Hacker News Reader

Comment by Mvandenbergh | Hacker News Reader

The problem with at least Facebook and Twitter (and possibly Google) in this analysis is that their attention-grabbing model requires them to analyse and interfere with the data which is coming through their systems. The only way not to do that would be for FB timelines and Twitter feeds to be strictly sequential feeds which at some point in the past they may have been but they are not now.

Once you start basing your business model on selectively making some content more visible, you have already crossed a rubicon between a strictly neutral common carrier and a publisher.

(There is an error in the quote where they use publisher when they mean common carrier, a publisher is precisely an organisation that does have full editorial control of what they put out which is the opposite of what they intend from the rest of the quote.

View on HN

Exactly. At face value, I'm firmly in favor of what Ravikant was saying, but even the simple example of a spam filter proves how it's impossible in practice: without basic spam filtering, something like Twitter would be absolutely unusable in the face of endless crypto and malware spam. At some level, Twitter/FB/et al have to do SOMETHING to shape/filter the volume of traffic they receive.

I don't think calls to regulate Twitter/FB would get much traction if all they were doing was spam filtering. To see it from a different direction: imagine the shit show that would happen if Google started "moderating" gmail content, even though it currently does very aggressive spam filtering.

Indeed. Spam filters, porn filters, swearword filters... all of those have precedents (in mainstream, non-social/internet media) and are perfectly acceptable by most of society.

In addition, notice that even GMail doesn't actually filter spam. It just hides it - it's an assistant, not arbiter of what you should see. It's more than happy for you to adjust the filters (for yourself personally).

Edit: GMail also does prioritization as of recently - but again, the purpose of that is to help users not filter content. Again, those algorithms are also adjustable for each user separately, they just have some "smart" defaults.

>GMail doesn't actually filter spam. It just hides it

This is not accurate. There is a long on-going issue with GMail (and others) wholesale blocking delivery of email from domains/IPs/providers it deems to be spam or otherwise inappropriate. GMail users get no indication/notification, and generally have no idea that email sent to them was simply not delivered.

Kind of. You can't turn their anti-spam off, you can't exclude addresses from it and, worse for me, Gmail can hard bounce if it doesn't like something in your server/IP/content in a way that's not fully documented or reproducible.

You can now. You can create a filter matching whatever criteria you want (including from:foo@example.com or a catch all), and specify matches 'never get flagged as spam'. Its not obvious, but it is there. And the usual limitations of filters apply, such as not being able to match arbitrary headers.

That only works on mail that actually gets delivered to your mailbox. Gmail can and will:

- Reject delivery of mail from sources it has determined to be spammy (e.g. residential IPs) - this mechanism allows the sender to be aware that delivery failed.

- Accept delivery and then silently drop the message somewhere between after accepting delivery but before appearing in the end user's mailbox.

In either case, there is nothing you (the mailbox holder) can do. In the second case, there is no way for anyone to know the message hasn't been delivered.

Things can get flagged outside of your mailbox as so spammy that they will never arrive. This is how every spam filter works that I have seen to date and google is no exception.

It’s still not so easy. At what point does “fake news” become spam?

Its pretty obvious what’s spam and what isn’t. “Fake news” is extremely subjective, so try not to anger half the country :)

Google does censor my email to some extent. It send a number of conservative emails straight to spam.

Political emails tend to get tagged as spam in general. I have someone close to me who is involved in local Democratic Party and progressive organizations. When you go to an event or donate money (say an award banquet), they share your name and spam the crap out of you. A lot of that mail gets tagged as spam as a result.

Conservative outlets work similarly, but they seem to have bigger online operations and are hard to get out of. My in laws get barraged with this stuff, and a mail provider like GMail or Microsoft gets a lot of spam signals from it I’m sure.

Political emails are spam. No matter from which direction they come from. So tagging them as spam seems just the right thing to do.

Political emails often are spam but not all of them. I subscribed to few newsletters from various political parties and I don't want google to block them from me.

It's only spam, if you don't want it.

If you explicitly solicit those emails, eg by signing up to a newsletter, it's not spam.

Some of the conservative emails I did not opt in, so could be considered spam. Others I did opt in and it still gets sent to spam.

On the other hand I get unsolicited email from some Democrat groups, and I never see such emails in spam.

I'm liberal, and lots of the unsolicited Democrat emails go to my spam folder.

