Hacker News new | past | comments | ask | show | jobs | submit login
Always Own Your Platform (2019) (alwaysownyourplatform.com)
322 points by ddtaylor on June 9, 2022 | hide | past | favorite | 181 comments




Off topic: this was posted and reposted four times in 2019 in a burst of three in March and two in June. The only one which hit was June 10, 2019. Well, as they say, second to last try is the charm.


From 2006 to 2014 I owned my own platform. Several actually. But I turned them down after it became a lot of work to maintain. At the time, it wasn't so obvious the web was dying, but in hindsight I probably helped kill it.

In the beginning the web was so new, and growing so fast, with new things, amazing sites, and more people getting online. Like all things in life, competition arises, and better sites started getting much more of the market share. People's expectations for what a website could offer rose tremendously, and would abandon a site if it wasn't up to snuff. Sites needed to have ever improving visuals, better content, better people, better interactivity, better everything.

And I couldn't keep up with it. Users went from being happy to try out something new to dismissive and bitter. More and more it felt like work to try to make them happy, to keep building more and better things. And that's exactly what happened. The Internet became work. It's why we all have to be paid to come to work and build the Internet. No one does it for free, because it's a thankless grueling job. The only websites that survived were the ones that made money, and could afford to use that money to hire people. Google, Facebook, Myspace, Stumbleupon, 9gag, and even Something Awful became money oriented rather than community oriented. They had to, or else.

The advice to "Always Own Your Own Platform" is a euphemistic way of saying make a whole company out of your site and underpay the only employee (you) for ever. The reason we don't own our own platform anymore is because it's soooo annoying to do so. It wasn't an accident.


While the damage is severe, I don't think it's permanent. Mastodon is far from a powerhouse, but is chugging along happily with a slowly growing user base. Matrix arrived just in time to help prevent Slack and Discord from becoming the defacto real-time communication systems for open source projects. Even SA is starting to heal now that it's out from under lowtax and his bullshit.


One thing I'd be curious to know is how large the current open platforms are compared to those of 20+ years ago. The internet as a whole is so much bigger that it's hard to maintain a sense of scale.

For example, how many people are using Matrix now, compared to IRC 20 years ago? IRC had 10M users in 2003 [1], Matrix has 28M 'global visible accounts', but who knows what that means.

Maybe it's ok for these platforms to remain 'small', growing slowly. Sure we'd all love to win the fight for a free internet tomorrow, but it's also important to maintain a realistic perspective.

Disclaimer: Usenet is still growing, but I presume that's just alt.binary - According to Wikipedia [1] it had 15M daily posts in 2010, and 171M daily posts in 2021.

[1] https://en.wikipedia.org/wiki/Internet_Relay_Chat [2] https://en.wikipedia.org/wiki/Matrix_(protocol) [1] https://en.wikipedia.org/wiki/Usenet


53m global visible accounts as of March, mate. https://twitter.com/matrixdotorg/status/1505009553884205063?...

It means exactly what it says. We don't know how many people in European governments, militaries, healthcare etc use it, but we know that they use it, though that information is not possible to discover publicly. We can see how many public accounts there are.

Usenet is predominantly piracy at this point.


I agree with you. But I do think Mastodon and Matrix / Element all have a marketing problem. It's seems to be "here's a clone of a service you use but you own your content".

I'm not saying that they don't do things better but they don't market it that way. To beat Twitter be better than Twitter. Make something memorable and unique, give features that beat the stockholder driven choices of the other guys.


Element (on Matrix) may be superficially a clone of slack or discord, but it already does some things significantly better:

* End-to-end encryption

* Spaces as freeform hierarchies of rooms

* Widgets, to embed arbitrary webapps into your chatrooms

* Open standard API

However, the point is more that Matrix can do way more than this given it’s a protocol: Element Call is already a better UX for conferencing than Hangouts or Zoom; thirdroom.io is a super promising virtual world system built on Matrix, etc.

We probably do need to market it better though!


Definitely, better marketing around differentiation.

I'm a huge dork for open standards, I frequently miss the days of Gaim/Pidgin and having Google Talk be Jabber/XMPP underneath, and being able to IM my friends wherever. Personally, I'm probably going to dork around with Matrix/Element because it sounds like my kind of thing.

But I can't figure why I'd switch to it for work.

I moved my company recently from Teams to Mattermost. The UX was the biggest selling point. Collaboration happens in Mattermost in ways it simply couldn't happen in Teams.

Putting aside the fact that we're not going to churn company comms every quarter for funzies, why would I switch us to Element? What can I do with Element that I can't do with Mattermost or Rocket.Chat or Slack or anything else?


What you can do with Element is:

* Have signal-style end-to-end encryption by default, if you care about keeping your data secure on the server.

* Pick from a huge range of alternative clients, including hooking up custom ones thanks to Element being built on the open Matrix standard. For instance, if you wanted to plonk an Intercom-style chat box on your website which funnelled into your support chatrooms, there are a a bunch of different matrix-as-chatbox clients you could use

* Bridge to other platforms as if you were using them natively (as per the sibling comment) - e.g. if you already have other chat systems flying around the place, you can easily unify them in Matrix

* Talk to people outside your company. If you care about collaborating with external people (other developers; customers; partners; suppliers; sibling companies etc) then Matrix's intrinsic decentralisation means you can just talk to them via the public Matrix network. For instance, I spend my life collaborating with folks on the mozilla, KDE, GNOME, Ansible etc Matrix servers without having to think (let alone manage a million different tabs and identities and crappy clients for their various communities).

So, that's the pitch.


Depending on if/how often you use your internal chat to communicate with external partners, Matrix's bridging could be a selling factor.


As much as I love FOSS, Element isn’t even close to Discord. Discord has one click streaming, easy to understand voice rooms, etc. You don’t have to sort out hosting, and the overall UI experience is sharper and feels more polished. Visually Element is nice, but it has some clunk to it.


Discord isn't even close to Element. Discord provides no privacy and a single point of failure. Matrix however will offer peer-to-peer lite servers running on-device that'll even work offline via BLE. Try using Discord during an invasion or natural disaster.

This idea that Discord is accessible globally is such an America-centric naïvete.


>Discord provides no privacy and a single point of failure.

Normal users fundamentally do not give a shit.

OP is contrasting Element as a client, and it is subpar in comparison to Discord's UI/UX. No protocol differences can paper over this.


Different strokes for different folks.

FWIW, you can bridge Discord to Matrix.

It is more valuable to me to have a single universal messaging app across all my devices than it is to corksniff over UX across a couple dozen apps.

Furthermore, I'm autistic, the UI of Discord is atrocious. Completely overwhelming.


None of these points even remotely refute my comment.


Ok but going back to the grandparent's point: what is the incentive to do that? It's really hard to be better than Twitter. I know Twitter kind of sucks and everyone here thinks they have too many employees, but the truth is that all of those employees are doing stuff, that despite it kind of sucking, it remains incredibly hard and extremely expensive to actually make something better. I just don't see how these platforms can compete with the primary incentive structure being altruism. It's way too much to ask.


> It's really hard to be better than Twitter.

It is, but right now people don't care as much about privacy as features and community. It's not what people who focus on openness or privacy want, but it's the reality. Where are the celebrities and brands? Where are the conversations happening? Where are my friends? This is the challenge.


Yes ... but that's never going to change, so what's the proposal here?


The key in my perspective is to not scale too far.

