Businesses which scale by hiring are much harder to grow and tend to be a lot less profitable. As a result, internet companies tend to see such businesses as not worth being in. Sure, there are lots of successful companies which operate this way, but Google sees it as an unacceptably high opportunity cost. Rather than pour resources into growing a scale by hiring business, they would rather put the resources toward a different business that they think they can scale with non-linear hiring.
Given this, it surprises me that Google is going after the enterprise cloud business. It is pretty well known that such businesses scale fairly linearly with sales and support staff. My understanding is that they know this and are attempting to scale the business this way instead of their usual strategy of ignoring the issue and hoping it will go away.
When something goes wrong, Apple, Amazon and Netflix (as examples) all manage to provide free, robust customer service.
Google can afford to support it's customers. They choose not to.
Ha ha. Last I heard, Google were actively trying to remove their engineers from client engagements, and replace them with third party "Google preferred suppliers". They really don't want to ever be face to face with a customer.
It's too easy to forget that if you do not take care of the users, they will leave and take the payers with them. Right now it's hard to escape Google.
But a little river carved the Grand Canyon, and as more people leave the services and provide easy to use alternatives, more people will follow. My guess is that, even though it might take 20 years if Google does not redirect their focus to be more equilaterally supportive of their users then Google will inevitably become a modern Yahoo.
Rim to rim, the canyon is pretty darned wide.
GCP can't get their act together to get a consistent enterprise sales model done....
Maybe the root issue was letting 4 private companies control the entire public debate with dangerous moderation practices.
This is the reason I was able to sway the C suite at my last position to avoid anything Google. I'm estimating it's probably around $9m/month revenue they lost from GCE and GApps.
They're simply not trustworthy, no/terrible support, and capricious about arbitrarily holding user's data hostage when some random automated abuse system decides you're horrible.
System Engineers don't let fellow system engineers use anything google for production purposes.
"The McNamara fallacy (also known as the quantitative fallacy), named for Robert McNamara, the US Secretary of Defense from 1961 to 1968, involves making a decision based solely on quantitative observations (or metrics) and ignoring all others. The reason given is often that these other observations cannot be proven."
That is indeed....your opinion.
To be a person is to not be reduced to an array of arbitrary scores in a worksheet.
They are happy to reap the benefits so it is fair to put the costs on them.
How are you helping?
This is not a new attack in either type or scope. It's merely more visible.
> How are you helping?
Not at all. I don't benefit from YouTube's business model so why do you think that I have some obligation to help YouTube?
Of course, since you think that I have some obligation, you surely do. What are you doing besides attacking people who point out that this is a foreseeable consequences of YT's business decisions?
I don’t think people expect much from YouTube.
Dealing with bots, and content moderation in general, is an arms race. You're fighting another person who could be just as smart as you in developing evasion strategies as you are in countering them. Getting it right all of the time is impossible.
Either way, yes, there are techniques for improving performance in the presence of an adversary but it doesn't negate the problem; anything you do will be countered eventually. Clever techniques only buy you more time before someone figures out a way around it.
For the company that loves encouraging obnoxious Captcha's to self-fund their AI/ML research and data harvest the web, they surely don't seem to like using it on their own platforms.
Google still has to routinely invent new techniques to stop spam. This problem is not solved - it's an ongoing arms race.
It is also a vastly easier problem - spam filters at google scale work because they see the same spam across zillions of accounts. This is not the case for youtube, where a spammer provides them with far fewer samples, and those are across a much wider space.
These problems are barely related.
(In fact, I wonder who these spammers are targeting. Surely the average YT user is smart enough to recognize blatant spam?)
Some random dude built an anti youtube comment spam tool. Youtube could absolutely fix this problem, they just don't care
5.7 - restricted terms
8.0 - username looks like base64
Gmail has the worst spam filter of any email provider I use today, excluding the one email server I have running on my domain that gets almost no email and has no filter. Or, well, spam filtering is a lesser problem on my server too because it doesn't block nearly all the email that I care about.
80-90% of the emails in my "Spam" folder are not spam.
It's worse than no filtering at all, since I still get exposed to the spam, but now I have to check two inboxes instead of just one, and then jump through annoying hoops to view images or access links in the second one.
Compromising a system that involves 100k people is generally harder than compromising a computer algo.