I've heard people in the political world talk about buying and selling email lists, which I thought was illegal, but seems to be common practice.

In the early 90's I gave something like $20 to the Sierra Club or a similar group. After a deluge in physical mail asking for donations from around the political spectrum, I never did it again. I left no forwarding address after moving either. Pretty dumb on their part, they traded all their goodwill for a short-term dollar.

Is it doing that because they are “conservative emails” or because they violate spam rules?

If those emails are flagged because they’re “conservative” then I agree with you that needs to he corrected.

However, if those emails are flagged because they violate well known spam rules then let’s help “conservative” email senders to follow the rules. Examples include not using single or double opt-in for signup, not honoring or providing an opt-out link, using emails from a bulk list that includes known flagged emails, not including a plain text copy of the email in the message, and originating from a flagged server. Reputable email services like MailChimp go to great lengths to educate email list admins how to follow these rules & curate high value email lists.

Needs to be more specific as well. Is it intelligent conservatism in the vein of Bill Buckley or a knuckle-dragging ad hominem attacks from the lowbrow right?

Gmail thankfully sends all the republican and democratic party emails I receive straight to spam.

One year, I made a concerted effort to troll some friends by getting them bumper stickers from the opposing party that they supported (e.g. dems got republican bumper stickers, repubs got democrat bumper stickers).

Since then, despite being sure to opt out of mails at the time of purchase, and clicking 'unsubscribe' numerous times, both parties continue to spam me.

Gmail filters these emails into spam for me. Thank goodness.

I'm led to believe that if enough gmail users mark emails as spam, this will get fed across to other users that haven't. So you can probably blame people like me for marking these messages as spam, for it occurring to you.

Or y'know, they could actually honor the 'unsubscribe' request and stop spamming me, and I wouldn't have to mark it (legitimately) as spam.

Then it doesn't censor your email at all.

You can freely view the emails in your spam folder, with one click. You can optionally move the emails to your inbox and train Gmail ("not spam" button) not to place them there in the future.

It's going too far to call it censorship (in any manner) when an email provider places an email into a spam folder that requires one click to view/review.

A lot of overt spam (when gmail is most certain that it's spam) doesn't end up in the spam folder. A counter is incremented and the message is thrown out.

Unless you are talking about email coming from the most blacklisted of IPs, that’s not true.

It's not censorship. It's basic hygiene.

At Google the top metric used to be relevance for a long time (until Amit Singhal was kicked out) for a good reason: it's a metric that kicks out spam, but doesn't pick sides.

Times and leadership has changed since though.

All good things...

> without basic spam filtering, something like Twitter would be absolutely unusable in the face of endless crypto and malware spam

I don't understand this argument at all. Is it so difficult for users to curate the list of accounts they follow?

Replies. It would be impossible to follow a thread. Even today it can get rough following a thread spawned from an elon musk tweet, given all the crypto-scamming.

I think there is a "I know it when I see it" type of argument to say, we will make the strongest effort possible to never censor anything that is not attempting to sell a product but we reserve the right to force advertisements to go through a specific channel.

For better or worse the types Of things which get called to be banned from a platform for being unacceptable speech rarely resemble the specific styles of advertising we classify as spam.

Another argument is let users be ther own spam control.have some idiot acquaintance from middle school who keeps spamming pyramid scheme and/or nazism? Fine, just unfriend them.

As long as you are given control over what content is pushed at you directly, there is no need to sensor the whole platform because of a specific undesirable.

That's interesting because we desperately need spam filtering on our phones right now.

Firefox on Android allows extensions. Including ad blockers.

I was referring to filtering spam phone calls.

Oh, those are annoying. I've been getting some lately, but not nearly as many as other people seem to be getting.

I think Google lets you report numbers as spam? (And, of course, caller id can be faked, but the important bit is that spam will differ statistically from legitimate calls.)

I think if there are going to be filters, then it would make sense for users to be able to opt in or out of filters. That way you can choose what you want to see, instead of having an almighty force that determines what you see. Having global filters that users can't turn off, besides to filter out illegal content, needlessly take away freedom from users.

Usenet dealt with this via killfile. There is no reason why you cannot do client-side filtering (compare /silence with /ignore in IRC). Client-side and server-side each have their pros and cons.

Usenet also had Cancelmoose and the Cabal (TINC).

Both individual and administrative filters are necessary.