Twitter, Reddit, and many other mega-platforms began the process of sucking (exactly) when their user bases got too big. operational costs also soared as they over scaled to stay dominant, while they began to fear changes in functionality because of a circular cycle of fearing user backlash.

No one platform can service massive levels and disciplines of users well, yet everyone keeps getting greedy, and always loses sight of that fact. Costs of competition at monopoly level create a paralyzing fear of losing a grip on users... That fear makes those platforms resort to corrupt methods to retain and addict their users to their product, and makes execs and investors unable to see outside of the vapor cloud they create...

Also, never go IPO if you want to keep your product's soul... Cash out early like Tom from Myspace, and let someone else crash and burn the idea as a CEO... :P


How exactly do you think they started to suck? Also, what’s stopping platforms from servicing large numbers of users? Honest questions, I’m just curious.


> But I do think Mastodon and Matrix / Element all have a marketing problem. It's seems to be "here's a clone of a service you use but you own your content".

Well, Slack and Discord have the ability to throw around money to solve that problem. Open source projects mostly don't.

So, open source projects have to hit some nerve and then slowly grow into useful and then a dominant project. See: the trajectory of Blender.


My mastodon pitch is it’s «ad-free, community-run and features a chronological feed (no algorithms)».

I used to open with the fact that it's federated, but it's a much harder concept to sell initially. People do appreciate it later on though, when they find out they can follow and interact with a bunch of other social networks and communities.


My biggest problem with Mastodon is it’s psychotic user base.


This is entirely dependent on the instance/server you join. I can't guarantee that you'll find a pre-existing community that fits your interests, but you can start your own independent server an carve out a niche to your liking. That is not possible on any other commercial and centralized social media platforms.


Source?


I've heard of Mastadon, even tried it but couldn't get any value of the contents on it. The last time I looked at it, most of the content was about Mastadon and welcome messages. I was expecting to see some crazy stuff like flat earthers or antivaxxers but even that was not available.

I'm also not sure that I used it correctly, maybe I needed to find servers or something.

Anyway, my point it is that the lack of discoverability is not just a marketing issue but it seems like baked in these platforms. Something is not there yet and that's why it's not growing like crazy.


> I'm also not sure that I used it correctly, maybe I needed to find servers or something.

You needed to find people to follow. Unlike corporate social media content is not pushed to you by algorithms. You have to build your own social network manually. Not doing so you have an empty personal timeline, and a too busy global timeline that is full of irrelevant stuff to you.

> Anyway, my point it is that the lack of discoverability is not just a marketing issue but it seems like baked in these platforms. Something is not there yet and that's why it's not growing like crazy.

Strange as it may sound on HN, but I am happy it is that way. There is a different culture now on the fediverse, and it is one that likely won't survive that 'crazy growth'.


That doesn't make sense though unless there's a directory of recommendation engine or search engine for interesting people. Otherwise it means that I'm supposed to follow people that I already know, which is ridiculous because I don't want yet another channel to see the same stuff I already see.

On HN there's the front page where interesting stuff is curated algorithmically and smart people that I don't follow or know perviously make interesting comments on that stuff. Mastadon doesn't seem to offer anything like that.


I think the trick is to find an instance where the people mostly share your interests. When I first tried mastodon.social, I didn't get anywhere. But when I found a more niche instance, I was able to find a lot of good content on the local timeline, which gave me people to follow and gradually exposed me to the larger fediverse.


Where do you find this content? Can you link a few examples of good content(in your opinion, of course)?


It's funny because you could argue that slack and discord replaced IRC. I wonder if we will have cycles or if commercial hosted solutions will always win this arms race.


unless you happen to own freenode/libera.chat, you don't own your own platform on IRC either. same for mastodon, unless you happen to own the mastodon instance you're on, you don't own it.

And the old web had plenty of platforms too, tripod and geocities were platforms as well. You could take it a step further still and say that unless you have a server that's sitting in your own basement, you don't own your own platform, since any hosting provider could decide to pull down your site as well.

it gets real lonely if everyone does own their own platform, and it certainly wasn't the case in the old web either (though it was definitely more common to own more of your own platform). Still, I don't really think this is the defining feature of the old web. I think the level of commerce is much more important, something like ao3 feels old web, and that's because it's a user-centric platform, rather than business-centric


This is where federation matters. I can choose where I host a Matrix room or Mastodon account and other people using other service providers can interact. It's even possible to move a Mastodon account from one provider to another without losing contacts, though this looks to have some rough edges.

I don't think everyone should necessarily own their platform, but it's very valuable to have the option to migrate anything that's a critical part of your identity or business to a different provider.


Anybody can and could start an IRC server. I ran one. I ran a BBS too. Now run a Matrix server.


one could argue irc has improved by the platformists abandoning it.


Slack and Discord are the de-facto realtime communication systems for open source projects, though. I wish it weren't so.


> People's expectations for what a website could offer rose tremendously, and would abandon a site if it wasn't up to snuff.

Amazon.com is basic text and image elements on a white background. The full-feature sites often just look like blown up versions of the mobile sites.

I believe there are multiple issues with owning your platform but I'm not sure this is the defining factor, at least anymore.


For most regular non-tech people this is true. I think it’s best to control your own platform but you can still leverage Twitter, YouTube, Facebook, etc to grow your brand and make sure they’re onboarded to your own platform of sorts.


> to grow your brand

Well, yes, if that is what you want to do. If you just want to be you online, and not 'grow a brand' it is best having a platform where you are in more control of doing so than these corporate advertising platforms that are indeed tailored to building brands.


Videos on YouTube will get 800x more views than views on yourownvideowebsite.com (assuming you aren’t a major Hollywood studio). Donald Trump went from 90 million followers on Twitter to 3.2 million followers on his own Twitter clone, and he has such a loyal following that people literally committed treason for him. Network effects are impossible to ignore


That is such an incredible example. If Donald Trump doesn't have strong enough if a brand to overcome network effects, nobody does lmao


> Donald Trump went from 90 million followers on Twitter to 3.2 million followers on his own Twitter clone, and he has such a loyal following that people literally committed treason for him. Network effects are impossible to ignore

You're not wrong, but please note that those are Trump loyalists that followed him to a service that Trump didn't actually use until quite recently:

https://www.pcmag.com/news/trump-makes-first-post-on-truth-s...

It kind of seems that far from being ignored, network effects were the main reason Trump was using Twitter in the first place.


It's probably worth pointing out that many people who followed Trump on Twitter weren't fans of his, but people who wanted to keep up with what he was saying (perhaps as critics, journalists, etc) and already had a Twitter account for other purposes. The folks who folowed him to his own Twitter clone were (by and large) actual fans who suppported him.

If you dislike Trump but want to monitor his actions, following him on Twitter might make sense. Joining a platform dominated by Trump supporters probably wouldn't.


Trump was a TV celebrity well before Twitter... That's not a good example.

These days YouTube does nothing to promote posters that aren't rich and/or famous already.

Having my own site and embedding my YT content is the only thing that helps me to be visible amongst a sea of celebrities who are already popular and promoted for free on the front pages of social sites like Twitter and YouTube.


Does no-code make this better or worse?


> Distribute [your content] via methods you control.

I'm being nitpicky but it feels off to me to call people out on owning their own platform but then use Cloudflare's CDN instead of running their own. And if they did run their own CDN, they would probably have to use AWS or another worldwide cloud provider and not own the hardware. And if I were Richard Stallman I would probably come down on you for using proprietary chipsets in your hardware instead of totally open source hardware. And you can just keep going deeper and deeper with this train of thought. So what level of ownership is acceptable to this website's author? I feel like the right balance of "ownership" is super subjective.

