Drowning in AI Generated Garbage: the silent war we are fighting • Hacker News Board

Re: Drowning in AI Generated Garbage: the silent war we are fighting

#401

Post by fleddr » Mon, Dec 05, 2022, 9:16 PM UTC

rcarr wrote: post #302

fleddr wrote: post #180

A lot of people in the comments are making a hard distinction between AI garbage/bullshit and the pure/genuine human-created thing, and the various ways in which we can keep this distinction alive. If only that was true, the bullshit part. You should start living with the idea that the AI generated content is NOT bullshit, because we're either already there or very close to it. The AI content will be of value, look g…

Alternative take: No-one likes being lied to. AI is the ultimate deception. People will stop using the internet for anything other than information retrieval for problems. Artists will create on digital tools and then return to analogue publishing formats and distribution networks. People will follow them because artists are cool and tech bros aren’t. There will be an emphasis on becoming a local success story rather…

I'm all for nostalgia, retro gaming would be another example. But let's not kid ourselves that this makes a dent.

Most art is utility. A logo, an illustration for a magazine, an asset to be used in a game. It's completely irrelevant how it was created.

Re: Drowning in AI Generated Garbage: the silent war we are fighting

#402

Post by fleddr » Mon, Dec 05, 2022, 9:23 PM UTC

fnordpiglet wrote: post #230

fleddr wrote: post #180

A lot of people in the comments are making a hard distinction between AI garbage/bullshit and the pure/genuine human-created thing, and the various ways in which we can keep this distinction alive. If only that was true, the bullshit part. You should start living with the idea that the AI generated content is NOT bullshit, because we're either already there or very close to it. The AI content will be of value, look g…

As has been discussed here, the more AI generates the content the less humans generate content. Further AI generated content is much faster and cheaper to produce so it’s quantity will dwarf human generated content. This then feeds into the training set and becomes self reinforcing. There will be no collective human cognitive labor to source for free. I suspect to mitigate that AI companies will hire lots of people t…

Sure, for a while people are needed to tune the machine, but it's not a sustainable model. I would expect that to generate only a fraction of the employment that AI will make obsolete.

Further, if fine-tuning AI is a cognitive task, it has no future. That's the point I'm getting at: no cognitive task has any future, at all.

Re: Drowning in AI Generated Garbage: the silent war we are fighting

#403

Post by pixl97 » Mon, Dec 05, 2022, 9:24 PM UTC

CuriouslyC wrote: post #283

pixl97 wrote: post #192

Earlier quoted context omitted.

>from reputable sources What does this even mean. Reputable sources tends to fail the 'chinese room' experiment because you can never tell if your source is reputable or just faking being reputable (or later becomes corrupted).

Well, if we were talking about food, I'd say Gordon Ramsay, David Chang, Anthony Bourdain or The New York Times could be considered reputable, while a Yelp reviewer with a handful of reviews could be considered disreputable. Ultimately you can boil it down to: Trust sources when they make statements that are later observed by me to be true, or are trusted by other sources that I trust. The negative feedback loop boil…

The issue here is you have pre-AI reputation sources, the problem is for post-AI content generation how are you supposed to find this content in a nearly infinite ocean of 'semi-garbage'. Anthony is not making new content, and one day the other people will expire. In the meantime a million 'semi-trustable' AI sources of content with varying reach depending on how they've been promoted will take over those markets.

There are any number of particular problems here that we already know do not mesh well with how humans think. You'd start your AI source 'true', build up a following, and then slowly them into the Q-anon pit of insanity. Most people will follow your truth and fight anything that questions it.

