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

Are large language models a threat to digital public goods?

arxiv.org

Post reply on HN
Print view View original

91–100 of 163 posts

Re: Are large language models a threat to digital public goods?

#91

Post by zirgs »

nologic01 wrote: post #77
version_five wrote: post #12

I see this as a rough parallel to "is the printing press a threat to illuminated manuscripts". Maybe it is under a very narrow view, but overall, improved propagation of ideas has led to improved dissemination of ideas, and it will this time too. People who's narrow world has been disrupted will perform all sorts of mental gymnastics to tell us how we're going to be worse off for it, but we won't. Ironically, interne…

True democratization of knowledge would mean LLM's are open source and usable by everyone (the way of Wikipedia). Even then you'd want attribution to survive. Eliminating or changing the incentives people have for doing things is the absolutely surest way to change behavior. Postulating that all published work can be appropriated at will by certain private for-profit entities is the death knell of the knowledge econo…

Art generators are open source and right now there are multiple ongoing court cases attempting to kill them.

Re: Are large language models a threat to digital public goods?

#92

Post by TeMPOraL »

vouaobrasil wrote: post #65
TeMPOraL wrote: post #11

Earlier quoted context omitted.

It's tricky, because in many ways, this is achieving exactly what I, as user, want computers to do for me: give me information I requested, and only that. I very much do not care about who discovered/created/published it, except only if it helps me quickly ascertain the trustworthiness of said information. I do not want to be forced or prodded to establish relationships with creators or communities. I do not want the…

> I very much do not care about who discovered/created/published it, except only if it helps me quickly ascertain the trustworthiness of said information. I do not want to be forced or prodded to establish relationships with creators or communities. I do not want their ads and upsells. And this is why humanity is going down the tubes...because you want something, you derive value from what you want, and yet you do no…

> And this is why humanity is going down the tubes...

On the contrary - this is exactly how and why humanity built a technological civilization in the first place. Note that I didn't say

> because you want something, you derive value from what you want, and yet you do not care about giving something back to who makes it.

Yes, because it would be backward and limiting to do that. Note: I never said I don't want to give anything back - I said I don't personally care specifically about the author/publisher. I don't want to establish any relationship with them. If I'm paying them directly for something, I'm paying them for that thing - not also for relationship (which really is a sales channel), not also for being advertised to. If I'm paying some intermediary, then rewarding the maker is the intermediary's problem, not mine.

Consider: do you compensate directly, and have an active relationship with, the person who bakes your bread (hint: people selling bread in bakeries are not actual bakers)? The company who supplied them with flour? The farmers who supplied the flour-makers with grain? Do you pay delivery drivers directly? After all, you're deriving value from their labor directly. Etc. Then there's an entire army of people whose work benefits you directly, and whom you don't even think much about, and rely on being compensated from the common pool (e.g. taxes) or stochastically.

The whole point of money is to allow exchanging value without forcing parties into maintaining an ongoing relationship. That's a feature, not a bug. And if anything is driving humanity down the drain, it's the idea that you should, need, or are even entitled to capture all the value you produce.

Re: Are large language models a threat to digital public goods?

#93

Post by _Algernon_ »

xigoi wrote: post #76
version_five wrote: post #12

I see this as a rough parallel to "is the printing press a threat to illuminated manuscripts". Maybe it is under a very narrow view, but overall, improved propagation of ideas has led to improved dissemination of ideas, and it will this time too. People who's narrow world has been disrupted will perform all sorts of mental gymnastics to tell us how we're going to be worse off for it, but we won't. Ironically, interne…

Putting all knowledge in the hands of a for-profit company is the opposite of “democratization of knowledge”.

It's not as if the rest of the internet is deleted just because an AI model has been trained on it.

Re: Are large language models a threat to digital public goods?

#94

Post by jerjerjer »

abc_lisper wrote: post #16

They used stack overflow to make their case, and report that user engagement has gone down after the release of ChatGPT. Could it not be the case that SO is less adept at finding related/duplicate questions than ChatGPT? Given the later's facility with the language, I would expect it to be. So I look at the paper to see if they accounted for that, and find this. "Second, we investigate whether ChatGPT is simply displ…

Some of my colleagues are using ChatGPT and blindly follow its advice. ChatGPT's answers often look more persuasive than human answers on SO.

