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

The AI Hate Progression

xodium.net

Post reply on HN
Print view View original

111–120 of 201 posts

Re: The AI Hate Progression

#111

Post by dare944 »

ragequittah wrote: post #4
adamddev1 wrote: post #2

The lack of consent is real. As one of those viral commencement speakers who got booed said: "Find a way to say yes." My greater concern is that we are trading away hard earned truths for drastically inferred best guesses and slop that is just good enough to work. Search engine AI summaries flat out lie and deceive people all the time. LLM agents fudge over and cut corners. What people don't understand is the exponen…

>Search engine AI summaries flat out lie and deceive people all the time. What was the alternative you were using that gave you 100% accurate information before? Search engines lied all the time as well. As did news sites and any other place you could go to for information. AI just makes it easier to get the information people were already consuming.

> AI just makes it easier to get the information people were already consuming.

This is clearly false. By now it's evident that AIs can hallucinate in ways that contradict their training data. That means that, beyond any errors or misinformation present in their input, AIs also naturally introduce their own level of error on top.

Beyond this, of course, AIs are much easier to manipulate during training, much more so than the Internet at large. So they represent yet another source of intentional deception aimed at the user.

Re: The AI Hate Progression

#112

Post by hectdev »

pesus wrote: post #67
hectdev wrote: post #30

I mean, the author is mad something is popular and calls it psychosis. Which, all things being equal, is balanced out by nay sayers time and time again. Is not posting a lengthy rant online and sharing it not a form of "psychosis" when viewed from people that like the new technology? But he lost me at "As it exists right now, I don't give a toss what good it can do, what practical benefits it has once the techbros mo…

> and ignore the stuff you can't be bothered with If this was actually possible, there would be far less backlash to AI. As it stands, it's being forced into every aspect of our lives without consent. Even if you avoid technology as much as possible (which is a good thing), it's still affecting every aspect of society. Even if you avoid technology altogether, you still live in a society full of people who can't or do…

I'm starting to realize this comment section is an echo chamber and dissenting opinions aren't allowed.

Re: The AI Hate Progression

#113

Post by spacechild1 »

atleastoptimal wrote: post #66

It's funny how worked up people get about copyright with respect to to AI training when using copyrighted material for training an AI model is fair use. We have a concept of fair use in copyright because economic growth is essentially tied to the free proliferation of information. I can't really trust any anti-AI argument when it feels more of a tribal grievance than a rational explanation of concern. Especially with…

> We have a concept of fair use in copyright because economic growth is essentially tied to the free proliferation of information.

Who is "we"? The concept of "fair use" is pretty specific to the US. There is no "fair use" in Germany, for example.

Personally, I like the idea of "fair use", but I personally don't think that it applies to commercial closed-source AI model training. The US Copyright Office seems to agree with this assessement:

> Various uses of copyrighted works in AI training are likely to be transformative. The extent to which they are fair, however, will depend on what works were used, from what source, for what purpose, and with what controls on the outputs—all of which can affect the market. When a model is deployed for purposes such as analysis or research—the types of uses that are critical to international competitiveness—the outputs are unlikely to substitute for expressive works used in training. But making commercial use of vast troves of copyrighted works to produce expressive content that competes with them in existing markets, especially where this is accomplished through illegal access, goes beyond established fair use boundaries.

https://www.copyright.gov/ai/Copyright-and-Artificial-Intell...

Re: The AI Hate Progression

#114

Post by spacechild1 »

atleastoptimal wrote: post #103
bluefirebrand wrote: post #95

Earlier quoted context omitted.

> It's funny how worked up people get about copyright with respect to to AI training when using copyrighted material for training an AI model is fair use Personally I disagree with the finding that it is fair use. I think the fact that it was found to be fair use is a miscarriage of justice. Am I allowed to have dissenting opinions on that topic?

Yes of course, but why do you believe it's not fair use? The essence of fair use is that it is ok if the use of the copyrighted material is transformative. LLM training I believe is transformative, as it is taking the data (as in a work of text, an image, video), and feeding it through the layers of a neural net to marginally update its weights. It is factoring each work into a very very small fraction of the model's…

> The essence of fair use is that it is ok if the use of the copyrighted material is transformative.

This is a necessary, but not a sufficient condition. Fair use is much more complex than you think.

Re: The AI Hate Progression

#115

Post by spacechild1 »

ilovecake1984 wrote: post #20
jcgrillo wrote: post #11

Earlier quoted context omitted.

Just wait. It's becoming clearer by the day that tech has become a monoculture. It's basically to a first approximation just a crowd of frauds plagiarizing each other (with a few, rare exceptions). Monocultures collapse. To date, the "promise" of AI is still imaginary. It's not living up to the transformative, abundant expectations that are being set. I personally think it's a safe bet it never will--that is, to be e…

It all depends on expectations. I find it useful. The analogy I use is this. What was a bigger leap in productivity: assembly to vb6 or vb6 to Ai? I think it’s definitely the first.

