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

Three sites made 215,128 “best software” pages for AI. Perplexity cites them

trellner.com

Post reply on HN
Print view View original

121–130 of 265 posts

Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them

#121

Post by kspacewalk2 »

locknitpicker wrote: post #63
SoftTalker wrote: post #54

Earlier quoted context omitted.

And it's all because of ads. The incentives in an ad-funded internet are just always going to lead to this sort of thing. The most important thing is getting the user to load your page, not actually satisfying their query. Let's hope the LLM model continues to be paying for credits, because any that move to ad revenue will become useless for real work.

> And it's all because of ads. Not really. A while ago there was a news piece stating that Israel was behind a series of fake think-tanks with very accessible websites which were created with the express purpose of feeding AI agents with alternative facts aligned with their foreign policy. If anyone has the link at hand, please post it.

If by "not really" you mean it's not all because of ads, some governments dabble in it too, you're right. But it's still overwhelmingly because of ads.

Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them

#123

Post by Retr0id »

xpct wrote: post #18

If I recall correctly, there were some papers which suggested that LLMs favor LLM-generated passages over human written ones. I can consistently reproduce this by asking Claude which code snippet it prefers: the one it generated in a different chat, or one that I refactored for my own needs and find more useful. It always picks its own :) I've also experienced that both Claude and Codex routinely include generated we…

I was giving local models a try recently, I think it was Qwen 3.6 I was trying at the time. I gave it a codebase and just asked it to review it. Its main feedback was that the comments and documentation were excellently written, but they were all Opus 5 slop.

Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them

#124

Post by bastawhiz »

xpct wrote: post #18

If I recall correctly, there were some papers which suggested that LLMs favor LLM-generated passages over human written ones. I can consistently reproduce this by asking Claude which code snippet it prefers: the one it generated in a different chat, or one that I refactored for my own needs and find more useful. It always picks its own :) I've also experienced that both Claude and Codex routinely include generated we…

I don't have an oai subscription to try, but I'd be interested to know if Codex picks Claude's code over a human's and vice versa.

Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them

#126

Post by csallen »

kspacewalk2 wrote: post #121
locknitpicker wrote: post #63

Earlier quoted context omitted.

> And it's all because of ads. Not really. A while ago there was a news piece stating that Israel was behind a series of fake think-tanks with very accessible websites which were created with the express purpose of feeding AI agents with alternative facts aligned with their foreign policy. If anyone has the link at hand, please post it.

If by "not really" you mean it's not all because of ads, some governments dabble in it too, you're right. But it's still overwhelmingly because of ads.

It's just incentives in general. Humans are self-motivated creatures, and in the absence of meaningful consequences they'll often do what they're incentivized to do, even if it harms others.

The internet is uniquely devoid of consequences (esp. reputational consequences, social faux pas, etc.) and makes effort expenditure minimal. So you get lots of bad behavior.

I think "ads vs not ads" is maybe the wrong way to model it. Ultimately people are just doing what benefits themselves across every dimension possible.

Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them

#127

Post by trimethylpurine »

locknitpicker wrote: post #63
SoftTalker wrote: post #54

Earlier quoted context omitted.

And it's all because of ads. The incentives in an ad-funded internet are just always going to lead to this sort of thing. The most important thing is getting the user to load your page, not actually satisfying their query. Let's hope the LLM model continues to be paying for credits, because any that move to ad revenue will become useless for real work.

> And it's all because of ads. Not really. A while ago there was a news piece stating that Israel was behind a series of fake think-tanks with very accessible websites which were created with the express purpose of feeding AI agents with alternative facts aligned with their foreign policy. If anyone has the link at hand, please post it.

You wrote "not really" and then described a literal ad.

Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them

#128

Post by embedding-shape »

cainxinth wrote: post #86
xpct wrote: post #18

If I recall correctly, there were some papers which suggested that LLMs favor LLM-generated passages over human written ones. I can consistently reproduce this by asking Claude which code snippet it prefers: the one it generated in a different chat, or one that I refactored for my own needs and find more useful. It always picks its own :) I've also experienced that both Claude and Codex routinely include generated we…

Take a random essay and add in a bunch of the phrases that LLMs love like “load-bearing,” “crucial,” structural,” and “woven,” and then submit the original and the edited version to an LLM and ask which is better. It will choose the second one virtually every time. They have ingrained biases that associate those words with good writing and arguments.

This is why using other LLMs as scorers for benchmarks and evaluations is such a bad idea, they'll have preferences you can't anticipate and won't understand immediately.

Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them

#129

Post by moron4hire »

rectang wrote: post #113
SoftTalker wrote: post #54

Earlier quoted context omitted.

And it's all because of ads. The incentives in an ad-funded internet are just always going to lead to this sort of thing. The most important thing is getting the user to load your page, not actually satisfying their query. Let's hope the LLM model continues to be paying for credits, because any that move to ad revenue will become useless for real work.

> Let's hope the LLM model continues to be paying for credits LLM vendors make this hard because you can't trust them with your session data. Yesterday you were opted out of training, then suddenly today you're opted in. It's an extension of the idea that they don't need to care about anybody's copyright. They don't care about preserving the security or privacy of customer data, because there is negligible incentive…

But don't you want Technojesus to come save you from never having learned how to invert a binary tree?

Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them

#130

Post by mbesto »

Aurornis wrote: post #8

I used one of the 12-month free Perplexity offers when they were everywhere. It felt slightly useful at first for simple queries where I didn’t want to go through the top 10 Google results manually. If I was looking for a specific recipe I remembered or a help page or user manual it would usually find it quickly. Then they started optimizing for speed of responses over quality of results. I can enter a query and see…

I use perplexity simply to avoid underlying vendor lock in.
Post reply on HN