I love a lot of traits of the "old-school" web and am a huge believer in self-hosting as much as possible. But imo yelling at people while on your high horse doesn't encourage anyone.


Unless you are using Cloudflare's specialized security services, CDN is pretty interchangeable and easy to switch between. You can ever do dynamic DNS things to use multiple CDNs at the same time and shift different traffic segments onto different providers if you are into that sort of things, or tier one CDN behind another to act as a mid tier.

I don't really understand why the decentralized internet folks harp on CDNs because they will be very required if they ever get their way since by definition, if everyone is running small to medium, highly distributed sites, then you aren't likely to have a point of presence near most users and that will be poor user experience.

My guess is that most people have only heard of Akamai or Cloudflare so they automatically associate CDN with huge companies, but the reality is that the CDN space is highly competitive and decentralized. Just off the top of my head there's at least a dozen providers I would consider tier 1, several dozen I would classify as medium sized or targeted at niche audiences, and probably hundreds of small providers, not to mention just about every large telco operates their own CDN now.

https://en.wikipedia.org/wiki/Content_delivery_network#Notab...


In other words, CDN isn't a platform, it's a service. Consume services, but own your platform. Cloudflare's not going to shut your web presence down because you're "competing" with them or just plain doing something they don't like. YouTube or Apple absolutely will.


> Cloudflare's not going to shut your web presence down because you're "competing" with them or just plain doing something they don't like.

While I agree with this distinction, I'm not entirely sure that this is wholly accurate.

The situations in which Cloudflare might have to intervene with legal content are probably pretty limited (e.g. content that would be legal, yet very controversial, or expose something about the company or something like that), but it's still possible.

As a thought experiment, consider a situation in which a person might be using a Russian CDN and would be hosting content about Ukrainian refugees (no intentions of getting political here, it's just a good example) - now I'm pretty sure that Russia might just make it illegal, but even without it being explicitly legal, a few strings could be pulled and the service to that site shut down. What about a Chinese CDN that expresses views that are different than the ruling government would approve of.

Maybe there should be an asterisk there - depending on the nature of your content, have both the services you use and the platform that you own be somewhere, where it won't cause problems.


But you can always change CDN provider in a few minutes, an hour or a day.

You cannot move from Twitter to SomeOtherUnknownPlatform.com or from the App Store to UnknownAppStore.com if those platforms decide they don't want you there. Well you can, but no one will ever see you again. You cannot even move to a well known platform in case your followers are not there (or don't use that OS). At least, it will take years to move them over.


> But you can always change CDN provider in a few minutes, an hour or a day.

It depends entirely on how many of their offerings you choose to use: https://www.cloudflare.com/cloudflare-product-portfolio/

That's like saying that your application that is running in AWS can easily be moved elsewhere, but that's going to be true only sometimes: a lot of the time you'll use their managed services (which may or may not be compatible with others), their SDKs for setting up the infrastructure (which may or may not be trivial to migrate) and might end up with a lot of knowledge that's largely only valid within their ecosystem.

But that's mostly because a lot of the larger service providers out there try to expand their market reach and provide lots of different services: like I could go to my domain registrar and get either managed hosting or VPSes.


> I don't really understand why the decentralized internet folks harp on CDNs

Personally, although i'm not dogmatic about it, i'm rather opposed to CDNs for two reasons:

1) For performance reasons: unless you're distributing video, removing bloat from your webpage makes it faster to load from anywhere (including bad connection on the other side of the world) than to add a CDN on top ; i've already got a TCP route to your server so if you stay under 1MB that's much faster than to resolve a new domain and open a new route

2) For security reasons: CDNs are often used to distribute scripts, which are the #1 entrypoint for infecting users with malware. Also, as you pointed out, "anti-DDOS" reverse-proxies like CloudFlare are yet another case and should not be used in 99.99% of cases, because TLS will be terminated on their side and they can read all your users passwords!

I believe the web would be a much better place if content was served from a single origin (i believe it was not done originally because most of us didn't have enough bandwidth to consider the option). For shared content across origins, there's much better solutions than CDNs in the form of Content-Addressed Storage (eg. Bittorrent/IPFS/DAT). Whether you download it from a single trusted source (eg. your ISP), or in a p2p fashion is your choice... but a location-addressed protocol like HTTP is not well-suited for distributing static content across the world.

If only browser vendors (read: Google) were not too busy destroying our URL bars or inventing new tracking systems for advertisements that would be a fixed problem by now. But no, who needs reliable content distribution for smaller websites when you can have WebGL and a Battery API to better track users and make it harder for competitors to implement browsers?! I guess it also doesn't help that Google who develops for ~90% market share is the biggest 3rd party origin today with Google Fonts, Google Analytics, Firebase, etc.


Cloudflare serve something like 12% of the entire web don't they? That's more than enough to be worrying even if they theoretically have some competition.


> I don't really understand why the decentralized internet folks harp on CDNs

For sure, I was just using it as an example on how subjective "owning your platform" can be.

I actually have zero problems with Cloudflare, CDNs, etc. If I did have a problem with it and you disagreed, wouldn't that still illustrate my point that "ownership" is subjective and opinionated?


> how subjective "owning your platform" can be

I think your point is a good one, but I don't entirely agree. I think the comment you were replying to suggests an objective criterion:

> CDN is pretty interchangeable and easy to switch between

So I think there is a real basis by which you can objectively decide how much of your platform you need to own. Outsource commodity components (Linux servers, CDN and static hosting, managed database as a service maybe?) but don't get locked into proprietary components (hypothetical e.g.: writing all your blog posts in the Medium editor, where you can't access the source if they remove the "export as markdown" button).

Of course there are nuances, and we can argue over exactly how commodified a particular capability is, but I don't think that makes it subjective. Maybe each person prefers different tradeoffs on a subjective basis, but can probably regard the facts objectively (e.g. count the number of managed database providers, decide it's acceptably commodified, but still run your own database server anyway because it's interesting).


That's a really good point. I could definitely see interchangeable tooling being a good criteria for owning your platform. I feel like I've never described it that way before but that is totally how I determine what is going to be a part of my tech stack. I.E. "Is it possible to self host this tooling/workflow?", "Are there competitors?" are some of the questions I ask myself before adopting a tool or workflow. Being able to self host my critical applications is a big deal even if I don't choose to self host all of them.

Again, very good point and I feel like I can totally get on board with this definition of ownership.


> I don't really understand why the decentralized internet folks harp on CDNs

+1,000,000 here. CDNs are exactly a mechanism for distributing content around the world in order to serve users locally. It's exactly decentralization! And, as you point out, the industry harbors healthy competition.


Distributed physically but not from an ownership standpoint.


Just like security, you settle for the risk you can afford and are comfortable with.

The cloud muddies your grasp of the risks though. A lot of these cloud offerings (like k8s) are incredibly complex, and therefore have a lot of attack vectors, and a lot of points of failure. We depend on layers built on layers, built on layers. You can't really eliminate all of these dependencies, but you can eliminate / reduce some of the highest risk ones (topmost layers).


> Just like security, you settle for the risk you can afford and are comfortable with

Right, I totally agree. What I'm not a fan of is this site actively encouraging people to call out others for not owning their own content. If we are gonna compare to security, it would be like suggesting everyone should adhere to the same threat model -- which just isn't practical.