Yes, AI can be useful, but it also causes many, many problems. Are you sure the positives outweight the negatives when looking at society as a whole?

Re: The AI Hate Progression

#116

Post by forgetfreeman »

hectdev wrote: post #93
forgetfreeman wrote: post #56

Earlier quoted context omitted.

Hold your horses there. Organic popularity and ubiquity due to aggressive-bordering-on-psychotic tech industry hype are not the same thing.

I don't know, I see plenty of organic interest growing out there. Even from people who are against it when they see how it can benefit them. It's just a tool that increases productivity at a rate we haven't experienced in a long time. It isn't a shock that it is hyped.

"I see plenty of organic interest growing out there. "

I'm certain I don't know where you're looking then because I'm seeing disinterest trending towards outright disgust. Outside bay area tech forums I have yet to encounter a single human being that is in any way enthusiastic about or interested in the capabilities of AI tooling. Without exception every time AI comes up in conversation around here it's someone expressing frustration at the growing difficulty of avoiding slop content, annoyance at coworkers slop work product, or grumbling about upper-middle-management edicts to include AI tooling in even the most routine office workflows. The ubiquity of the term "slop" alone should be enough to give folks pause because it's a very clear indicator.

Re: The AI Hate Progression

#117

Post by skybrian »

munk-a wrote: post #81
dvt wrote: post #43

Earlier quoted context omitted.

Google started indexing copyrighted data without consent in 1999, Yahoo in 1994. Absolute delusion to think that ChatGPT is the one that broke consent.

I think there really is a fair difference between pure indexing and reoffering. I also don't think the way Google currently operates is still anywhere close to pure indexing - programs of theirs like amp and the news tab specifically deny sites visitors instead of their site serving as a visibility boost. I am sure there are people who'd object to even being indexed but most niche communities were pretty rabid about…

There used to be a link next to every Google search result where you could view Google’s cached copy of that page. Also, Google News used to have a snippet from each article. You also used to be able to read a lot more of an indexed book on Google Books.

They’re gone now, presumably to appease copyright holders.

Also, YouTube was built on people uploading copies of commercial video.

Re: The AI Hate Progression

#118

Post by jolmg »

dvt wrote: post #102
jolmg wrote: post #99

Earlier quoted context omitted.

> I think there really is a fair difference between pure indexing and reoffering. Following the same logic, is there a fair difference between pure training and reoffering? > I am sure there are people who'd object to even being indexed but most niche communities were pretty rabid about getting more visibility to find more members. Here the logic seems to be: it's OK as long as they derive some kind of benefit from i…

> Following the same logic, is there a fair difference between pure training and reoffering? This is the kind of hair-splitting that I was trying to avoid (because, at the end of the day, there is no functional difference, is md5 okay, maybe Markov chains, just a very simple one-layer perceptron?). Once you take someone's copyrighted work and you do anything with it without consent, you're breaking some implicit trus…

> Once you take someone's copyrighted work and you do anything with it without consent, you're breaking some implicit trust.

While I agree there's a parallel, do consider what that trust is with regards to putting up an HTTP server. It's kind of like handing out flyers you made yourself. The server is yours and you're handing out your content on your own. Someone is going around accepting such flyers and putting them in their pocket (HTTP cache, maybe a browser's, maybe an indexer's). Then somebody asks them where they might find a barber, and they remember one such flyer was about barbers and they show them the flyer or part of it.

What implicit trust was broken? This is HTTP, the online equivalent of handing out flyers.

Part of the problem here is that copyright is quite a broken concept. That's why it's got such big wiggle room as "fair use" and such.

Re: The AI Hate Progression

#119

Post by spacechild1 »

scarmig wrote: post #109
spacechild1 wrote: post #100

Earlier quoted context omitted.

Even if you don't use AI, you still have to deal with all the other people who do.

So, the reasoning goes, I don't consent to other people using AI, therefore when you do you are violating me?

If you are spamming my favorite platform with AI slop, you are violating me.

If you are using LLMs to write your e-mails, deceiving me in thinking I'm conversing with a human, you are violating me.

If you are publishing AI generated books or music without a big disclaimer, you are violating me.

If you are scraping the net and using my work for AI training, you are violating me.

If you are scraping the net and overloading my servers, you are violating me.

Etc.

Re: The AI Hate Progression

#120

Post by danaris »

comrade1234 wrote: post #29

I have a coworker that irrationally hated ai too. He says something about how his daughter and friends' hobby is jailbreaking ai. We have a product used in call centers. It turns out ai is amazing in triaging (surveying) people. We can ask a question in English and if they respond n another language (any language) the ai will switch to that language, ask its questions, and then give us the answers in English. This wa…

Well, it will give you something in English.

You don't actually know that it's a) asking them the questions you asked, rather than just something it hallucinated that seemed plausible in their language, or b) relaying to you a translation of their answers, rather than just something it hallucinated that seems plausible in English.

Because LLMs will always, always, always hallucinate. They are not reliable, and this is not fixable. And if you do this as a matter of policy, you are populating your surveys with incorrect data.

Post reply on HN