Anything can be spam because spam is in the eye of the beholder, but nonetheless, politics is generally not spam. You can have a politically-neutral policy against undesirable content that passes the smell test. It does require that you a) use reasonably objective terms in the definition of "spam", and b) that you resist the inevitable pressure to reinterpret those terms to cover that which others find undesirable political content.

Actually politics generally is spam - in the sense that much of it is knowingly dishonest self-promotion of individuals and social groups.

Just because politics isn't trying to sell you a penis extension doesn't mean it isn't trying to sell you something far worse.

It doesn't even matter how political content is published. What matters is that currently voters get more protection from a faulty toaster than from lies knowingly generated by politicians and corporate PR outfits.

The creation of deliberately false "talking points", fake news, smear campaigns, organised trolling and astroturfing, fake reviews, fake feedback, and other kinds of cognitive pollution should be banned from all forms of media.

Which is not to say this is easy or straightforward. But attempts should be made - because lying to voters and consumers is absolutely toxic to genuine democracy.

There’s something fascinating when you equate penis enlargement spam with political speech.

Btw, laying in politics has happened for like forever. Is there any evidence that it’s worse than in the past?

Both are messages coming to you. Just because they have the right to metaphorically knock on your door doesn't mean both campaigners and vacuum cleaner salesmen don't get the door slammed on them and doormen won't tell both to scram if they try to visit apartments unsolicited. The difference in categort broadness seems classically strawmen as well.

Not true at all.

Spam is unwanted email that you did not sign up to receive and sent from someone who you do not know.

He oversimplified the issue. Phone companies don't make money through ads. FB/Twitter/Google make money by "getting into the weeds". They chose a business model with much more reward (hence much more risk). They're basically running ads businesses. You post about shoes, they sell you shoes. You talk about unemployment, they sell you a candidate ? Gitlab/Microsoft et al have slightly less dependent business models that aren't dependent on ads (for the most part).

People warned that this would happen when Facebook (YouTube, Steam, etc.) started sliding in this direction.

And years before Facebook was even created, a middle way was also proposed where a carrier like Facebook wouldn't be legally responsible for the content they carried, unless they started to select and/or moderate that content, at which point they would lose the carrier status, becoming a simple broadcaster, which would be legally responsible for everything that they would help to broadcast.

> unless they started to select and/or moderate that content, at which point they would lose the carrier status

But since the homepage feed (or any medium really) displays contents in a certain order, some selection must take place.

Typically some algorithm (usually a recommender system together with some business logic) is used to determine which contents from all that's available to you are actually shown to you and in what order.

Bias seems to be an unavoidable part of the design to me.

There are shades of grey. However, we have leaked footage of a Google co-founder saying (at an all-hands meeting no less) that the outcome of a democratic election conflicts with Google's values. There is a lot of room for interpretation there, but there are signals from Google in particular (eg, donation streams; leaked video; the occasional scandal bubbling out) that their management might be seeing the world through a partisan lens.

Abstract ideas of unavoidable bias are only of academic interest; the right wing of politics is justified in seeing Google as a direct political threat. That would not be justified if Google had a strict "no political talk, no political campaigning, we are the Switzerland of the internet" style policy for their workplace.

I honestly don't see the problem with Google using its position to affect peoples' viewpoints. If we're going to allow other industries (such as energy) to hire lobbyists or advertisers to change peoples' viewpoints, and even worse, going straight to the decision-makers with lobbyist $$$, then criticizing Google is hypocritical.

>the outcome of a democratic election conflicts with Google's values.

What's wrong with this? The outcomes of the previous Presidential elections conflicted with many other companies' values. Every political election's outcome conflicts with some company's values, because companies stand to gain or lose depending on the policies enacted by that politician.

> If we're going to allow other industries (such as energy) to hire lobbyists or advertisers to change peoples' viewpoints, and even worse, going straight to the decision-makers with lobbyist $$$, then criticizing Google is hypocritical.

I'm curious that you would assume that as being the default position ?

I'm not making a moral case. I do think there is a moral case as well but it is a very complicated do-unto-others-as-you-would-have-them-do-unto-you style one with some nuances that isn't going to fit into one comment. The case is that Google is potentially a direct threat to the right wing of politics. It would be prudent for the right wing to respond by trying to break Google up and neutering them as a platform, so that there are several successful competitors in all their markets. Realistically it is possible that the moderate left wing could be convinced as well - nobody is served by the risk that an entity as powerful as Google becomes an active propaganda platform. If they aren't even professing neutrality internally then they are on the way to becoming one.