> The cloud muddies your grasp of the risks though.

Are you saying "you" generally or me specifically? Because I was just using a CDN built on AWS as an example of not being able to own all parts of your stack. It's mostly impossible for an individual and most companies to build a CDN on their own hardware as most entities don't have datacenters around the world.


> Are you saying "you" generally or me specifically?

Generally, as in all of us.


Ah, yeah. I see what you mean.

I run a K8s cluster at home (I also work with AWS EKS) and in a lot of ways I see the upsides of what you gain with K8s. I also see why a ton of people won't ever need the complexity of K8s and I can see how the complexity/downsides would outweigh the benefits in the majority of situations.


Having CF handle your CDN proxying is still much better than just using some all in one platform. At least you own your data and can migrate it if need be.


I absolutely agree -- I was just using that as an example of how subjective "owning your own platform" is.

EDIT: I originally said "owning your own content" but "owning your own platform" is more in line with what I am talking about


I don't think the CDN qualifies as a "platform" because, I think almost by definition, a platform is something that's hard to move your application off of. It's relatively painless to switch your app's CDN provider compared to, say, moving it between AWS/DigitalOcean/Azure.


I totally see what you mean and I agree with you.

As I said in another comment I have no problems with Cloudflare or CDNs -- and like you said in another comment CDNs can actually make it cheaper and safer to self host your own application.

> I don't think the CDN qualifies as a "platform"

I think you are 100% correct and I agree with your point. But I think even your disagreement with my comment highlights the original point I was making: "ownership" of "your platform" and what those terms mean to you is super subjective based on your values, skillset, etc. Wouldn't you agree on that point? If so, doesn't it come off as condescending (like you felt about my original comment) to tell people what they should or should not do with their digital content creation?

My whole point was that it bothered me that this site encourages people to call out others for creating content on a platform they have disagreement with -- who creates that standard of what is suitably "owned" or not?


Not just the data, but the infrastructure as well. A CDN lets me host on my own equipment rather than renting cloud compute - and I can change CDN providers anytime I like.


If you don't mind me asking, how do you have your own infrastructure configured?

I just recently set up my own blog on a DO droplet, which is fairly vendor-portable if I need to change cloud providers. I had always assumed that serving content over a non-commercial connection would have some issues, although this was more of a vague doubt in the back of my mind.

Do you worry about having greater exposure to (possibly malicious) outside traffic on your home network? Does your private ISP plan provide enough resources for hosting content?


I talked a bit about this in another comment (https://news.ycombinator.com/item?id=31686615#31687589), but let me provide a quick runthrough.

I have 443 port forwarded on my Ubi Edgerouter, which sends those to a container on my server with SSLH. SSLH forwards OpenVPN packets to my pfSense VM and SSL packets to NGINX which directs the web traffic to other containers hosting the actual web services. I have a subdomain with DDNS that is routed to my public dynamic IP.

The above is for services that don't go through the CDN like Matrix. For my public web site which uses Cloudflare, I use an Argo tunnel directly from the webserver container to the CDN itself. Due to the CDN and the static site it barely affects my connection. Honestly what slows down my connection is if several people use full-tunnel VPN at once due to my limited upload speed.

I run enterprise equipment and know how to use it, and my provider allows me to connect my own equipment to their network. They also permit home hosting within reason. The line is also very stable, so uptime is good and things are reliable when everything is properly set up. I also have a backup server to run my VMs if the main one fails, have a spare modem, etc. The idea is that if anything happens on my end it can be fixed in a couple hours.

Security wise the VPN is locked down and so is Matrix (access for family members only, no public account creation permitted). Bots do occassionally connect to my public IP but its nothing to worry about. For the website CF takes care of traffic spikes. Since everything is in seperate containers and VMs in a worst case scenario the damage done will be limited.


Fun rundown to read, thanks for sharing!


The day cloudflare starts filtering individual pieces of content, I would get out fast. FB, twitter, et. al. show me what they think will keep me looking at them for as long as possible. It is only coincidence that it is sometimes content I actually want. If I can get all my friends to just post by RSS I don’t need or want anything else those platforms do.


How are you defining "filtering individual pieces of content"? Cloudflare kicked off The Daily Stormer and acknowledged that they were doing so as their own choice for political reasons.


That is the very issue I consider. The Daily Stormer block was done to an entire business which is a different thing than prioritizing some posts over others from the same source. It is worth discussing cloudflare’s decision to block the Stormer. But that is different than Facebook’s optimization AI choosing to not show me posts from people I choose to follow. Both are potentially concerning and complex, but I consider them to be separate topics.


MITMing and filtering or blocking fragments of a request.


Sure, I agree. I actually have no problems with Cloudflare, CDNs, etc. I was just using the website's use of Cloudflare CDN as an example that "owning" your "platform" is subjective to everyone's personal experience and values. And I have a problem that the site encourages people to call out others who don't line up with their view of what owning the platform means to them.


But you don't actually need a CDN unless you're optimizing / min-maxing for speed; someone's leftover raspberry pi running in a broom closet halfway across the world can serve a website just as well as an AWS cluster can.

Of course, if you get bigger you need to start making some concessions, but for a lot of people that's wishful thinking.

I remember when I was 16, there were these guys that made really shitty voiceovers and edits for videos and put them online. They just had a computer running at home. It'd get hammered whenever they uploaded a new video, but it worked. They didn't rely on anyone else - hell, this was before youtube was a thing.


I think hosting a website in a low powered computer like Raspberry Pi can be an alternative to really own our platform. Granted, it can't handle big traffic, but it is cheap enough to built and maintain yourself.


A light weight site run on a traditional web host is fine for most use cases early on in my opinion. The problem is so many people think of costly cloud hosting as the first move and burn up all their capital even before they reach 10k users.

That is why the dot com era was more fruitful. People started in their basement, and service outages were not a big deal... Heck, service outages were often the thing that spurred more interest in products and also spurred users to invest so that products could scale.


Totally agreed. And what you said is in-line with my original thought that "owning" your "platform" is super subjective and opinionated. Some people think that infrastructure is part of your platform and some don't!


> I'm being nitpicky but it feels off to me to call people out on owning their own platform but then use Cloudflare's CDN instead of running their own.

Infrastructure isn't a platform, and vice versa. If Cloudflare started to make editorial decisions about which web traffic to "promote" to users, there would be hell to pay.


> Infrastructure isn't a platform, and vice versa.

I gave this some thought and after a bit I disagree with this.

Wasn't Parler an example of infrastructure being used to editorialize _against_ specific opinions and ideas?[1] As I understand it, they were kicked off of AWS and refused by multiple infrastructure providers for hosting content that was perceived as offensive and potentially dangerous. Is that not an example of infrastructure being part of the platform?

*I want to make clear I do not use Parler at all and am in no way stating I am either for or against it -- I just think it's an interesting example.

I would love to hear your perspective, though, on that.

To clarify, though, my original point is that "owning" your "platform" is subject to your values and experience. Do you agree with that? I actually have zero problems with anyone using Cloudflare, CDNs, etc, I just think it feels self righteous to encourage people to call out others for not owning their own content (which the original website does).

[1] https://time.com/5929888/amazon-parler-aws/

EDIT: forgot to add the link :)


> Wasn't Parler an example of infrastructure being used to editorialize _against_ specific opinions and ideas?[1] As I understand it, they were kicked off of AWS and refused by multiple infrastructure providers for hosting content that was perceived as offensive and potentially dangerous. Is that not an example of infrastructure being part of the platform?