Using that information I would implement a system that would figure out the topic and prevent bots flagging that specific content only.
Companies that are providing such an amazingly affordable service because they're just skipping out on doing moderation don't get to use "Well, doing it the right way would be too expensive" as a defense - that's how you end up with Uber. Uber broke laws, Uber shouldn't exist at this point, something like Uber should exist, but Travis Kalanick should have been fined into near non-existence and not currently be sitting happy on 2.8 billion. We, as a society, need to have standards.
So you'd rather have the status quo of a few decades ago (ie. large media organizations acting as gatekeepers), rather than the democratized ecosystem we have now?
I'd prefer neither system, and I don't think it's a binary choice - but if forced to choose between the two I'd probably go with the monolithic status-quo machine (even though I'm extremely progressive and they hated us).
But we mandate it because that's the world we want to live in.
The real conversation here is about Alphabet and Facebook's margins. If they're not doing enough, they can spend more and do more.
I'm not sure, but I don't think the answer is that it obviously won't materially change YouTube's profitability.
They have a monopoly on search, that they use to censor any competing video hosting. If they go and start censoring their own service too, this is a very serious problem.
If the profitability of Youtube is on the table, then so should be antitrust, and Google should not be allowed to own it.
Even more important from Google's point of view is it was the way to push Chrome to dominance. And a way to keep it there ahead of any potential competitors.
Which is valuable because of the data and ability to block ad blockers. That is, Google's ad ecosystem.
How did YT push Chrome?
It's similar to how "DOS ain't done until Lotus won't run", only with YouTube and other browsers.
Google optimizes YT for Chromium-based renderers. Or rather, Google pessimizes YT for other renderers.
It would cost 2.5% of their profits if they had an entire US workforce and every video needed to be viewed in its entirety.
There are a plethora of much more conservatively sized video uploading sites that vet and have specific contracts with their content producers.
Clearly you wouldn't need to review all of it. You could use AI to identify things for review (as opposed to remove them), limit it to people with a number of downvotes or views. Hell, you could just pay people to handle the reviews/escalations/appeals.
Although, Google totally could afford it. It would cost less than 2.5% of their profits to pay US minimum wage to review every video.
So you think employing those hundreds of thousands of workers would incur no office space costs, no IT system costs, no management, no service staff, no HR, no payroll, no recruitment, no training. Wow. And your talking about Google profits, but YouTube is a business itself that needs to make money. Why does the rest of Google need to subsidise it?
Well, my numbers do allow for a $50MM/year in the various IT/management/training. I would think you could have people working from home, so office/etc. costs would be minimal. But I also made them US based and reviewing every random cat video's full length that is only seen by 1 person. Some savings could be achieved.
> your talking about Google profits, but YouTube is a business itself that needs to make money. Why does the rest of Google need to subsidise it?
Okay, fine. YouTube could internally have added the costs of reviewing every video's full length with US based people in 2021 it would have only cut in half their increase in profit over 2020.
That’s hilarious. The $50m was a rounding error you forgot to account for. Given a work force of several hundred thousand, that would just about pay for an office chair each, let alone any other equipment or infrastructure, or office space to put the chair in.
I have no idea why minimum wage workers would get an office chair worth a grand, or whatever you're assuming. You can have people do the work from their own couches and phones.
For $50MM/year I could set up the infrastructure to manage the process.
Share-holders don't have a natural right to profit off of societally destructive behavior.
YouTube does not care. They only care when large popular cable TV networks or large partners leave YouTube. When that happens, that is a big frown for advertisers. Small users, creators or live-streamers have no chance against being listened to by YouTube. The algorithms won't ban the partners throwing cash at them but the small users will get banned automatically and YouTube doesn't care and won't care.
That is how it is on YouTube and their private platform with their rules or ToS and we have known for years that they continue to do this and ultimately, they will NEVER change.
Articles about YT say that 500 hours of video are uploaded every minute. Moderation has to be automated. You cannot throw more humans effectively at this problem.
So about 10,000 moderators required. This costs under a billion per year. Given that Youtube brings in $20B + of revenue, I'm sure this would not kill their profits.
https://www.bbc.com/news/technology-52642633
So you now have a situation where you are making thousands ill and in doing so will incur a cost much higher than 1B per year.
But yes, it can be done.
