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

The 100k whys of AI

lcamtuf.substack.com

Post reply on HN
Print view View original

61–70 of 111 posts

Re: The 100k whys of AI

#61

Post by tasuki »

thinkingemote wrote: post #19

On HN many comments under many threads are about whether the submission was written by AI. You could say I have noticed a pattern in Hacker News comments! In these comments there's a common pattern where some users argue that they do not agree that the submission was LLM written and they often focus on specific details to refute it (e.g em-dashes) and some users see the overall pattern clearly that it's totally obvio…

> The take away I get is that it's okay to notice patterns and it's okay to not notice patterns. Remember that other people may be noticing patterns and associations in things that you might miss. Be charitable.

I wish the people (often wrongly) accusing others of using a LLM to generate whatever were more charitable. Yes we notice patterns. But then we also notice patterns where there are none.

Re: The 100k whys of AI

#62

Post by Planktonne »

This is exactly why it is perfectly possible to identify AI-generated prose/images; it's not that any one word or sentence is the tell, but that it all sounds/looks the same as the other generated stuff.

At this point, I think the people who struggle with identifying the AI feel are telling you that they don't really engage with media much.

Re: The 100k whys of AI

#63

Post by TrackerFF »

dlenski wrote: post #2

A nice illustration of the homogeneity of LLM responses. Another way to describe this effect would be… If you ask humans to write 1,000 books, you're asking 1,000 different humans with different experiences and different skills and different moods (etc.) to write those books. But if you ask LLMs to write 1,000 books, you're probably only talking to 3 or 5 different models, tops. And they've all trained on the same or…

LLMs are great at producing average.

We see this with their GenAI music equivalents. All the music these GenAI models produce is exceptionally (aggressively, even) average.

It is the most polished average you'll ever find. Never awful (anymore), never fantastic. Just bang in the middle.

Re: The 100k whys of AI

#64

Post by Mikhail_Edoshin »

ekianjo wrote: post #34
dlenski wrote: post #2

A nice illustration of the homogeneity of LLM responses. Another way to describe this effect would be… If you ask humans to write 1,000 books, you're asking 1,000 different humans with different experiences and different skills and different moods (etc.) to write those books. But if you ask LLMs to write 1,000 books, you're probably only talking to 3 or 5 different models, tops. And they've all trained on the same or…

prompts will give very different results. this is where you do the work.

A controller has to be at least as complex as what it is supposed to control.

Re: The 100k whys of AI

#65

Post by NitpickLawyer »

spwa4 wrote: post #57
NitpickLawyer wrote: post #50

Earlier quoted context omitted.

> 1000 variants of existing art. This is very naive. I can almost guarantee that some combinations of 20 * 50 features will hit on something that has never been written before in that specific combination . And if that's still not enough, increase the number of features. Add more randomness, add more steering, add random steering in random chapters, change it up, and so on.

> Add more randomness, add more steering, add random steering in random chapters, change it up, and so on. That doesn't work for AI models. The whole training process depends on the basic principle that if you take the average of 100, in this case book cover designs, that the average is less like randomness than any individual cover you've used to make your average. So the output will, by necessity, be closer to the…

> That doesn't work for AI models.

Of course it does. I know it does because I've been using variations of this workflow since gpt3.0. In fact it's the only way it can work, since by design LLMs work from left to right. You can't expect it to produce original stuff if you don't give it the anchors for what original means. It'd be like going to a new bar every night and asking for a "beer that you haven't had before". There's no information to work on there.

Re: The 100k whys of AI

#66

Post by Lerc »

dlenski wrote: post #2

A nice illustration of the homogeneity of LLM responses. Another way to describe this effect would be… If you ask humans to write 1,000 books, you're asking 1,000 different humans with different experiences and different skills and different moods (etc.) to write those books. But if you ask LLMs to write 1,000 books, you're probably only talking to 3 or 5 different models, tops. And they've all trained on the same or…

I wonder how much variation there would be if you got a single model to produce a couple of gigabytes of tiny children's stories.

Might be an interedting research project.

Re: The 100k whys of AI

#67

Post by xnx »

No one should like slop autogen books, but this is barely more damning than being upset that all the garments have 2 legs when they search for "pants".

Re: The 100k whys of AI

#68

Post by pxagntuvzt »

Lerc wrote: post #66
dlenski wrote: post #2

A nice illustration of the homogeneity of LLM responses. Another way to describe this effect would be… If you ask humans to write 1,000 books, you're asking 1,000 different humans with different experiences and different skills and different moods (etc.) to write those books. But if you ask LLMs to write 1,000 books, you're probably only talking to 3 or 5 different models, tops. And they've all trained on the same or…

I wonder how much variation there would be if you got a single model to produce a couple of gigabytes of tiny children's stories. Might be an interedting research project.

There is one already: https://arxiv.org/abs/2305.07759 https://huggingface.co/datasets/roneneldan/TinyStories

6.5GB of tiny stories, as requested. ;)

Re: The 100k whys of AI

#69

Post by spwa4 »

NitpickLawyer wrote: post #65
spwa4 wrote: post #57

Earlier quoted context omitted.

> Add more randomness, add more steering, add random steering in random chapters, change it up, and so on. That doesn't work for AI models. The whole training process depends on the basic principle that if you take the average of 100, in this case book cover designs, that the average is less like randomness than any individual cover you've used to make your average. So the output will, by necessity, be closer to the…

> That doesn't work for AI models. Of course it does. I know it does because I've been using variations of this workflow since gpt3.0. In fact it's the only way it can work, since by design LLMs work from left to right. You can't expect it to produce original stuff if you don't give it the anchors for what original means. It'd be like going to a new bar every night and asking for a "beer that you haven't had before".…

The point was to take a random combination of story elements. Pick one each {King,dad,CEO} {betrays,kills,loves} {his enemy,the king,a foreign prime minister} and feed to an LLM.

The output will not be an intricate well designed epic storyline, but a cookie-cutter boring snoozefest.

BUT you can give that to a bunch of humans, who "insert their life experience" (ie. parts of their training data, translated to LLM terms) and sometimes out comes Game of Thrones, Star Wars, ...

Re: The 100k whys of AI

#70

Post by hackingonempty »

Even the authors name seems to be generated in many cases. Look at how often "Bright" appears: Andrew W. Bright, Nolan Bright, Bright A. Jeffery, Pamela Bright, Thomas Bright, Daniel Bright, Mayan Bright, Henry Brightwood, Leo Brightham, Milo Brightspark.

There's also Molly Wonder, Elliot Wonder, Professor Pax Wonder, and Theo Wonderquill

Don't forget Lucas Thinkwell!

Post reply on HN