Hacker News new | ask | show | jobs
by eropple 2671 days ago
People are pressuring YouTube to hire and train moderators to competently and soberly evaluate context and take intelligent action.

Nobody is pressuring YouTube to blow channels away because one of their ML algorithms hit a probability threshold.

1 comments

It's simply absurd for YouTube to manually moderate comments at the scale they currently operate. If you force them to do that, it won't be profitable, and you'll end up with YouTube blowing channels away because they can't afford to host them anyway.
Why? Why is it absurd?

You are thinking at human scales, and that is understandable, but Google doesn't think at human scales and it's only "absurd" if you think that Google has the inalienable right to the smallest possible cost of goods sold, even if that means offloading their externalities onto everyone else.

It is probably obvious that I do not. You shouldn't, either.

At Google's scale, trained-but-unskilled workers are not expensive. They are not cheap, but they are not expensive. And Google makes a lot of money. This is a common throughline from large societally-threatening, socialize-our-externalities-but-never-our-profits companies from Facebook to Google: "doing something correctly, or even trying to, would just cost too much money, so we should continue our societal-termite ways!" Until these unwatched monsters--and that is, I stress, the default state of the corporation, it is only the threat of the society that grants them their charter taking it away that adds even a speck of decency to them--prove, prove, that they somehow just can't survive by reducing incomprehensible net revenues to merely gigantic, then I will continue to operate on the understanding that they don't want to. Which I tend to think is a much, much more realistic thing.

I don't care. They fix their product or YouTube delenda est. Either is preferable to the current situation.

> And Google makes a lot of money.

They make money by not spending it when they can get the same outcome for free[1]. Also, Search and Adwords make money, YouTube is getting by[2] (relatively). Why should other divisions subsidize a loss-making YouTube? Some channels don't make enough money relative to number of comments to be financially viable (no matter how cheap the moderators are) - Google has simply outsourced this decision to individual channel owners.

1. Google user's do a lot of things for free already, e.g. Map POIs

2. My guess - they don't breakout YT's income/expenses in fincancial reports https://www.marketwatch.com/story/the-sec-wants-to-know-why-...

I understand that. I also understand that stuff like YouTube is effectively becoming the public square of the twenty-first century and if a company wants to own that, they can deal with not making all the money off of it that they could possibly, theoretically, make.

People matter more than corporations. Society matters more than corporations. I'm comfortable asserting that it would be better for Google to close YouTube down than to let an organ of growing central importance to society at large become what it's obviously starting to become; something less damaging than that neglectful caretakership can arise in its wake.

Are you seriously suggesting that disabling comments on a certain type of video content is more damaging to society than losing a global engine of content creation and community?

YouTube benefits society immensely by sustaining a very expensive 21st century public square. If we as a society want to have that - and I at least very much do - we can deal with not making comments on all the videos we could theoretically comment upon.

I am seriously suggesting that this is not something that can be algorithmically determined. I'm quite OK with all manner of content not having comments enabled. I'm not okay with unthinkingly stupid false positives all over the place harming creatives' (actual creatives) ability to feed themselves, and those false positives are overwhelmingly caused by bad heuristics and objectively dumb algorithmic decision-making.

Feeding humans into The Machine, having The Machine make context-free, alarmingly inaccurate, and functionally beyond-appeal decisions--because the appeal process doesn't scale either, we are so frequently told, when it isn't just "drop the appeal on the floor--is bad. If Google has no other answer than Feed The Machine, then The Machine should be considered inimical to humans and should be dismantled.

But, of course, The Machine is not necessary; that's a convenient fiction to paint the problem as a dilemma of "no YouTube" and "some unaccountable algorithm runs YouTube and decide what you can see, free to lead kids from Let's Plays to Nazi agitprop and pedophiles to their spank bait." It's just that the Machine is cheaper, you know? And that's really, and literally, all.

So, quality instead of quantity? How horrible.