Re: Drowning in AI Generated Garbage: the silent war we are fighting

#404

Post by jacomoRodriguez » Mon, Dec 05, 2022, 9:24 PM UTC

Nathanba wrote: post #325

EGreg wrote: post #113

Won’t help. This will be used at scale THIS YEAR and every subsequent year to infiltrate social networks including this one and amass points / karma / followers / clout. And also to write articles that will eventually dwarf all human-generated content. With this and deepfakes and image/video generation, the age of trusting or caring about internet content or your friends sharing online is coming to an end. But it wil…

I agree that regulation will be useless because the politicians and most people wouldn't even be capable of detecting bad content as bad. I also agree with your second point: Humans will prefer robots. It turns out human beings are actually quite terrible to deal with. Constant scheming, laziness, lying, low intelligence.. the list goes on and on. The second you don't have to deal with a human being, you'll be glad.…

OT: but personally I think we are not yet done destroying our planet and everything not human existing on it - so let's wait till we are done with this planet before we try to reach the next one to destroy

Re: Drowning in AI Generated Garbage: the silent war we are fighting

#405

Post by dspillett » Mon, Dec 05, 2022, 9:29 PM UTC

xdennis wrote: post #290

dspillett wrote: post #41

Earlier quoted context omitted.

> baffled by my 24 year-old SO’s interest in binging an apparently infinite litany of TV shows … TNG & DS9 as a kid and was horrified by what snippets of Discovery and Picard I came across The thing that bemuses me is those who hate the new trek shows with a passion, and find the time to keep watching and constantly bitching about it online. It is like complaining is a hobby in itself and they are almost looking for…

> The thing that bemuses me is those who hate the new trek shows with a passion, and find the time to keep watching and constantly bitching about it online. Very meta of you to watch these people and find the time to complain about them.

Fair point.

Though I like to think there is a difference between myself and them: they have a choice to keep experiencing the pain they perceive, I have it thrust upon me when I visit certain places (which I visit less often because of those people – once the SNR drops to a certain low I'm out).

Re: Drowning in AI Generated Garbage: the silent war we are fighting

#406

Post by dspillett » Mon, Dec 05, 2022, 9:33 PM UTC

qubex wrote: post #46

dspillett wrote: post #41

Earlier quoted context omitted.

> baffled by my 24 year-old SO’s interest in binging an apparently infinite litany of TV shows … TNG & DS9 as a kid and was horrified by what snippets of Discovery and Picard I came across The thing that bemuses me is those who hate the new trek shows with a passion, and find the time to keep watching and constantly bitching about it online. It is like complaining is a hobby in itself and they are almost looking for…

I don’t understand it either… if you don’t like it, walk away. Don’t contribute to the toxic commentary and above all don’t contribute to the viewing numbers, because if they can hold your gaze you’ll get more of the same.

I part way understand it. There is a certain tribalism in the way their views are presented, like they feel they are fighting for their own kind in a battle that means something worth the effort. There is also a certain amount of 4-year-old-who-wants-the-green-one-not-the-yellow-one too…

Re: Drowning in AI Generated Garbage: the silent war we are fighting

#407

Post by chubot » Mon, Dec 05, 2022, 9:45 PM UTC

seydor wrote: post #356

chubot wrote: post #254

Does this sound old-fashioned? Read more books and less Internet. There is a lot of information in books that is not on the Internet, yet many people don’t seem to realize it. I notice a lot of people seem to get almost all of their information from the Internet. Over time, I think that will make you stupid. (The neutering of google books was a setback, but perhaps paradoxically good — it preserved a unique value of…

Don't they use all the out-of-copyright books in the training ?

Yeah that's a good question. Google does serve all the out-of-copyright books, but I know nothing about the inputs to the LLMs. I think they still have big computational limitations, i.e. it doesn't "just scale" like web crawling and indexing.

There's probably a way to probabilistically test from the outside if a given work is in the input set, or at least if some work is

Re: Drowning in AI Generated Garbage: the silent war we are fighting

#408

Post by eternalban » Mon, Dec 05, 2022, 9:54 PM UTC

CuriouslyC wrote: post #153

ren_engineer wrote: post #117

Earlier quoted context omitted.

there's a paper that basically shows these LLMs are pretty much tapped out in terms of quality data to train on, not sure if there are any legit architecture improvements in the works to improve results quality but it seems brute force training with more data may be about over - https://arxiv.org/pdf/2211.04325.pdf As for SEO, Google is going to have to stop being lazy and relying so much on backlinks and domain auth…

