Helpful:Made by humans

Creating reasonable-sounding bullshit used to be hard.

Estimating the quality of something you found online used to be easier.

Because we only aimed to fool search engines, not each other.

Because organisations used to care cared about consistency and being trustworthy, and consistent(ish) because of it.

Because people argue, and think, and criticise.

Because each person has their style, which helps us estimate how useful their words are.

GenAI doesn't allow most of that.

Because it is bad at doing so at all.

Because it is tuned to care about quantity over quality.

Because it's rewording, so even when it's mostly sourcing reddit and the likes, it does so imperfectly and without the cadre of people correcting what they got wrong.

So it is not only becoming harder and more time-intensive to find good information but also harder to dismiss low-quality information as such.

It's worse than it was in the early nineties, when search engines were terribly clunky things.

Which all wastes so much human time, but somehow we're calling all of that a feature.

And - due to how people work right now - probably misleads more people, leads to more mistakes, with no responsibility taken.

And it's barely even frowned upon - yet.

We somehow trust it more even though it's shown to be flawed and most things it does.

We don't check, we feign ignorance, or even call it a feature.

So we're increasing the rate at which we repeat incorrect things, at alarmingly steady rate.

Which is on us.

Unlike the LLM, which does not understand what is true, we can evaluate truth - and choose not to.

Caring about truth would negate much of the time we saved, after all.

I say we take back responsibility.

I say we aim to have have quality well above the 'always 20% wrong' machine.

I'm not saying that there are no mistakes. I'm saying that there probably are, and owning up to that.

When something is uncertain, or a generalization, it should be mentioned as such. Things in ongoing-research status tends to be marked as (verify), or even hidden altogether.

All mistakes in here are at least genuine human misunderstandings, and tend to self-correct over time.

Weird grammar, bad arguments, come from choice, or a lack of our time.

And you get to complain, I get to go "shit" and correct it.

And you ought to get a resource that has some value to you.

...also, I have my own opinions, that are not the average of everything out there.

I have free will - I will not be flavoured by my makers and then lie about it.