> But most of the time we spend on the web today is no longer on the open Internet - it's on private services like Facebook, Twitter and LinkedIn.
Ultimately there are three good reasons for this, which the author doesn't address at all.
The first is spam and abuse. The problem with a purely "open web" is that there's no solution to those, because of bad actors. The only solution we've found so far are centralized organizations which can run sophisticated machine learning and hire thousands of moderators.
The second is aggregation and discoverability. What good is it to publish if nobody finds you? Matching what people want to read, with the content they'll like (and not turning it into a firehose) is a really hard problem which, again, open standards basically do nothing for.
And then of course the third is monetization. There's a lot of amazing content on, e.g. YouTube, that wouldn't have been created if it weren't for monetization, because it takes content creators time that they'd otherwise have to spend at a job.
Now if we can create open standards that work as well as centralized solutions for all three of these... then we're talking.
The pre-giant-conglomerate solution to those things was the (lack of) scale.
Aggregation and discovery was low-volume stuff like web rings and humble links.
Abuse had a much lower payoff due to the lower scale. Communities were mostly small, and abusers were small. Bigger communities were bigger targets, but could afford to pay people.
If ten thousand people wanted to attack one of them, they could, for sure - but people having public discussions on FB or Twitter aren't immune to a group that size either. Scale and centralization hasn't had much benefit there.
Monetization is certainly the difference. But is it worth it in its current form? If you want to reach a huge audience, these mega-platforms make it much easier to do so. You don't even have to persuade that many people, if the algorithm picks you up and starts dropping you in people's feed!
I'd argue that the blind algorithm-driven promotion is the single biggest problem in the big social media and search platforms (Google, FB, Twitter, even Amazon 3rd party sellers).
--
But I think the interesting thing people usually miss is that the old smaller internet is still there, for people looking for niche discussions and communities outside of the huge platforms.
> The first is spam and abuse. The problem with a purely "open web" is that there's no solution to those, because of bad actors. The only solution we've found so far are centralized organizations which can run sophisticated machine learning and hire thousands of moderators.
Not true. It's the only solution when your focus is growth, and so you permit open signups from anywhere needing only something like an email address. This enables brigading, bots and abuse with fake and anonymous accounts.
If a web of trust were involved, you have an introduction chain for every user, and a bad actor revealing themselves casts suspicion on the whole chain. You can revoke the whole chain in one go which is far more effective than picking bad actors and bots off one by one. The investment needed to sneak in is also considerably higher.
So the status quo is not the only solution, it's the lazy solution focused on growing as fast as possible with no regard for the challenges that raises.
> The second is aggregation and discoverability. What good is it to publish if nobody finds you? Matching what people want to read, with the content they'll like (and not turning it into a firehose) is a really hard problem which, again, open standards basically do nothing for.
Relying on a web of trust for signups makes the second problem far worse. If you're only creating a service for your friends and family it works great, if you want to expand it to the general public it is impossible. The encrypted email folks have been trying to get the web of trust to work for almost 30 years without success.
> If you're only creating a service for your friends and family it works great, if you want to expand it to the general public it is impossible.
Not impossible at all. Send out invitations and let them pull in their groups, as but one option. GMail was initially invitation only if you recall, and it was crazy popular even then.
> The encrypted email folks have been trying to get the web of trust to work for almost 30 years without success.
Web of trust has been working fine for all of those 30 years, so I really don't know what you're talking about.
What hasn't "worked" is commercial buy-in by email clients and, more importantly, email providers who want to mine your data. People choose the free, easy email options (gmail, yahoo, etc.) without realizing how much personal information they're disclosing.
So to claim that the commercial interests who are interested in violating your privacy and mining your information haven't bought into the web of trust is a pretty solid endorsement, not a downside.
GMail was offering email, a service people already wanted and could access. Email is the first social media platform, Google was just offering a fancy new client to an existing platform. A better example would be Google Wave, or Google Hangouts, or Google+, or Google Talk.
The fact you think the Web of Trust is working fine when it is still basically impossible to send an encrypted email to anybody even when you have your client set up properly is somewhat mind boggling to me. A solution is not working if it is impossible to use.
> Google was just offering a fancy new client to an existing platform
The point is that invitation-only works just fine if the incentives align.
> The fact you think the Web of Trust is working fine when it is still basically impossible to send an encrypted email to anybody even when you have your client set up properly is somewhat mind boggling to me.