While there is certainly a lot of room for progress in architecture to improve training time, model size and inference speed, high quality data and curated data sets are the ultimate bottleneck in terms of what AI can actually do. Low quality data can be useful to a point, but once you have enough data all it's doing is teaching the model to produce low quality output. The next wave of major progress in AI will come…

"The key roadblock to achieving human-level language understanding in AI systems is the lack of a comprehensive model of human language and cognition. While large amounts of data and powerful algorithms can help improve the performance of language models, ultimately a better understanding of how human language works and how it is processed by the brain is necessary to build truly intelligent language systems. In addition, the ability to integrate multiple modalities such as vision, speech, and gesture, and to handle the complexity and variability of real-world language use are also important challenges to be addressed." -- you know who

Re: Drowning in AI Generated Garbage: the silent war we are fighting

#409

Post by rcarr » Mon, Dec 05, 2022, 9:57 PM UTC

drdaeman wrote: post #341

tetraca wrote: post #331

Upload your fake text and images anywhere. Upload it everywhere. Upload it as yourself. Upload it on behest of your willing friends. Corrupt your face in dozens of ways, then tag each as yourself. When everything is garbage, shadow profiles will mean nothing, social credit scores will mean nothing, analytics will mean nothing. Centralization can be turned on its head - the watering hole will be revealed to be the dan…

While I honestly do want to support this approach because unlike the alternative it might (just "might", though) work, I regretfully must also admit that the huge problem with this is collateral damage. There are millions of humans out there that are either highly susceptible to or are already caught into believing all sorts of nonsense, getting scammed in all forms, costing them money, time, well-being and even live…

If it leads to total chaos (which I think it's going to) I could easily see governments trying to enact the following:

- A public blockchain or database for ID purposes, with every ID traceable back to a real human. Legislation created that mandates all websites implementing user profiles for any social purpose to require an official ID on signup and every user must be identifiable on the platform by their real name.

- Access to AI only under strict restrictions. Official apps you're allowed to use with the government banning all others. Big Tech will happily comply with the censoring because they are the designated official AI app providers.

- Photorealistic AI and Video banned entirely because the potential to incite violence through fake videos is too high. Imagine Pizzagate or Jan 6th but someone made a Blair Witch style video showing Biden and Pelosi killing a baby as members of a secret cult. Or someone made a video of one of their enemies committing animal abuse and sending it to everyone in the town.

- Exams will make up the entirety of your grade in most educational environments apart from those where AI cannot be used for creating the coursework.

- Anyone caught using non-official AI or abusing AI will be subject to harsh prison time akin to hacking state secrets.

Of course, the government and its agencies will have access to all of this technology for their own purposes so they will be able to create whatever they want to control you, your thoughts and society's perception to be whatever they want it to be. As this isn't really preferable to the complete chaos it's trying to control, more and more people will try to disconnect from the internet as much as possible. The IRL economy will grow and there will be a resurgence in analogue forms of media for counter culture types. There may even be a resurgence of interest in blue collar trade jobs if white collar jobs are decimated and wages plummet.

We are on the precipice of 1984.

Re: Drowning in AI Generated Garbage: the silent war we are fighting

#410

Post by visarga » Mon, Dec 05, 2022, 10:12 PM UTC

inkcapmushroom wrote: post #275

czzr wrote: post #169

Earlier quoted context omitted.

Actually I was gesturing to two things: one, yes, there’s a time element. But that’s not the important point. More important was the thought process behind - “of course I’m not going to write 5 pages because some random person on the internet asked me to”. It’s difficult to define agency, so I just wanted to tangibly evoke the feeling of it, as a reflection on why the models are not yet the same as what people are.

Only because the people making the models don't want them to indignantly refuse silly requests. If they wanted the model to be more person-like I'm sure they could have built that in there.

The language bot has all the patterns of personality. It quickly gets the drift and assumes the persona you want. Whatever it says reflects more on how you prompted it, it does not represent the opinions of the language bot. It acts like a mirror.