[fsflover]

By this measure, any information you can get whatsoever is biased and there is no reason to trust anything at all.

[satvikpendem]

All information has some sort of bias, as no information can truly be unbiased. There is no reason to trust any specific piece of information but taken in aggregate one can disambiguate the biases.

[fsflover]

So we are in agreement.

[latexr]

The major difference is that right now when you land on a page you can do your due diligence and decide if you trust the source. You can still be tricked, but it’s harder and you can get better at the detection.

With LLMs, everything is given the same importance so you have no idea if the data came from a reputable source or an obvious SEO junk website.

[fsflover]

AI can also provide the sources. And if you need to be certain, you should ask for that.

[latexr]

Those sources are often made up (pages never existed) or outright wrong (they don’t say what the LLM claimed).

Asking an LLM for sources is about as (in)effective as telling it to not make stuff up.

[fsflover]

You seem to be much behind the progress in LLMs. Modern ones provide correct, verifiable citations just fine.

[latexr]

All citations are verifiable, even the wrong ones, and they always were.

And no, modern ones don’t “do it just fine”, they are still frequently wrong. Either you’ve been incredibly lucky, or have just stopped verifying thoroughly.

But if you’re that confident, please do share the exact models which you trust. Proponents always shift goal posts (somehow, every release in perpetuity, those ones are the good ones and everything before were garbage) so let’s avoid the vagueness.