But, clearly YT's assessment is that the harm of a "bad" video being served on their platform greatly outweighs the harm of incorrect takedowns (which can be appealed). I will call this "The Principle Of Can't Unsee."
[0] https://www-nytimes-com.translate.goog/2017/08/22/world/midd...
It's not like they're going to suffer for it.
even classifiers with 99.99% precision means hundreds of thousands of videos and channels will be incorrectly marked as abusive every year - we just need to accept that and hope that the automated systems are improved over time.
Side note on communication here.
Any time someone uses a phrase like "Please educate yourself on $THING", they lose a lot of credibility, at least in the context of that communication. Why?
1) It comes across as incredibly dismissive of both any knowledge that that individual may have - especially on HN, I'm pretty sure we're all familiar with YouTube's scale - and of the idea that there may be missing information, on either the situation or context for the other persons point of view.
2) It's quite condescending as well - and if someone thinks that you are coming at a conversation from a combative place, they won't want to listen to what you are saying. Why should they listen if they think you don't care?
3) It's often used to paint broad strokes in places broad strokes may not be appropriate, and the phrase being attached to that makes it lose credibility even when it is appropriate.
---
Ok, now a response to the actual content:
I don't know why we've decided to allow things to exist "at scale" willy nilly. If something becomes a net negative "at scale", don't allow it to be "at scale".
No one is entitled to a business model.
Hear! Hear! We shouldn't have to give over society to the fattest, most insatiable stomachs masquerading as technology companies.
I've taken to thinking of this sort of scaling as being the digital equivalent of race to the bottom (i.e. cheap mass-produced garbage products).
Because we allow things to exist by default. Because we're a free society.
If you want an exception, then recognize and carve that out within the democratic process, not by preemptively applying authoritarianism.
You mean like I'm advocating for now? We've let it exist, and I am now advocating for it to be broken up - like I've been for a while.
Edit: I should specify - I am advocating for platforms at the scale of Facebook/Youtube/Twitter to be broken up in general, with an explicit preference towards federated services. We've let them run, and it's time for this generations round of trust busting. I'm am advocating for it to be run through a legal manner though, and discussed and acted upon via democratic means.
I think that all over comments on my gp we've gotten off topic, my email is available in my profile for anyone who wants to discuss this sort of thing further in any manner.
Mastodon seems promising as a social media replacement, and PeerTube has the tech for YouTube replacements. Both of these two are using ActivityPub, which is on Wikipedia here: https://en.wikipedia.org/wiki/ActivityPub
The problem comes with content. Matrix is in the strongest position at the moment imho - moving chat platform has the least network effect, and it has the most effective (only?) bridges between "standard" centralized services out of the three. Mastodon has a decent presence, especially of tech people, but not so much outside of the tech or "got banned from twitter" bubbles. PeerTube is in the weakest position of the three afaict - it's mostly again the tech bubble and the "got banned from youtube bubble", but with a much worse distribution.
The tech on Matrix and Mastodon at least seems well established though[0], and I think both would succeed if not for the massive network effects of YouTube/Facebook/Twitter/etc.
My honest opinion is that even if a federated solution doesn't exist or isn't far enough along yet for any of the above it would be better to break down the current state of things and force the creation of one.
[0] I've not used Peertube really, so I don't know how far along it is.
Thanks for the info.
Network effects make this a losing battle, and that will not change until it is forced to - this my advocacy for it to be done via law.
I should not that I have and do occasionally watch content on peertube, there's just sadly not much there.
The conference looks quite interesting, I'll have to try to watch some of the talks this year!
>No one is entitled to a business model.
When it's just ordinary American conservatives being censored by social media, somehow nobody says things like that. Ban Trump and allow the Taliban? There's nothing wrong with that, they're private individuals, they can ban whatever they want. What do you mean nobody is entitled to a business model? Go away with your freeze peach.
<us-pov> FWIW, I see much more domestic harm in letting unchecked lies be spouted by US politicians and having that beamed into our minds non-stop then in allowing people to see atrocities committed by people on the other side of the world.
Yes that goes both ways, I'm an independent.
Edited again: Got too deep into politics on that first edit. Suffice it to say one party is worse than the other in my opinion, and that may color my POV. I still expect to see those who are lying or encouraging violence banned/fact-checked/whatever - on both sides equally.