Google could have avoided this situation by not explicitly championing political views inside their organisation.

Also the energy situations you cite aren't really comparable, the companies are only lobbying for things that make them more money and they don't have the same sort of power as Google in the political sphere.

It would be totally hypocritical for the right wing to break up Google. The Democrats tried that with Microsoft back in the late 90s, and as soon as Bush took office the case was dropped because Republicans don't believe in enforcing anti-trust law.

>the companies are only lobbying for things that make them more money

Every company does this if they can, and it's either going to help or hurt some political side. The case you're making here is that Google is bad for Republicans. Maybe, but coal companies are bad for Democrats (they give money to Republicans to help them win races), so why is this OK for coal companies, but not Google? I don't see the difference. As long as other companies or industries are allowed to influence politics with money, it's perfectly OK for Google to influence elections however they want, and it would be wrong to break them up because, as I said before, the Republican party is opposed to anti-trust law.

> It would be totally hypocritical for the right wing to break up Google. The Democrats tried that with Microsoft back in the late 90s, and as soon as Bush took office the case was dropped because Republicans don't believe in enforcing anti-trust law.

Circumstances were different - Microsoft wasn't doing anything particularly political. They aren't pro-Republican. This is the difference between politically attacking an entity because it is a corporation (a bad reason) vs attacking because they are politically active (an acceptable reason).

That is the central point. Google are removing potential defences against a political attack.

> Maybe, but coal companies are bad for Democrats (they give money to Republicans to help them win races), so why is this OK for coal companies, but not Google?

It is OK for Google, they can donate to whoever they want to. The issue is if they are going to be an partisan actor they control too much information and have too much influence on how people gather information.

>It is OK for Google, they can donate to whoever they want to. The issue is if they are going to be an partisan actor they control too much information and have too much influence on how people gather information.

If they can donate to whomever they want, they are also morally correct to control information however they want. Giving money to politicians is bribery, and is much more direct than merely controlling information on the internet. Personally, as long as bribery is legal, I have no problems with Google using a different tactic. It's much more ethical to try to shape peoples' opinions at large than to directly bribe politicians.

That argument ignores scale though, giving money to politicians directly may well be unethical, but it is a path that is open to everyone and is at least somewhat out in the open. Compared to that, Google basically is the internet for a large chunk of people and tracking how they use their index is practically impossible.

Compared to news media where the actors are highly partisan but there are strong voices and opportunities to be heard for all points of view. The alternatives are a lot thinner for Web search and Youtube; and most people would be shocked if it did turn out they were actively pushing a message.

Besides, I'd expect political donation laws to come under attack to. It is a very political question. Google should have stuck to strategies and pronouncements that are neutral so that they were less likely to get involved in partisan politics.

It doesn't really matter whether you see it as ethical or not; what matters is that Google has huge and largely unchallenged reach in a field and appear to be official stances by management on social issues that they do not need to. This makes them a legitimate political target.

Every industry sees the world though a partisan lens to some extent. Those criticisms also apply to energy, education, repair, etc

Exactly. Criticisms about Google's politicking are unjustified as long as lobbyists are allowed in other industries.

Would you mind providing a source for this leak?

It was that one from late last yer. Might have been https://www.breitbart.com/tech/2018/09/12/leaked-video-googl... . There wasn't anything there that was particularly scandalous; it was just a really interesting that this was the state of affairs inside Google.

I mean, it's OK if management have political views, as long as those views don't influence search results.

As long as the experimental process is rigorous and the people are incentivised to use the right metric (and the metric isn't politically biased), then it should be fine.

However, upon making that argument I find myself concerned at the possibility that Google's corporate interests (and perhaps some political interests) may shape which questions get asked, and thus the direction the service takes.

It's quite analagous to Chomsky's views of news organisations in Manufacturing Consent.

Everyone was fine with chronological order.

In that case one can just spam the same (or similar) post every second and your feed will likely be filled with that post.

I know I'm setting up a bit of a straw man here but my point is even chronological order can be exploited.

> I know I'm setting up a bit of a straw man here

Especially as this thread has already mostly said spam filtering (which your example would easily trigger) isn’t in opposition towards a goal of neutral status.

And if you did that, people would unfriend/unfollow you. Problem solved.