Still, I find ChatGPT answers much harder to validate. In programming Q&A, the answer is usually a series of calls. When you try to apply the human solution (assuming it's at least somewhat correct), the error often lies in input data mismatch or some minor changes in requirements when the answer is not perfectly aligned. Rarely the calls themselves are wrong - they might be deprecated if it's an old answer, but generally, they do at least exist. With ChatGPT, you never know which particular part of the answer has been hallucinated.

Also, as SO has more stringing question requirements, you are forced to construct a minimal case to reproduce the error. Often, composing an SO question got me straight to the cause of the error.

Re: Are large language models a threat to digital public goods?

#95

Post by _Algernon_ »

detourdog wrote: post #8
martingalex2 wrote: post #5

This is a very intriguing study as it would appear the fact that LLMs need public data to train on, its incentive on the marketplace of ideas is to reduce the very thing that gives it its power. It contains the seeds of its own destruction, destroying the web and open data ethos, and yet another data point pointing at another AI winter.

For reasons I can't articulate I see LLMs as a vehicle for removing the creators from their ideas. This is very different than search engines. If a search engine generates traffic for documented ideas it creates a community. An LLM based internet seems to remove the creator and shim itself in between for the sake of business.

Google has done this for years, by quoting what it thinks you want to the top of the search results. This is just the next step.

Re: Are large language models a threat to digital public goods?

#96

Post by 2muchcoffeeman »

hdior wrote: post #50
2muchcoffeeman wrote: post #36

Earlier quoted context omitted.

The problem with this sort of comparison is that up until now, new ideas came from humans and technology advances merely helped to spread ideas, or greased the wheels. Now you can generate new ideas. Often without any skill of your own. This is bad at first glance because look what happened when smart phones became popular: no one remembers phone numbers of their family, people can’t remember directions when driving.…

Skill generating ideas is over rated. Implementation is when skill matters. I can imagine post scarcity Star Trek like ideas but not implement them. My skills at CE and SWE allow me to implement reasonable solutions there. Chasing ideas for the sake of chasing ideas is a form of bike shedding and premature optimization.

You can only say this right now where there are an abundance of ideas and less skill to implement them. But even implementation requires creative thinking and the ability to generate ideas. Now we want to take away the need for the first part. This is NOT like going from handwriting to printing or radio to TV.

Something new is going on.

I use ChatGPT for work in a limited way. If you haven’t tried it, I suggest that you do. This potentially paradigm shifting technology.

Re: Are large language models a threat to digital public goods?

#97

Post by wilg »

The paper title is clickbait and therefore the HN comments are focused on various low-effort reactionary noise, but it seems like an interesting study.

I would not be surprised to find StackOverflow usage dropped significantly because of ChatGPT. It's simply a much more effective tool for getting help with typical programming problems. Not as good of a resource for expert-level or architectural advice, but that's okay with me. Basic "debugging via internet" is much easier to do with an interactive service with lots of knowledge.

It's often pretty helpful just pasting an entire error message with backtrace into ChatGPT and seeing what it thinks.

Re: Are large language models a threat to digital public goods?

#98

Post by spacebanana7 »

FeepingCreature wrote: post #80
bradgessler wrote: post #72

Earlier quoted context omitted.

What happens after LLMs kill off SO and then seek more updated training data? It seems like these “deaths” are either temporary or something else will pop-up that continuously improves and trains an LLMs with proprietary data. There’s def a shift in the social contract of the internet. We’re shifting from “publish a few things you know a lot bit in exchange to read stuff other smart people have shared” to “solve a pr…

Would be cool to have projects write documentation in a form that's annotated for LLMs, ie. premarked topic breaks.

I think open source might benefit a lot from being so easy for LLMs to provide answers for. Hopefully they’ll be able to reason straight from the source code.

Re: Are large language models a threat to digital public goods?

#99

Post by selestify »

somenameforme wrote: post #90
selestify wrote: post #78

Earlier quoted context omitted.

Why is that a concern? When’s the last time you needed to remember a phone number or directions for travel?

Because of the implied effect on things like attention span, memory, and cognition in general. It's the same reason that mastering fundamental mathematics is critical, even though we have calculators. Trends in the Flynn Effect are relevant here. [1] The Flynn Effect is the observation that "real" IQ values were increasing, dramatically, for decades. In recent decades (since births in ~1970), this effect has not only…

What reasons are there to think that the reversed Flynn effect comes from a lack of memorization of rote knowledge like phone numbers, as opposed to attention-sucking social media apps?
Post reply on HN