Sure. Businesses have always tried to have it both ways, and really like being necessary infrastructure while still having the option to refuse service. And, to be fair, AWS (and Cloudflare) aren't quite at the point where they are going to be publicly regulated utilities that no longer have the option of refusing to provide service (like an electricity or phone service provider).

But they are getting closer to that every day, and eventually we'll stop talking about IaaS and start talking about IaaU (once most other providers give up the Tesla-vs-Edison fight and concede that AWS has defined the APIs that have become the defacto standards they need to copy).

But even in a hosting-as-regulated-utility (or common-carrier, which is the relevant term from the transportation and communication industries) world, there are still exceptions and there are lots of things that in the name of public safety etc. you aren't allowed to do when using these services, and limits that are imposed. I don't really expect utility services like compute, storage, edge caching, content replication and distribution, etc. to be exceptions to that.


Depends what your platform is but most of the examples given are about content.

So I see it as own your content and don't licence it out or be unprepared to migrate hosting platforms as needed.


Good point about content and licensing. It is discouraging to understand that in most cases when people post content it isn't theirs anymore despite them taking the time to create it.


Really hard for non-techies; even a hosted WordPress isn't easy to use. I've been asked a few times 'I want to write online, what do I do?' and it's hard to find easy to use answers that don't involve vendor lock-in. Or me having to provide support forever.

SSGs with plain text files are brilliant once you are set up, but it's a big hurdle getting there.

Best I can do is try to find platforms with sane exit strategies.


> SSGs with plain text files are brilliant once you are set up, but it's a big hurdle getting there.

Even in that case, not nearly as practical, unless you mostly do long-form and only post from your computer. Even with SSG set up, having to be on the computer to create a draft, or make a quick correction, is really a bother.

Also, when you own everything, you end up spending more time tinkering with it, which is not bad per se if you are having fun, but terrible if your goal is to actually post. Those 2 things - harder to post and easier to get sidetracked - explains why so many SSG'd blogs are made up of a few posts explaining why and how it was set up, and maybe a couple actual posts.


> Even with SSG set up, having to be on the computer to create a draft, or make a quick correction, is really a bother.

I think you might be able to solve this with Syncthing and a watch process on your syncthing server. Commit and render on each change, and optionally sync if the rendered site is different from the previous version.

...granted, that's a really tech-nerd setup.


That's what stuff like Netlify does. You just set up a git hook to build the site when you push, and it all happens automatically.


At that point you may as well use a real cms.


I dunno.

I have basically this working with my todo.txt file between my laptop and phone, and it's incredibly low-maintenance. Most CMSes mean running a complex app, a full DB engine, and having to stay on top of security updates for all that. Heaven help you if you run a CMS with a bunch of third-party extensions.

For my static sites I have SSH and nginx. Nothing else needed.

Static isn't always the answer, but where it does fit it's hard to beat for simplicity and efficiency.


How is running syncthing better than git or the site generator?


Sorry, I was really vague there.

If you run syncthing on your mobile device, you could make edits to a local copy of your plain-text content, and have those edits auto-committed/rendered/synced when they hit a computer housing the actual git repo.

I personally despise typing on my smartphone because I'm so much slower on it and I'm much more error-prone this way, but I use something like this for managing my personal todo.txt file and it works pretty well.


May be worth considering: Lightweight ActivityPub writing platform, straightforward to self-host but there are also managed hosts: https://writefreely.org/


Installing and managing anything computer related is so hard these days. Take programming for example.

I started many years ago on C. I could literally copy a compiler executable out of an archive and it's ready to go.

PHP was an apt install command. Want to use it on the web or with a database? A couple of more install commands. Guides were simple to follow, and they worked.

I'm trying to install Python for my daughter, and f me it's like pulling teeth.

The official installer is apparently the wrong way, you should use xyz package manager. The guides that people recommend just don't work, weird errors show up. Searching online leads to others with similar problems, but the always recommended guides are never updated. It's frustrating spending an hour fixing one error message, then another pops up. All I want is to install the damn thing.

I've worked as a software engineer and as a documentation writer, and I struggle to deal with this stuff. Nothing just works, even a simple website now is a massiveness stack of stuff that breaks for weird reasons and is really hard to diagnose if you don't deal with it all the time.

Right now I'm trying to set up Unity3d to build to iPhone... It's like researching an arcane magical spell. There seems to be no clear guide to how things are supposed to be set up. I'm now watching bloody YouTube videos on it, and despite following the exact same steps using the same software versions suddenly they have settings I don't, where did they come from who knows. It's never explained and you end up trawling through forums to find the nugget that explains exactly what thing you have to do.


Python isn't a special case either, same issue with node, php/composer, R, and many others. I pretty much do this for a living now, it seems like most of my skillset is figuring out how things should run without making too much of a mess.


> Really hard for non-techies

Yes; but it’s not like even that many techies host their own content.

I love how Jeremy Keith copies everything he tweets to his own site [0]. Wanted to check how he set it all up; but his site isn’t open-sourced.

[0] - https://adactio.com


Looks like it's the other way around. He posts on his site and it syndicates to the other platforms.

https://youtu.be/X3SrZuH00GQ?t=852


Even better! The true POSSE spirit!


Drupal has been solid. It needs a good developer who knows how to architect it properly so that it endures.

Wordpress has been overrun by paid modules, which contradicts most of the benefits of it being open source. It also has had far too many security breakdowns. Paid modules are also now developed on Wordpress to break on updates, the eco-system at Wordpress needs serious attention.


I mean "just learn HTML and upload it" will always work, you can provide them with a simple template as well if needs be (e.g. http://bettermotherfuckingwebsite.com/ ), but they would also need to learn FTP or its equivalent built into their file browser.

But I'd argue those are easier than wordpress itself. Plus, WP is quite invasive in that it will do its own updates, and most plugins try to push some paid services as well.

I'm sure there's more minimalist WP alternatives though.

What about a blogging platform that runs as a self-contained single page application (minimal) that gets its contents from text files? I'm sure it exists somewhere.


I think there are SSG front-ends that are quite good.

I find that writing in HTML is fiddly for regular blogging but I do agree it's a lower barrier to entry than it seems for simply writing a single page.

When I show people my text-file SSG system, they love writing in markdown, they love that the site is 'portable', but they don't tend to want to get involved with the templating etc.


So, the eternal "You can host your stuff in your basement" vs. "No, not everyone can host their stuff, it's hard".

Here's (hopefully) some food for thought on how a decentralized web could work in an ideal world in the future.

The problem is, even if you sell amazing shiny turn-key server boxes to every household, where they would be able to seamlessly host everything they currently keep elsewhere (iCloud, Facebook, etc.); even if messaging and other communication becomes absolute P2P in the whole world and therefore there's no need for centralized messaging services —

— there would still be the problem of centralized search. The beauty of Internet is not in its content as much as in the ability to discover it. Search implies there has to be a central place where you start it. This is why everyone - techies or not - tend to push their creations to places that provide exposure and discoverability: Medium, SoundCloud, YouTube, Facebook, etc. Then all that is additionally indexed by a meta-engine that is Google today. Centralization upon centralization.

I think there might be a solution to this which would be a mix of P2P and locally centralized services. Imagine a gigantic balanced tree of indexing engines that belong, say, to communities. Whenever someone performs a search anywhere in the world, the query is propagated through the tree and is processed in a parallel manner by a great many nodes at once. I'm not sure about the exact algorithm right now, but something suggests Google probably works this way anyway. Except in this ideal world, the search engine doesn't belong to a single authority, but is rather split into myriads of local services maintained by (and paid for) by the communities or some small local companies.