I used encrypted PGP email from within gmail using a browser addon for years. It was easy, so "impossible to use" is simply false. The incentives to make it easier simply don't exist because commericial interests don't value your privacy or systems that are robust against abuse, so it gets no investment and whatever makes growth easier and faster gets all the investment.
The fact remains that the original claim that "the only solution we have" are centralized organizations with moderation and machine learning is simply false. Web of trust is a solution to the broad violation of privacy and the abuse allowed by open, anonymous systems.
If people truly valued those things, then they would clamour for software that would make the web of trust easier to use.
How many encrypted emails do you receive annually?
There have been many post-mortems on the Web of Trust. Few mention "corps can't steal my data" as a failure, the failure is entirely in the impossibility of building the web of trust with enough people to gain critical mass. In fact one danger of the web of trust sometimes mentioned is the way it leaks your social graph just like a social network, only without the privacy protections provided by services like Facebook.
I've had my keys published and PGP set up on all my email accounts for years and literally the only encrypted emails I've ever gotten are from Facebook, and only because I went through a fairly arcane process to set that up. I even crypto-sign all of my email as a hint that I can do PGP. I doubt anybody has ever verified my signature on even one piece of email.
Not just a fancy new client - it was still the days when cloud storage wasn't ubiquitous and you saved stuff by sending yourself email, and Gmail had a completely monstrous storage offering compared to other free email providers. It was as much proto-Dropbox as it was email.
> The first is spam and abuse. The problem with a purely "open web" is that there's no solution to those, because of bad actors. The only solution we've found so far are centralized organizations which can run sophisticated machine learning and hire thousands of moderators.
False: there are plenty of old-school forums that still work just fine, with a handful of very hands-off moderators.
There's a trend toward censorship in the big players, because it's more profitable to get thin-skinned users on their platform, but this is not at all the only way to create a healthy community.
> The second is aggregation and discoverability. What good is it to publish if nobody finds you? Matching what people want to read, with the content they'll like (and not turning it into a firehose) is a really hard problem which, again, open standards basically do nothing for.
I generally find things through a few bloggers who have a high signal-noise ratio in what they post.
The idea that Facebook/Twitter/etc. do a better job than my bloggers is laughable. I left these platforms for a reason.
> And then of course the third is monetization. There's a lot of amazing content on, e.g. YouTube, that wouldn't have been created if it weren't for monetization, because it takes content creators time that they'd otherwise have to spend at a job.
On the contrary, I think the best content is almost always made because the person making it cares about what they are saying. "Hey guys make sure you like and subscribe! And by the way I've been playing Clash Of The Clans!" That's not content, that's advertising, and it's the poisonous result of monetization. Paul Harrell makes little-to-no money on his channel, Erowid survives on donations, Wikipedia authors are unpaid, Mother Jones and ProPublica are purely donation-based... and these are far better sources in most cases than any other sources of information on the subject. There is no case where ad-supported content does better than non-ad-supported content that I know of.
I want to suggest that all of this has more to do with you and your individual preferences than the way things work for everyone else.
* what's the scale of the forums that work fine?
* what if other people want to consume more content than "a few bloggers" produce and recommend?
* when you say "does better", what do you mean?
* is the best medicine made by pro bono pharmacists? are the best buildings made by off-duty carpenters? are the best movies made for free? with the best special effects? are the best novels free? the best graphic designs? would everyone agree with your assessment of what is 'best'?
"I left these platforms for a reason." Yes, and other people stayed for reasons too.
> * what's the scale of the forums that work fine?
Why, as a user, do you care? As a user, this question is irrelevant to my interests.
> * what if other people want to consume more content than "a few bloggers" produce and recommend?
They physically can't. There's simply more content in existence than a human being has the physical ability to consume. And you'll run into problems with depression, medical issues caused by sitting, and internet addiction long before you run into your physical limits to consuming content.
Quantity isn't the problem, quality is.
> * when you say "does better", what do you mean?
Finds higher-quality content that is more entertaining and/or informative tailored to my needs.
Yes, my needs are different from other users, but the whole value added that ad-based sites claim is that they spy on users to tailor content to their needs. My claim isn't that the content I find is better than anyone else's, it's that my way of finding content does a better job of finding content that fits the users' needs than ad-based sites do.
