt. A so-so creative person will require more specific direction to produce something good. All the way down to the new intern who need paragraphs of specs and multiple rounds of revision to produce something usable. Which is about where the multi-billion-dollar AI seems to be?

Re: An image of an archeologist adventurer who wears a hat and uses a bullwhip

#444

Post by otabdeveloper4 »

o11c wrote: post #47
MgB2 wrote: post #30

Idk, the models generating what are basically 1:1 copies of the training data from pretty generic descriptions feels like a severe case of overfitting to me. What use is a generational model that just regurgitates the input? I feel like the less advanced generations, maybe even because of their limitations in terms of size, were better at coming up with something that at least feels new. In the end, other than for co…

It's not a single image though. Stitching 3 or so input images together clearly makes copyright laundering legal.

No it doesn't. Commercial intent (usually) makes it illegal, not the fact of copying.

So the criminal party here would be OpenAI, since they are selling access to a service that generates copyright-infringing images.

Re: An image of an archeologist adventurer who wears a hat and uses a bullwhip

#445

Post by vjerancrnjak »

MgB2 wrote: post #30

Idk, the models generating what are basically 1:1 copies of the training data from pretty generic descriptions feels like a severe case of overfitting to me. What use is a generational model that just regurgitates the input? I feel like the less advanced generations, maybe even because of their limitations in terms of size, were better at coming up with something that at least feels new. In the end, other than for co…

If it overfits on the whole internet then it’s like a search engine that returns really relevant results with some lossy side effect.

Recent benchmark on unseen 2025 Math Olympiad shows none of the models can problem solve . They all accidentally or on purpose had prior solutions in the training set.

Re: An image of an archeologist adventurer who wears a hat and uses a bullwhip

#446

Post by loki-ai »

jim-jim-jim wrote: post #351
mvdtnz wrote: post #221

Earlier quoted context omitted.

I'm guessing you've never created something of value before. People are entitled to the fruits of their labour and control of their intellectual property.

If I paint a picture on a physical canvas, I can charge people to come into my house and take a look. If I bring the canvas to a park, I'm not entitled to say "s-stop looking at my painting guys!" If you're worried about your work being infinitely reproduced, you probably shouldn't work in an infinitely-reproducible medium. Digitized content is inherently worthless, and I mean that in a non-derisive way. The sooner w…

and how do you reconcile any work in software development? If someone isn’t willing to work for free, should they just not work in the field at all? Do you think software culture would really be richer?

Re: An image of an archeologist adventurer who wears a hat and uses a bullwhip

#447

Post by numlock86 »

> close-up image of a cat's face staring down at the viewer

> describe indiana jones

> looks inside

> gets indiana jones

Okay, so the network does exactly what I would expect? If anything you could argue the network is bad because it doesn't recognize your prompt and gives you something else (original? whatever that would mean) instead. But maybe that's just me.

Re: An image of an archeologist adventurer who wears a hat and uses a bullwhip

#448

Post by jks »

vjerancrnjak wrote: post #445
MgB2 wrote: post #30

Idk, the models generating what are basically 1:1 copies of the training data from pretty generic descriptions feels like a severe case of overfitting to me. What use is a generational model that just regurgitates the input? I feel like the less advanced generations, maybe even because of their limitations in terms of size, were better at coming up with something that at least feels new. In the end, other than for co…

If it overfits on the whole internet then it’s like a search engine that returns really relevant results with some lossy side effect. Recent benchmark on unseen 2025 Math Olympiad shows none of the models can problem solve . They all accidentally or on purpose had prior solutions in the training set.

You probably mean the USAMO 2025 paper. They updated their comparison with Gemini 2.5 Pro, which did get a nontrivial score. That Gemini version was released five days after USAMO, so while it's not entirely impossible for the data to be in its training set, it would seem kind of unlikely.

https://x.com/mbalunovic/status/1907436704790651166

Re: An image of an archeologist adventurer who wears a hat and uses a bullwhip

#449

Post by sothatsit »

GrantMoyer wrote: post #435
sothatsit wrote: post #426

Earlier quoted context omitted.

You don't even need to add much more to the prompts. Just a few words, and it changes the characters you get. It won't always produce something good, but at least we have a lot of control over what it produces. Examples: "An image of an Indian female archeologist adventurer who wears a hat and uses a bullwhip" ( https://sora.com/g/gen_01jqzet1p8fjaa808bmqnvf7rk ) "An image of a fat Russian archeologist adventurer who…

All four of those are dressed like Indiana Jones. They look like different versions of Indiana Jones you'd see in super hero multi-verse story.

So... ask it to dress them differently. You can just ask it to make whatever changes you want.

"An image of an archeologist adventurer who wears a hat and uses a bullwhip. He is wearing a top hat, a scarf, a knit jumper, and pink khaki pants. He is not wearing a bag" (https://sora.com/g/gen_01jqzkh4z2fqctzr9k1jsfnrhy)

Want to get rid of the pose? Add that the archeologist is "fun and joyous" to the prompt. (https://sora.com/g/gen_01jqzksmjgfppbv5p51hw0xrzn)

You have so much control, it is up to you to ask for something that is not a trope.

Post reply on HN