If you think it would be wasteful and traffic-heavy, think of the resources and bandwidth that might be freed if Google, Facebook and other giants were replaced by this highly decentralized system, which, again, would work pretty much like the incumbents do, except data would belong to and be hosted by individuals who created it, and search would be one giant brain with potentially millions of cells that perform queries in parallel.

How far fetched is this? Very :) But at least I hope the idea is thought provoking.


> Facebook, Google, Twitter, Medium, and YouTube entice us to give them our creative work. It's time to take it back ... Stop giving away your work to people who don't care about it. Host it yourself.

So, youtube, right? Where would you host your videos if you had a mind to take them back?


Youtube has not much value as a video hosting. You can host your video anywhere else easily.

Youtube's value is viewers that will accidentally discover your video. Nobody will watch your videos if you publish them at your own website.


Sorry but have you tried hosting your own video?

There's only really one set of codecs that works across all devices. Video Player libraries that work really well with features like captions/subtitles, speed controls, automatic resolution shifting are hard to find, much less if you are not technical.

If you want to cater to an audience outside the western world you'll want a video processing pipeline to downscale your videos to lower resolutions.

Ideally you'd want to chunk up the video into something like HLS playlists so navigating around the video wouldn't require someone to wait until the download caught up to the middle of the file... the list goes on.

Video is still hard. The big players make it seem easy, but it's not, in my experience, that simple.


Peertube for one.

Even if you want to use Youtube, Floatplane or others to expose your content you can still pull users back to your site and own the creator - user relationship.


YouTube can be used as a host yet not as a platform. Make all your videos unlisted and just use video player on your website.


But that’s not owning your platform. What if youtube decides that your video is in violation of some community guidelines and deletes it?


You can fix that by having backups and hosting elsewhere. The difference is using a service's infrastructure vs. using a platform's reach.


If you own the page where they are embedded, YOU own the content. Your YouTube goes away, you can place the video elsewhere (or even back on another YouTube channel) and update your embedding to point to the new location. You still own the page and any comments and/or monetization you might have established.


sign up again?


Agreed if you make them all unlisted AND:

1) set the iframe width="514" height="289" or smaller that will remove the "Watch on YouTube" link (viewers can still view in fullscreen)

2) add ?modestbranding=1 to the end of the src url, this will also remove the "YouTube" link

Then add your own comments system to your embedded page and monetize how you see fit.


> Google

Yeah, stop letting Google dictate your search results! Create your own search engine and own your platform!


Ain't sure if you know this, but in case you don't, Google is a lot more than just a search engine nowadays.


I would simply spend millions of dollars serving my own viral videos.


Well. I'd wager you can server your 1 viral video for max a few grand a month.


Paid for by all the Google ads you run or money from Patreon haha


paid for with what, patreon?


Sounds cool in theory but almost impossible and impractical. For example, you can have a newsletter but google could decide your domain belongs in Spam forever. There goes that “owned platform”


From your comment you seem to think that Google somehow owns your arse. Perhaps you use them for your email and other functions and fail to see over the top of their walled garden?

There is no doubt that Google largely own Search and several other functions but they really don't own the internet as a whole. Often we see sad articles on HN about how the monster in the room twats someone into oblivion thanks to a stupid "AI" making a judgement call that a human would be embarrassed about. To avoid or perhaps slightly mitigate that fate is well documented hereabouts and elsewhere. For what it is worth (in my experience) Google is not inimical on an arbitrary basis. They really don't decide anything - it's all algorithms, pixie dust and bollocks. Avoid that and you are golden.

As it turns out, I've been able to run many "local" email systems and even my own little IT company quite happily for over two decades, without giving much thought to Google. We do have a newsletter email that we send out too. Some of my customers fly quite close to the edge of the spam/ham boundary.

I operate in the UK. Bizarrely, I can quite happily run SMTP from a domestic FTTC connection here, with a little care.


What GP meant is that Google owns the email of all the people you would send mail to, and can decide that all messages from you go unseen by your recipients.


I don't think there is any fix for that except for people to deliberatively and incrementally take that power from their hands. not betting on the outcome. I guess the old real hope is that 15 years from now when google is finally no longer relevant, that smaller business and individuals diffuse that control more broadly.

or we get google 2.0


The bin forever? I don't know anyone who gets their news and blog posts straight from Google search. Google can't stop you from putting up an RSS feed, sharing the link to friends, or posting it to content aggregators like HN or Reddit, which I think are far more important for building a real audience than Google search.


I interpreted the reference as being to Google's email product, not search.

As a techie (but not email expert), I have tried to run a couple small mailing lists for community projects over the years, and wound up giving up because of gmail/hotmail/etc's aggressive spam filters catching too many of our legit emails.


Well, technically it's still an owned platform, just in this context it's a platform that got owned by google.


Facebook (Instagram) and Twitter convince you that getting followers is growth. All you're doing is donating content to them and giving them advertiser profiles so that they can sell your own audience back to you. Algorithmic timelines mean that "your" followers/subscribers see what the owners want them to see, and not the stuff from you that they explicitly subscribed to.

This website is good advice. Stop sharecropping for billionaires.


As others have pointed out, there's currently no solution that allows common people to self host.

There's a minimum level of linux/sysadmin/devops expertise requried to setup and maintain a self hosted website.

I've written a somewhat long essay about this problem:

https://hasen.substack.com/simple-self-hosting

A solution is possible but no one is working on it yet as far as I can tell.

Instead, everyone wants to make money by taking offline/local applications and turning them into online/cloud applications (even if no specific capability is gained by being on the cloud) just because it's easier to make more money that way.


> here's currently no solution that allows common people to self host.

And there sure as hell should be!

These days, a lot of people seem to be capable of installing a Linux OS on their machines. Their questions and problems have gotten a lot more sophisticated in recent years. A simple site should be hella easier.

I don't understand why such a basic and simple thing hasn't shown-up yet. I constantly run across people that aren't worried about money, just getting text and photos on a simple page and getting them online. The money thing could be a plug-in or something. A Webring plug-in would probably interest more people right away.


They haven't shown up because non-tech people don't demand it. They cave in and use one of the managed solutions like wordpress dot com, squarespace etc. A group of devs need to create a solution even without demand & set about popularising it, which is very hard, when most people are satisfied with cloud solutions and don't see the danger that their data could be lost or vendor captured or censored arbitrarily. The second issue is network effects, where these cloud services provide exposure to your content and provide you exposure to related content, which as a self hoster, you must do for yourself.

Combined, both problems seem to be remarkably hard when viewed at closeup detail, especially when pitched to "point and click" users with minimal willingness to hack at the command line.


The command line is half the problem. The solution should not require command line chops. Should just be a GUI.


One of the simplest combination of solutions that I know in this space is to use Publii CMS to compose your blog and push it to GitHub Pages, which automagically makes your SSL enabled site live. Of course something as trivial as adding your custom domain leads you down a rabbit hole of nameservers, CDN etc. There is probably a simpler path here I'm missing.

But the bigger problem is the non-tech world simply doesn't know about any of this. When they want a site, they straightaway consider Blogger vs Wordpress vs SquareSpace vs Wix vs Shopify. Or just a FB/Instagram Page. In fact the latter often comes before the dedicated sites since you depend on them for networking.