The problem is, at a fundamental level, ad-motivated spying doesn't even have the goal of meeting users needs--the goal is to meet advertisers' needs, and that only incidentally meets the needs of users sometimes.
> * is the best medicine made by pro bono pharmacists? are the best buildings made by off-duty carpenters? are the best movies made for free? with the best special effects? are the best novels free? the best graphic designs?
Are the best of any of these things supported by ads inside a walled garden platform? You do make a good point: none of these things are typically supported by the business model you're defending.
> would everyone agree with your assessment of what is 'best'?
If you don't agree that "like and subscribe and by the way I've been playing Clash of the Clans lately" is such garbage that it's questionable to call it "content", then we disagree at a fundamental level so there's not much of a discussion to be had. Suffice it to say, I think most people agree with me.
If most people agreed with you, who would be using all these platforms you've said are inferior and bad? If people didn't find the platforms met their needs better than the alternatives, why would they create and consume within them?
I'm not even defending ad-supported anything--I'm defending creators getting paid, and I find it hard to believe you think that's unnecessary. I like the Patreon model wherever possible. But I also think it's ludicrous to pretend that most people would prefer manual curation of limited content, but they just end up using the platforms they do based on... coincidence? conspiracy? they're not as well-informed as you?
Something about this reminds me of how television was seen before "prestige TV" became a thing.
> If most people agreed with you, who would be using all these platforms you've said are inferior and bad? If people didn't find the platforms met their needs better than the alternatives, why would they create and consume within them?
Because first to market advantage is huge in social platforms.
> The only solution we've found so far are centralized organizations which can run sophisticated machine learning and hire thousands of moderators.
I’m pretty sure HN, the site were literally posting on, does not have either of these, or at least not at that scale and yet it’s miles better than anything on fb or twitter so I’m not sure I buy this argument.
> I’m pretty sure HN, the site were literally posting on, does not have either of these,
As far as I know, HN does have two full-time employees whose sole job is moderation. I believe they also have some automated spam detection involved too.
> yet it’s miles better than anything on fb or twitter so I’m not sure I buy this argument.
This can be explained almost entirely by the fact that bad actors go where the greatest concentration of good actors is. Small communities are relatively safe with little effort required on their part, because they simply aren't a big enough target to be worth a bad actor's time.
For a long time, it was taken for granted that "Mac didn't have viruses". It wasn't because Mac OS was somehow more virus-proof than Windows. It was simply that Mac was like 2% of the marketshare and thus just not a worthwhile platform for hackers to target.
> I’m pretty sure HN, the site were literally posting on, does not have either of these, or at least not at that scale and yet it’s miles better than anything on fb or twitter so I’m not sure I buy this argument.
True. I like HN. By the look of it you like it too. When you say HN is miles better than fb or twitter I'd say that is a subjective judgement. I suspect that only a fraction of the people currently spending their time on fb or twitter would spend any time on HN. They're looking for something else. Something they don't get here.
On another note I get the sense that the vast majority of users of those walled gardens couldn't give a hoot about the concerns raised in the essay. Those platforms currently deliver something of value to their target audiences (however that may be defined). So as long as they do that and no alternatives arise to disrupt their stranglehold things will remain as they are.
The moderation problem scales in relation to the size of the userbase. For small communities like Hacker News a single moderator is no problem, but larger communities end up needing more and more moderators until you have so many that you need meta-moderators moderating your moderators. Once you've grown beyond a small community bad actors will constantly attack your system, often with complaints that they are being unfairly repressed and bad faith reports on good users.
Pretty soon there are news reports about how badly you discriminate against certain groups and world governments start poking their nose in your business. Plus you hit that endless stream of edge cases where no matter what solution you choose someone is going to compare it to the holocaust. Even clear simple policies get bogged down in edge cases extremely quickly. For example, you might have a "no nudity" rule on your site. Very easy, but now you're getting lawsuits from the breastfeeding mother's association on one end for taking down their content and the anti-pornography league on the other end because you allowed a picture of a bare ankle to remain up.
I believe this is addressed by the complaint about "growth at all costs". If your objective is to scale massively, then bad actors and trolls are indeed a problem. Facebook et al indeed favor growth at all costs, and therefore must deal with complex and large scale moderation schemes.
If you keep it relatively small it may be manageable, because "scale" cuts both ways ;)
The problem is reaching that critical mass where the social media service becomes useful before hitting the point where trolls make your life a living hell. A social media service isn't useful if none of your friends or family are on it, and it's extremely dubious if only one friend is on it.
