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

No elephants: Breakthroughs in image generation

oneusefulthing.org

Post reply on HN
Print view View original

201–210 of 373 posts

Re: No elephants: Breakthroughs in image generation

#201

Post by JFingleton »

genewitch wrote: post #177
x187463 wrote: post #3

This is a before/after moment for image generation. A simple example is the background images on a ton of (mediocre) music youtube channels. They almost all use AI generated images that are full of nonsense the closer you look. Jazz channels will feature coffee shops with garbled text on the menu and furniture blending together. I bet all of that disappears over the next few months. On another note, and perhaps other…

I have the Gemini app on my phone and you can interact with it with voice only and I was like oh this is really cool I can use it while I'm driving instead of listening to music. I can never think of anything to talk to an AI about. I run LM local, as well

Have it interview you (as-in a job interview) on your specialisation. Works your interviewer skills.

Ask it to teach you a language.

DnD works really well (the LLM being the game-master).

Re: No elephants: Breakthroughs in image generation

#202

Post by Animats »

__loam wrote: post #170
Gasp0de wrote: post #121

Earlier quoted context omitted.

I love using LLMs to generate pictures. I'd call myself rather creative, but absolutely useless in any artistic craft. Now, I can just describe any image I can imagine and get 90% accurate results, which is good enough for the presentations I hold, online pet projects (created a squirrel-themed online math-learning game for which I previously would have needed a designer to create squirrel highschool themed imagery)…

If you use this technology, you're actively harming creative labor.

Creative labor is going the way of manual labor.

Re: No elephants: Breakthroughs in image generation

#203

Post by immibis »

AnthonyMouse wrote: post #197
hnbad wrote: post #39

Earlier quoted context omitted.

Copyright is a logical consequence of property rights. I'd agree that property rights hold back industry and trade but if you want to abolish property rights, you first have to decommodify the essentials like food, housing, public infrastructure and healthcare, because unleashing the market when it has control over all of these is going to have some very undesirable consequences.

Copyright isn't a property right. Property rights are rivalrous. If you own a sandwich and a thousand other people want to eat your sandwich, only one person can, so property rights exist to define who gets to choose who gets to eat the sandwich. Writings and discoveries are non-rivalrous. To quote the first head of the US Patent Office: > He who receives an idea from me, receives instruction himself without lessenin…

The definition of capitalism is a system in which all sorts of things that are not property are artificially made into property and given artificial property rights which can be traded.

Re: No elephants: Breakthroughs in image generation

#205

Post by nialv7 »

nowittyusername wrote: post #141

There is circumstantial evidence out there that 4o image manipulation isn't done within the 4o image generator in one shot but is a workflow done by an agentic system. Meaning this, user inputs prompt "create an image with no elephants in the room" > prompt goes to an llm which preprocesses the human prompt > outputs a a prompt that it knows works withing this image generator well > create an image of a room > and th…

None of your observations say anything about how these images are generated one way or another.

The only thing we currently have to go off of is OpenAI's own words, which claims the images are generated by a single multimodal model autoregressively, and I don't think they are lying.

Re: No elephants: Breakthroughs in image generation

#206

Post by vunderba »

nowittyusername wrote: post #141

There is circumstantial evidence out there that 4o image manipulation isn't done within the 4o image generator in one shot but is a workflow done by an agentic system. Meaning this, user inputs prompt "create an image with no elephants in the room" > prompt goes to an llm which preprocesses the human prompt > outputs a a prompt that it knows works withing this image generator well > create an image of a room > and th…

As somebody who actually tried to build a multimodal stable diffusion chat agent about a year back using YOLO to build partial masks for adjustments via inpainting, dynamic controlnets, and a whole host of other things, I highly doubt that it's as simple as an agentic process.

Using the prompt to detect and choose the most appropriate model checkpoint and LoRa(s) along with rewriting a prompt to most appropriately suit the chosen model has been pretty bog standard for a long time now.