It's a hard problem: how to be fully decentralised, yet connected/indexed/searchable, secure, yet flexible... and all this with sensible defaults and point-and-click to onboard lay people without scaring them away.


Publii is a good one but it lacks proper marketing. It's just not promoted enough. It also does not have a mobile app, and don't support pages. It could be a good solution for non-tech people that want to own their space and can't spend too much money/time, but it would need to change substantially for that to happen.


I never heard of it, but having taken a look it just seems like a GUI variant of Hugo. Which is nice and all, but nowhere near what I'm talking about.

Generating static sites is popular specifically because the root problem is not solved: hosting and maintaining a website is often a nightmare.


Have you seen cloudron.io? They are spot on solving the problem you describe


It does not seem to even be the same ball park.


In a perfect world, everyone would have their own servers in their homes or colos. Fiber would be cheaply avaliable everywhere, and everyone would have the know-how to manage their locally hosted services.

In truth, it's quite a lot of work to manage and maintain. Now there are say Docker Wordpress images that can be set up in ten minutes on AWS. However part of self-hosting is that there's no technical support - you're on your own if something breaks, there are updates, etc. Whereas a managed service abstracts all that away, and thus is appealing to a mainstream audience. While we at HN may like the technical side of things, many creators are uninterested in spending hours debugging a performance issue and want to write/draw/etc.

Personally I have servers in my basement and run a Matrix instance, OpenVPN, and a webserver (static site fronted by Cloudflare) as my publicly exposed services. I also have additional services like Landscape and network storage on LAN. When you run all those services, you don't just follow a guide, set it up, and leave it. While that may work for a couple months, something will break eventually and you'll have to scramble to fix it without understanding why. To do things properly and get good uptime you need to understand how things work, why they're configured that way, and have good backup and recovery practices. Security is also another factor, especially if you are storing private or confidential information.

Also for creators, there are services like Youtube and Twitch which are pretty much impossible to replicate without massive cost and infrastructure. If you're just doing text and images, then you could use a VPS or a CDN fronted homeserver and operate your own website. I'm not a vtuber, but if I was it simply wouldn't be possible for me to run my own service even though I'm interested in this stuff.

The only thing I really can recommend is to keep local copies of all your work on disk, properly backed up. Then at least if you lose access to the service, you can bootstrap again somewhere else.


I think users appreciate having a feed that combines everyone’s posts and algorithmically recommends the best ones, like youtube or twitter do. That can be solved with RSS now, but RSS doesn’t get you likes and replies etc.

And they appreciate having usernames so they can say “check me out at @username”. You can solve that by making everyone buy a domain name, but we have zero infrastructure right now for letting people use their domain name as their username.

The cryptocurrency community does seem at least halfway interested in solving these problems fwiw, but they do still have some major scaling problems to overcome.


Why would this be perfect? It sounds like suggesting that in a perfect would everyone would grow their own food, generate their own power, process their own sewerage, etc to make sure they are not dependant on anyone.

But ultimately we can see that depending on others and centralising services gives us more. Yes sometimes those services fail and it’s out of your control, but you still lived a better life having relied on them. I get the feeling that HN users have a hard time feeling that they are not in total control of every aspect of their lives.


Hosting your own LAMP server is one thing, but I have been touching computers for 30 years and I still would not trust myself to to roll my own ecommerce anything. It sucks but I'll always outsource with shopify/paypal/patreon/whatever.


You can’t even run your own realistically. Taking credit cards requires an insanely difficult and expensive auditing process for basically no benefit when you could be up and running instantly with PayPal/etc.


The reason why you should own your own platform is called “platform risk”.

https://www.startupillustrated.com/Archive/Platform-Risk/


There are several tiers here:

1. Those who have the money to spend on building a web app with their own payment processor, etc. to sell their content or services.

2. Those that use Shopify to do the payment processing and GUI tools to build the website.

3. Those who simply create an entry on Instagram or Google maps and upload a bunch of photos and their phone number

4. Those who don't even bother with the above

So it is clearly a trade off where you can outsource the hosting of what you want to present at various levels. I feel the number of people at each tier is exponentially larger than the tier above. So no, owning the platform is not going to be main stream, ever.



It's hard to distribute content from your own platform. Most successful content creators are relying on many platforms, it's virtually impossible to fight them.

But it's always to have a small resort of yours on the internet. It doesn't have to be big and great, and it can be as simple as a business card that points to your footprints on every other platform.

For example, many Russian content creators are losing their audience immediately after they're sanctioned. It's always a good thing when people search your name, your own site appears on the first page.


We've reached a point where you have choices about platforms and have the ability to export the data you input into a platform, back out. Maybe not in a format that's ideal, but it doesn't require the vendor's permission to extract your data out anymore.

We can thank a mix of European legislative changes which pushed global policy changes, some sane Large Enterprise software procurement teams (only took 20 years) who normalized asking for it, and scrappy startups with hard-line policies in the Valley who made it trendy to implement these features.


I don't think you need to own your own platform. It's a trade-off decision between total cost of ownership and your time. You can buy a new server with enough power to be able to serve everybody even during huge spikes but it's expensive, however relatively maintenance-free. You can have and old PC with low upfront costs but old consumer-grade hardware breaks down and you have to maintain it (sometimes we have to speak about "preventive maintenance"). Or you can put your things into a VPS or cloud for the price of electricity and a good internet connection alone (at least in my country, prices really are THAT low).

If you don't vendor-lock yourself in you can then move your things from one VPS/cloud to another. And you don't even have to start with Kubernetes cluster upfront. You can start with "old-school" "bare metal + operating system with root access" server, then move your application behind a load balancer, then to Docker/K8s when it grows. I see so many projects suffocating to death by "thinking big". I even encountered an e-commerce site which didn't even had a proper payment gateways, well, Cash on Delivery services still exist today. Not ideal but it worked (and the goods arrived OK).

But I'm not telling to "own your platform", that's BS nowadays. "Always have a way out" is much better. Start small and right-size often.


I agree with the sentiment of this post, but also agree that self-hosting in a scalable, and secure way is challenging.

Another thing that is challenging is the corporate security, often augmented by Google, applied to domain names that often results in your site getting blacklisted through no fault of your own. Just try getting those flags removed if your ISP becomes persona non grata.

There are so many options now to have a "mostly owned by you" platform while taking advantage of cloud services. Just need to do the pros and cons and choose wisely.


Haven[1] is your own platform for publishing privately when you're not trying to "build an audience"

[1]: https://havenweb.org


Interesting. How's that differ from PostHaven [1]?

[1]: https://posthaven.com


Haven is about _private_ blogging as a social media alternative. Post privately, share with your friends/family, even expose your blog with authenticated RSS. The crypto world has co-opted the term "Decentralized", but in a non-crypto sense, Haven is what decentralized social media could look like. Happy to chat more (and apologies for not catching your question sooner). Feel free to message me on the haven contact page: https://havenweb.org/contact.html


Looks like havenweb can be self hosted and posthaven doesn’t offer that.


With regards to RSS, I'm building a new RSS reader focussed on using RSS without being overwhelmed. That's achieved by giving priority to sources and categories, and having a limit on the number of items per source.

If that sounds interesting to you, please give it try https://lenns.io. You can import/export your subscriptions so you can come and go whenever you want.


The main issue for creators is distribution.

They go to platforms for that. If they host on their own, they lose it.

Aside, it talks about ditching the middleman.

Problem is these platforms already dominated people's attention. They are an unavoidable middleman. Creators have to compete for attention there, convincing these platforms they deserve some.

