It's the problem. There is not a more important problem on the web.
All that crying about google results deteriorating? It's them trying to filter out exponentially more noise being generated every time unit. LLMs made it so much worse.
You want people to stop caring about something? Flood the web with plausible and crackpot arguments about it from both sides and watch people walk away from the topic in resignation. It used to cost something to pay a legion of trolls, now it takes a couple racks of GPUs and a manageable electricity bill, especially if you're a state actor.
Noise kills utility of any information channel, twitter, HN, google, the web, you name it, it suffers. He probably couldn't have predicted LLM jammers, but I worry the internet's information capacity is about to peak.
Sure, but that’s huge. And probably the reason some (or at least me) are becoming so disillusioned with the web. It’s why I find it hard to believe/trust what I read from news to health advice to product reviews. The noise numbs us to the point where it all starts to feel worthless. This may be one reason ChatGPT et al feel so novel. They get rid of the noise. Of course, they are all built on the back of that noise.
> This may be one reason ChatGPT et al feel so novel. They get rid of the noise.
I think I disagree. LLMs are much better at increasing the noise than they are at getting rid of it. Worse: they are really good at making it look like they get rid of the noise.
To me, LLMs have the potential to break the Web. Instead of search engines crawling the Web and allowing people to search its content, we may have to go back to trusting people: "I read this blog because I know who writes it". The day we can prove that LLMs were used to add tens of thousands of mistakes in Wikipedia (with a political agenda), will we still be able to trust Wikipedia the way we do today?
LLMs have the potential to systematically and automatically destroy the Web.
Yes most of what is on the web is not trustworthy so a Google search will not show trustable sources.
Thus I look at known trustworthy sources e.g. traditional media where their biases are known. Or places like here which get some known reputation.
Reviews long established places like Consumer Reports and Which . Reviews on sites I assume most are false - but I look at negative ones and see if there are some common problems that seem reasonable.
I assume all influencers are liars or at least advertorials.
One good rule is looking how the site/author gets paid for doing what they do. If not obvious assume that it is getting clicks.
baq|2 years ago
All that crying about google results deteriorating? It's them trying to filter out exponentially more noise being generated every time unit. LLMs made it so much worse.
You want people to stop caring about something? Flood the web with plausible and crackpot arguments about it from both sides and watch people walk away from the topic in resignation. It used to cost something to pay a legion of trolls, now it takes a couple racks of GPUs and a manageable electricity bill, especially if you're a state actor.
Noise kills utility of any information channel, twitter, HN, google, the web, you name it, it suffers. He probably couldn't have predicted LLM jammers, but I worry the internet's information capacity is about to peak.
technotarek|2 years ago
palata|2 years ago
I think I disagree. LLMs are much better at increasing the noise than they are at getting rid of it. Worse: they are really good at making it look like they get rid of the noise.
To me, LLMs have the potential to break the Web. Instead of search engines crawling the Web and allowing people to search its content, we may have to go back to trusting people: "I read this blog because I know who writes it". The day we can prove that LLMs were used to add tens of thousands of mistakes in Wikipedia (with a political agenda), will we still be able to trust Wikipedia the way we do today?
LLMs have the potential to systematically and automatically destroy the Web.
pasc1878|2 years ago
Yes most of what is on the web is not trustworthy so a Google search will not show trustable sources.
Thus I look at known trustworthy sources e.g. traditional media where their biases are known. Or places like here which get some known reputation.
Reviews long established places like Consumer Reports and Which . Reviews on sites I assume most are false - but I look at negative ones and see if there are some common problems that seem reasonable.
I assume all influencers are liars or at least advertorials.
One good rule is looking how the site/author gets paid for doing what they do. If not obvious assume that it is getting clicks.
JKCalhoun|2 years ago
I think (given the context of 1995 when we still had land-line modems, etc.) he nailed it.
_dain_|2 years ago
almost everything I do at work, I learned online, not from an in-person teacher