Re: No elephants: Breakthroughs in image generation

#207

Post by smusamashah »

ARandumGuy wrote: post #172
smusamashah wrote: post #135

I am waiting for when I could provide these a scene snippet from "Hitchhiker's Guide To Galaxy" (or any book) and it could draw that for me. The gold planets, the waking up on the beach, total perspective vortex etc. I like the book, but there are quite a few scenes which are quite hard to visualize and make sense. An image generator that can follow that language and detail will be amazing. Even more awesome will be…

I've seen stuff that echoes this sentiment before, and I have to say I don't understand this desire at all. Why would I need a computer to show me what something in a book looks like? I already have an imagination for that! Books are fundamentally a collaborative artform between the author and the reader. The author provides the blueprint, but it's up to the reader to construct the scene in their own head. And every…

You make it sound like an unfair wish, which I would say is unfair itself. I like the book, I like visualising it in my head, and I fantasise what a scene would look like. AI won't generate a true visual. It's all fantasy anyway and AI can actually do it the way I have it in my head. It will solidify that thought.

For this book in particular, I read the comic version and I didn't like the visuals very much. I have a different idea of babel fish. Vogons look different. I would love to see the visual that's in my head on paper.

Re: No elephants: Breakthroughs in image generation

#208

Post by smusamashah »

thwarted wrote: post #186
smusamashah wrote: post #138

Earlier quoted context omitted.

Vortex may be not so much but there are other hard to visualize things. I am on the third book, and I have no idea what Beeblebrox's two heads look like. Second head is often mentioned in passing. Sometimes its mentioned as its always there, other times it feels like its just pops out of somewhere, otherwise, it's like it doesn't exist. There is the scene when they see themselves on the beach on first rescue by the s…

We have alt text for images, you want alt images for text. You can see other people's interpretation of Zaphod's two heads by watching the BBC HHGTTG show (Mark Wing-Davey) or the movie (Sam Rockwell), among other renditions, which offer completely different interpretations, none of them canonical (not the least of which is because there was no canonical version of HHGTTG according to DA). I'm sure there are multitud…

I agree that people interpret things differently and visualise differently and that is my point. I want to see the concept from my head in a solid visual form. Some concepts that are not clear, like those that I mentioned having trouble with, I want to see any kind of visual representation at all. I will probably not like some of those, but there these tools can help generate a bunch of variations tailored specifically for me. I can choose one and carry on with that. I can come back and more details to that as I read further.

If someone can show me exactly what I am thinking of, won't that be amazing.

Re: No elephants: Breakthroughs in image generation

#209

Post by AnthonyMouse »

immibis wrote: post #203
AnthonyMouse wrote: post #197

Earlier quoted context omitted.

Copyright isn't a property right. Property rights are rivalrous. If you own a sandwich and a thousand other people want to eat your sandwich, only one person can, so property rights exist to define who gets to choose who gets to eat the sandwich. Writings and discoveries are non-rivalrous. To quote the first head of the US Patent Office: > He who receives an idea from me, receives instruction himself without lessenin…

The definition of capitalism is a system in which all sorts of things that are not property are artificially made into property and given artificial property rights which can be traded.

That is the definition of capitalism people use when they want to apply the term capitalism to something that sucks.

Allowing people to own physical items as business inventory or production equipment and compete with each other for the customer's dollar is entirely possible without the existence of copyright or patents. You would then be relying on some combination of open source, charitable contributions and patronage, industry joint ventures, personal itch scratching, etc. to create writings and inventions, but books and the wheel were created before patents and copyrights were.

Re: No elephants: Breakthroughs in image generation

#210

Post by lawlessone »

nowittyusername wrote: post #141

There is circumstantial evidence out there that 4o image manipulation isn't done within the 4o image generator in one shot but is a workflow done by an agentic system. Meaning this, user inputs prompt "create an image with no elephants in the room" > prompt goes to an llm which preprocesses the human prompt > outputs a a prompt that it knows works withing this image generator well > create an image of a room > and th…

huh, i wa thinking myself based on how it looked that it was doing layers too. The blurred backgrounds with sharp cartoon characters in front are what made me think this is how they do it.
Post reply on HN