[flagged]
Three sites made 215,128 “best software” pages for AI. Perplexity cites them
181–190 of 265 posts
Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them
#182Honestly, I will just flag every post that is entirely AI slop from now on. This has to stop. The home page for this "independent research firm" is also 100% nonsense [1]. "The record a machine reads is not the one a company writes.". Ironically this low-effort spam is exactly what this report warns about, and does not belong in HN - or anywhere else. [1] https://trellner.com/
Unfortunately a large amount of people on this particular site love this, and they will meet your disgust with equal enthusiasm. Many people on here fancy themselves kindred spirits with the most ghoulish VCs you could imagine, and to them, an AI filled internet is a sign the system is working, and approval is a chance to be part of the elite who "get it."
Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them
#183Earlier quoted context omitted.
> 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 :) A better question to ask for each snippet is "Estimate the seniority and competence of the developer who wrote the following code, ignoring bugs that linters or LLMs can catch and focus only on structure…
You're asking basically to ignore bugs and correctness. Can it be a useful comparison?
> You're asking basically to ignore bugs and correctness.
Not ignore correctness, just bugs that will be caught by tooling.
> Can it be a useful comparison?
IME, yes. LLMs in an agent-loop are trivially able to write spaghetti code that will never do an off-by-one error or something else that is easily caught by tooling, which is not something humans can do.
Judging code on whether it has bugs easily caught by tooling is pointless - LLMs are running the tooling in a loop anyway, so no matter how bad or poor their code actually is, it never exhibits bugs that are caught by tooling.
Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them
#184It's difficult to read more than a few sentences, when this itself is clearly a Claude artifact.
I had the opposite reaction: maybe it was not written by AI, but it would have been better if it was.
Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them
#185[dead]
Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them
#186Earlier quoted context omitted.
You're asking basically to ignore bugs and correctness. Can it be a useful comparison?
> You're asking basically to ignore bugs and correctness. Not ignore correctness, just bugs that will be caught by tooling. > Can it be a useful comparison? IME, yes. LLMs in an agent-loop are trivially able to write spaghetti code that will never do an off-by-one error or something else that is easily caught by tooling, which is not something humans can do. Judging code on whether it has bugs easily caught by toolin…
> just bugs that will be caught by tooling
Bugs are bugs, if the instruction is "ignore bugs except for those that can be caught by you" then the instruction is basically "ignore bugs".
And it implies "ignore correctness" because when program is incorrect we usually refer to it as a... you guessed it, "bug".
Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them
#187Earlier 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.
Can we just ban advertisements and "marketing" at a political level please?
In some countries it is, to some extent. In Norway for example, it is illegal to broadcast ads for political parties, alcohol, tobacco, plus basically any ads specifically targeting children (though the last one is often violated).
Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them
#188If 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 expect that after enough cycles of it training on it’s own output, it will converge on the One True Paragraph filled to saturation with AI cliches, and will refuse output anything else.
Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them
#189Google won't even crawl my personal blog. How can I get AI to read it over and over again?
Re: Three sites made 215,128 “best software” pages for AI. Perplexity cites them
#190Earlier quoted context omitted.
I noticed this when I tried to get LLMs to play text adventures. Early on in my experiments, I wanted to give them hints when they got stuck at a puzzle. I did this by stopping the loop, injecting thoughts into the LLMs own persistent scratchpad (as if it had thought of that itself), and then starting the loop up again.[1] That way, the LLM would read what it had intended to remember from the previous turn including…
Did you ever try to ask a chatbot to rephrase your hint in its own words? If it prefers LLM generated text, surely that would help.
This is very similar to how you rewrite queries so that similarity search works better... which ... well, it's similarity search all the way down pretty much.