If one side still has it happen more under those circumstances, I guess that says more about the side then the circumstance. </us-pov>
Also, free speech isn't immunity from repercussions. If I say something incredibly dumb, I expect to be mocked. If I say something hateful/dangerous, I expect people to avoid being associated with me - including being banned under certain circumstances.
I wish we wouldn't equate "free speech" with the first amendment. IMO "free speech" is a much broader concept than the specific legal protections outlined in 1A.
Mocking people for saying things you disagree with ("dumb things") doesn't violate 1A, but that doesn't mean it's aligned with the general ethos of free speech.
I feel I failed to communicate all aspects of my POV above - I definitely agree with this. Also yeah, this did get off topic and I apologize for pushing that even further just now.
In general, I try to advocate more for polite conversation and actual exchange/malleability of ideas - I just also expect for not everyone to agree with that, and to get mocked if I say something stupid.
The bigger thing is association - If I say something that someone else finds reprehensible, I bear them no ill will for trying to stop themselves from being associated with me, regardless of if that's through banning/mocking/public press conference[0]/a music video[0]/anything else.
[0] I'm glad I don't have a big enough platform to worry about these though.
Well, people are facing repercussions for viewing pro-Ukraine material. There's no immunity from repercussions, right?
(Or more precisely, people are facing repercussions for violating the algorithms that Google chose. But they're still repercussions.)
The point is that there's a double standard. If social media censors someone you hate, it's "there's no freedom from repercussions" or "they're a private group so they can ban anyone they want" or "their system, their rules". If they censor someone you like, this suddenly changes to "nobody is entitled to a business model" and "Youtube has no right to operate like that".
The issue is when automated banning leads to there not being a human in the loop, or when "well it's just the algorithm" becomes the go to excuse.
Please be less smug.
YouTube has no right or requirement to operate at that scale. They've consciously decided to scale past what they're able to manage well. This is no different than a developer who produces terrible non-functional code but defends it to their boss by pointing out the line count is really high.
I'd be impressed by their scale only if they managed to maintain some level of quality also.
so only you can be smug, eh ?
> They've consciously decided to scale past what they're able to manage well.
That is not true. They are doing quite well considering the overall scale.
> I'd be impressed by their scale only if they managed to maintain some level of quality also.
So you assess quality by random one-off/anecdotal articles always highlighting the negative ? How do you know what their quality is when you can never know the number of abusive videos correctly blocked ?
You're being very condescending when it's totally unwarranted. I assure you people are aware of the scale of YouTube.
But why should we give YouTube a free pass? You're talking about the service as if it's a force of nature rather than something human created and managed. Why are they entitled to operate at a scale beyond their own control, all while extracting huge profits?
Some people just want to throw their hands up: "Dealing with infinite uploads is hard!" "It doesn't scale!" without realizing the youtube is responsible for the very problem they created. They designed it this way!
If anything, YouTube's scale gives it a better chance of being able to address this issue, since there are likely fixed costs here that YouTube can better fund (obviously investing in better automated screening, but also setting up the workflows for human moderation, etc.) I think the only benefit a smaller scale competitor has is that they're less visible and unlikely to be "caught" for hosting content they "shouldn't". But if the rules applied equally to everyone, I think YouTube is strictly at an advantage in being able to implement better human moderation.
If I were designing from scratch, I would probably use a Peertube/Mastodon type federated system.
Each host would responsible for the videos they host - meaning they would have incentive to limit videos uploaded to any given host to an amount they could handle, or only allow trusted uploaders to their host.
Then you would have the federation so that you could allow hosts that you didn't have issues with, while blocking those you did. Eventually, there would evolve safe/block lists - like with email and adblockers - so it wouldn't even require that much effort on the parts of those running the hosts.
Secondly, for the folks saying why should we allow youtube to scale beyond what it can handle, or why should we accept the 99.99% precision ?
It is because I do not see a futuristic society in which humans are doing work like "manual reviews" - to progress humanity we need to automate such manual labour so I support the fact that companies are testing the limits of automatic content moderation - and working to improve it - I see it as a win overall. Adding human reviewers would be a step in the backward direction IMO.
Why are you assuming that HN readers don't understand YT scale?
You do know that HN has readers who have built things at YouTube scale, including YouTube itself, right?