I remember back to the days of umpteen different mutually incompatible Instant Message apps each with a different subset of your friends. They all ultimately failed as social media platforms and the only ones left are either integrated with larger services (SMS, FB Messenger, etc...) or are domain specific microservices (Slack).
Because for a long time things were interoperable. Many chat services ran off jabber/XMPP. You could easily use adium, pigeon, psi, gajim, and even the AOL IM and iirc yahoo messenger would all communicate interoperably.
In fact, you could even jabber chat with those using Google talk/hangouts until Google decided not to play. Before that iteration was usenet, IRC and ICQ.
Now you have WhatsApp, google chat or whatever it's called now, Snapchat, Facebook messenger, tik tok, zoom, kakao talk etc. But nobody wants you to leave their garden. And Because somebody popular makes a post using some platform, the masses rush there and doesn't leagu leave until they're told to by the new YouTube personality du jour.
So no, i tend to think in the process of millenialing chat services, they made them simple for them to understand, and kept them in the garden so they wouldn't get lost..
Opt in systems solve these in open social media platforms, but introduce their own vulnerabilities such as information/reality bubbles and difficulty in growing networks.
> The second is aggregation and discoverability
Advertising takes many forms, including word of mouth.
> monetization
This is true enough, but most content on the internet isn't monetized by the creator via the platform, and there are many examples of creators who monetize independent of the platform. It also depends on how low you want to go w/ what you call the platform. Some pay for a platform in order to monetize. I don't have stats on this, so maybe it's experiential bias.
The only correct solution to spam and abuse is to empower governments to charge people with offences when they commit them. This means national isolation, since the US can't extradite a fraudster in Russia and vice versa.
Aggregation and discoverability is easy. Talk to people. Don't think about publishing, think about being part of a community.
Monetisation can be solved by doing it the old-fashioned way. Standards aren't necessary; just make it possible to do small anonymous transactions. Don't ask them to log in and sign up before you'll take their money. Just take their money. Internet busking.
Not all spam is illegal, neither do I think it should be. National isolation would therefore not solve the problem entirely (if even for the most part) nor is it usually a desireable solution.
How do you find that community? Discoverability is a huge problem and word of mouth is by far not enought to replace search engines.
Your proposal to monetization seems to be an approach that hasn't succeeded in any way on the internet, despite people trying various ways.
My personal feeling on the tasks is that your solutions are not very helpful to tackle any of those problems.
>>The first is spam and abuse. The problem with a purely "open web" is that there's no solution to those, because of bad actors. The only solution we've found so far are centralized organizations which can run sophisticated machine learning and hire thousands of moderators.
This is completely false on 2 fronts. The idea that these major services have solved spam and abuse is simply provably wrong.
Further there are a number of ways stop spam and abuse on decentralized systems, in some ways these are better than on the centralized platforms.
>The second is aggregation and discoverability. What good is it to publish if nobody finds you? Matching what people want to read, with the content they'll like, is a really hard problem which, again, open standards basically do nothing for.
Here is a valid point, but there are some cool tech coming out that solves this, things like Lbry for example
>And then of course the third is monetization. There's a lot of amazing content on, e.g. YouTube, that wouldn't have been created if it weren't for monetization, because it takes content creators time that they'd otherwise have to spend at a job.
This is amusing, given the massive levels of demonetization on YouTube i think you are overstating the value of YT ad revenue in the modern area. Most YouTubers today seem to make the bulk of their money from outside sources like Sponsored Videos', Merch Sales, Direct Donations, etc.
Ultimately there are three good reasons for this, which the author doesn't address at all.
The first is spam and abuse. The problem with a purely "open web" is that there's no solution to those, because of bad actors. The only solution we've found so far are centralized organizations which can run sophisticated machine learning and hire thousands of moderators.
The second is aggregation and discoverability. What good is it to publish if nobody finds you? Matching what people want to read, with the content they'll like (and not turning it into a firehose) is a really hard problem which, again, open standards basically do nothing for.
And then of course the third is monetization. There's a lot of amazing content on, e.g. YouTube, that wouldn't have been created if it weren't for monetization, because it takes content creators time that they'd otherwise have to spend at a job.
Now if we can create open standards that work as well as centralized solutions for all three of these... then we're talking.