> my point is even chronological order can be exploited

The question is, how easy can it be?

The straw man, is indeed, that you can filter such, client- or server-side. If you see the same shit the whole time, you ignore it all. Easy; even IRC clients with scripting had such features in the 90s. Here is a list of techniques used in e-mail filtering [1]

[1] https://en.wikipedia.org/wiki/Email_filtering#See_also

I think it depends on how the recommendation algorithms are implemented. (I have no idea about how Facebook and Twitter recommendation algorithms work)

Let's say for example you have a community contributed labeling system that allows users to label content. A recommendation algorithm might then be written that simply pulls the top 10 highest number of viewed pieces content that have the same labels as a piece of content the viewer just viewed. In this case, there is nothing in the algorithm that evaluates the content itself. The evaluation was done by users of the software, not the software itself.

Simple "neutral" algorithms are generally easy to game. If you choose an algorithm that you already know will be gamed for harassment or whatever then you're essentially choosing that outcome.

The labeling algorithm is just a relatively naive example. More complex "neutral" algorithms can be designed. Mitigations for abuse that are "neutral" themselves may be added before/after initial implementation. (e.g To extend the labeling example, a moderation system could be added that allows users to flag inappropriate labels.) Also, I don't think it is always possible to foresee in what ways a system/algorithm might be abused; therefore, it might be necessary to address said abuse once it reveals itself.

I don't think it's been convincingly argued that ranking or recommendation algorithms necessarily represent any sort of editorial bias. The purpose of these algorithms is to rank content based either on a prediction of what the individual user would want to see based on their existing history or on which content is more broadly engaging in general. In other words, if YouTube decides I really like to watch Ben Shapiro owning the libs with FACTS and LOGIC, that doesn't mean YouTube is making an editorial decision to promote that content, it just means YouTube has profiled me as some sort of conservative. If YouTube profiled me differently, they would be equally happy to recommend TYT videos or whatever.

There may be some editorial bias introduced if the userbase of a given social media site is themselves skewed in a particular direction. For instance, there seems to be some skew in the content that shows up on my Tumblr dashboard that seems fairly irrespective of anything I've subscribed to. But even that isn't necessarily damning for Tumblr, because they're not necessarily making an editorial decision to promote that content.

These platforms (these meaning the ones with the most market cap) need to be advertiser-friendly in order to grow. They inherently require editorial bias.

I'm not an expert on advertising, but I suspect these concerns are often overblown. There are YouTubers who get demonetized--presumably because their content isn't "advertiser-friendly"--who still manage to get individual sponsorship deals. If these companies were serious about being a "common carrier", they would invest more money into developing an advertising market where advertisers could pick and choose the content they wanted to advertise and less money into demonetization.

demonetized video still get ads, it is a two tier system where ad space on demonetized videos is lower in price.

Even if that’s true—which I somewhat doubt—YouTube basically makes demonetized videos disappear in terms of recommendations. They don’t admit it but every YouTuber I’ve seen get demonetized confirms that views basically stop after demonetization, aside from subscribers.

There's a very simple solution here that avoids any and all bias and hidden agendas: simply make everything transparent.

- Users get to create their own profile. For instance in your profile you get to selected whether you would like to see content geared towards men/women/any. And this would follow for a whole slew of other topics, similar to the opaque profile companies are creating on users today without their consent or input.

- All recommended videos have all their profile tags openly shown. And the reason for recommendation is also completely open.

The cool thing about it is that this also works really well with the current neural network based recommendation systems since they output a weighting of things. Simply transform those weightings into a nice output and the user can see exactly what's going on and exactly how they're being viewed. As a nice aside this would also remove a big chunk of the incentive for increasingly aggressive spying as companies try to get ever more extensive profiles on each and every person that uses their products.

One downside here is that users might not be consciously aware of their own preferences. This could again be resolved transparently. "Hey PhilWelch, we've noticed you like videos about underwater basketweaving. Would you like for this interest to be ticked in your profile? [Yes, No, No + Never ask again]

That would be awesome for advanced users but unfortunately bad for mass ergonomics. A shockingly high percentage have never cracked open the settings page even once.

Why would it be hidden away in some settings instead of right in your face with a nice fat button "Change how my videos are recommended..."? Even beyond that though, most online platforms today offer very limited general customization - and so there's not much reason for users to go digging through settings. They are designed to work primarily with the default configuration, and so that becomes a self fulfilling prophecy.

No, if they instead worked to develop better user moderation tools, it would work just as well as it does now, just without their (appearance of) liability. Things like word filters, subscribable block lists, etc.

> The problem with at least Facebook and Twitter (and possibly Google) in this analysis is that their attention-grabbing model requires them to analyse and interfere with the data which is coming through their systems.

Not to mention a completely different topic - ads for scams. When I create a new ad on FB, they keep it in review for at least a couple of hours. But every day I see dozens of ads that are obviously dishonest. Some financial scams, ads for some fishy schemes, some of them badly translated.

When you are a publisher and accept an ad, obviously it's very tempting to claim all responsibility is with the advertiser (it is not, although details vary between jurisdictions). But even if you didn't have any responsibility legally, it still feels terribly wrong from the ethical point of view to get rich by helping scammers cheat some poor chaps of their money.

Not on a daily basis, but I see ads for scams on youtube as well.

Different topic, but the strangest thing I witnessed on youtube was ads for audible during a stolen audible book... So the owner of the content payed the thief (indirectly, but still) for placing an ad during performance of the stolen content. And youtube couldn't care less, as long as money moves...

This is where twitter screwing over third party developers comes in to play. I would have preferred if twitter main feed was sequential and then there were thriving third party alternatives to it. Twitter could have easily monetised by taking a slice of ad revenue from third party developers, or by charging API fees to them.

They could make the algorithm customizable and let the user to control it. That would be much better than the current blackbox where you get uninteresting stories promoted because of misclicking once, and have no way to tell the algorithm about its mistakes.

> They could make the algorithm customizable and let the user to control it.

It often gets it wrong, but that's the general idea, isn't it? You don't customize it by checking some boxes or adding some keywords, you customize it by giving attention to some results more than to others.

Obviously, there's the added layer that they aren't optimizing for your satisfaction but for your business value to them.

This business model is the reason that FB, T and G are dominant platforms worth discussing. Products that are purely transactional are easy to compete with, splitting up markets and driving down prices. This is why true "common carriers", like ISPs, have fought desperately against net neutrality and to capture more customer data -- they want in on the value capture that comes from shaping content, because they are rapidly being commoditized as carriers of bits.

IMO this is a moat in the eye of a lot of libertarian tech philosophy. Most software products become competitive by building networks -- with end users (e.g. social networks), with developers (e.g. open source), or indirectly via content (from media aggregators to spam filtering). Once you're in the business of curating a network, you are implementing values whether you acknowledge them or not.

Publishers don't exercise editorial control? Come on now.

The problem is also that they are advertisers, and I'll be damned if I want ads for my company appearing next to "Illinois Nazis" or ISIS or whatever other garbage.

There's another problem: it becomes a race to the bottom if it's nothing but trolls and garbage content.

And the problem this creates for society is that we censor content that isn't advertiser friendly. We've already seen where this leads with broadcast television, lowest common denominator trash that is safe for advertisers, escaping that advertiser optimum made youtube good in the first place.

I subscribe to a lot of history channels so I've seen the effects of this filtering directly, history contains a lot of talk about nazis and hitler and you can't even mention the big H without being demonetized.

But this is just another example of how advertising is a cancer on society.

> The problem is also that they are advertisers, and I'll be damned if I want ads for my company appearing next to "Illinois Nazis" or ISIS or whatever other garbage.

I don't understand the problem. Why can't Toyota or Proctor and Gamble say it only wants to bid on a certain white list of channels or videos? The only downside is a smaller reach for the advertiser and a smaller "pie" for the "content" company (be it Instagram or YouTube).

I think the main problem is the "content providers" be it Instagram or YouTube want to maximize the ad revenue and any suggestion I have such as strictly reverse chronological timeline by default goes against the idea of "growth hacking".

I would personally love to be a fly in the wall in Netflix or Spotify headquarters. Because in my naive mind, Netflix makes MORE money when people subscribe to Netflix but never watch anything. But then if people watch Netflix a lot, they are more likely to talk about what they watch to others and encourage others to subscribe to Netflix?

Any company probably wants to make sure they make more money the more customers they have? Kind of sounds obvious but I don't think it is the case with every company. I'd imagine there are some businesses where there is an ideal size and if you are bigger, the additional customers kind of follow a law of diminishing returns? I can't think of any examples but would love to hear if anyone reading this can...

Reply on news.ycombinator.com