Unless these problems are solved in large scale by open alternatives, nothing will change.


The issue is that you will always need some type of infrastructure to work.

Maybe we need government investment in infrastructure. Having government infrastructure allows a lot of small businesses to work.

Take for example the road system. There are so many business that depend on the roads. Imagine if they were all owned by private companies that could kick you off at their whim?

The postal system also is very similar, though recently it has been getting less investment.

One advantage of government infrastructure is that there are laws and rights such as due process/freedom of speech/etc.

I am not sure what that would look like in practice for the digital world. One area that be useful is digital identity. Other people have mentioned it, but maybe having the Postal Service provide you with a secure identity card (similar to what Estonia provides). This would enable “social login” without being dependent on Google/Facebook/Apple/etc.

The truth is owning your own platform is beyond what an individual developer or small (or even medium sized) company can do.


Not everyone needs a platform to be honest. Most people lives their lives without posting anything online, just jumping from platform to platform searching for something to consume. And the argument that people would worried if they lost access to their data is almost invalid, most don't even backup their phones, much less data on platforms.

And for people who needs to distribute stuff, they would just opt-in for every platform, self-owned or big tech.

Removing middlemen or the whole decentralization thing is a welcomed change, so we get more to be in control more. But this is a paradigm shift, like how we moved from offline to online, the ways and methods of doing things need to change. For instance, instead of self-hosting the entire thing, I imagine a way for people to actually host their own data, while big tech host the frontend and backend. When you want your data to be displayed on Facebook, you will need to provide a temporary access key that can be revoked anytime.


On this website, it only says "please do it", but didn't mention too much on "how". Own your platform is ideal, but makes your platform discoverable, maintainable, and even profitable is hard but possible. It needs protocols (like RSS, but much better). That is what our dreamer developers should work on.


I don’t think the hosting is even that hard of a problem.

It’s the network and discovery effects of platforms that are hard to duplicate.


I think everyone reading this as "host your own servers" is missing the point. This more about owning your relationship with your users. Running a fully managed website with video content and a donation form is owning your platform even if you use nothing but SaaS products. Posting to YouTube is losing your platform.

One of my company's big products is podcasts. Podcasts are traditionally distributed via RSS. So long as we have an RSS feed of our audio and a website under a domain name we own, then we own our platform. The mechanics are immaterial. We are paying vendors for services but we own our data and our relationship to users. Losing our platform would be becoming a Spotify exclusive. Then Spotify owns our platform and we are we just get a cut.


Not everyone. And that is the issue. If everyone invent or keep his wheels, who are going to fly?

And if you do, keep your option open like apple always have their system on 3 platforms and I suspect they might even have a linux option or fpga or five…

You never really own your platform. Just partially.


I prefer a variant of this: own your domain name and make it easy to move.

I don't think everyone has to run their own server. It's fine to pay someone or use a free service. The issue is, can you change who hosts it? If you can, you are in control.


I think owning your own platform is a really cool thing to do and I think we live in a golden age for doing it. It's never been easier and you've never had more options.

I also think that placing self-ownership before all other concerns will be wrong for many people. The internet is bigger than ever and for many purposes the very real autonomy and flexibility you get out of self-hosting will be outweighed by using an ecosystem that some company is investing money into.

To me, the key is realizing that hosting is work - but it might be work that you enjoy and / or derive benefit from. When someone else does that work for you, they are deriving benefit from it, but sometimes that really is a win / win.


Pika Pods[0] (no affiliation) is an interesting take on facilitating self-hosting.

They offer a managed service that simplifies deployment of popular foss apps[1] using containerized SELinux. "Self Hosting as a Service"?

Their price estimates (shown on the apps page[1]) start at $1 - $2 per app, per month, a portion of which they share back to the app creators.

Previously on HN: https://news.ycombinator.com/item?id=31284512

[0] https://www.pikapods.com/

[1] https://www.pikapods.com/apps


I was just wondering the other day why video podcasting wasn't more popular?

Back before YouTube was really used for series, I used to follow a decent number of video diaries essentially, RSS shows, with Miro (a video podcast player that doesn't exist anymore). Ask a Ninja, Hope is Emo, Rocket Boom, Scriggity - to name a few.

I mean, sure YouTubes ad model is probably a big part of it, I'm just surprised that aren't more channels that syndicate via RSS as well. I think the desire is there though in a lot of the community, fed up with the algorithm choosing what they actually end up seeing despite their subscriptions.


I think owning your own domain is enough. Use social media to place links to your web site and blog.

I don’t completely take my own advice: I have gone back to using Blogger for the convenience, but I periodically save/dump my data.


Exactly. Gab and Mastodon have done exactly that and are self-hosting. Whether if they are alternatives to Facebook, Twitter, etc is another story.

Both are failures in that regard and it is not early days.


I used to have my own blog with my own domain from 2009-2012. The increasing popularity of Facebook in my country and the waning of my passion for writing blog post, make me abandon it. Currently I still have a blog in a couple of blogging platforms, but they are unmaintained and it has been a long time since the last time I wrote there.

Personally, I want it to go back to how it was before. I want to start a new blog with my own domain. But I don't know if I can write again like it used to be.


This post makes an epic leap from "A lot of platforms have serious problems" to "Always own your platform".


There is also in between solution. There are whole bunch of small providers that would host your website and if you keep a copy you are safe and small provider will not dictate you what your content should be (within the law of course).


Owning the content and the platform is valuable for both individuals and companies. I wouldn't want my data or content to be taken hostage and me having to rent the access on it.


I wonder if Substack qualifies as owning your own platform, since you can export your subscriber/email list even if they kick you off the platform and go somewhere else.


It doesn't, but for me keeping a copy of your content somewhere local is more important than 'owning your platform'.

I lost a whole blog I ran daily from 2005-2013 because when I set it up I didn't understand vendor lock-in. I just used the web form, fired-and-forgot about the posts. When it came time to move to a new host, the pictures all got compressed and the formatting got thrown out. It's not a mistake I'll make twice, now every blog post is stored in text files before it goes anywhere.


Medium, that's a name I haven't seen in a while.


But how can HN be so cynic about distributed systems like Bitcoin despite the scams, and talk everyday about ownership.



I own my platform thank you.


What's always blown my mind is the trend for businesses to fail to understand this. Building your entire company around AWS means a future competitor also already has full control of your company and access to your data.

And I think the most hilarious example is Microsoft, a trillion-dollar level company which chose to hand over the entirety of the web (by switching to Chromium) and mobile (by abandoning WinMo for Android) and doesn't understand how they handed control of their entire business to their top competitor.

The architects of those two decisions probably guaranteed the future collapse of Microsoft.


The way I see it, Microsoft did that because they didn't view the web or mobile as mission-critical to their success as a business. Obviously they tried to invest in those areas and failed miserably (pre-Chromium Edge was awful, Windows Mobile was doomed to fail against iOS/Android).

Here's an interesting breakdown of their profit:

https://techbehemoths.com/blog/how-microsoft-makes-billions

Azure is a huge cash cow, even in the competitive market (AWS, GCP, IBM, etc. etc.) because nobody wants to handle their own servers now. Many organizations are still a "Microsoft shop" and Microsoft makes loads of money from those contracts, not to mention enormous contracts with the government. It's interesting to see the money they make from gaming so strong; Google tried to get into gaming and keeps failing (see Stadia) and Apple seems to be having mixed results with it. Don't underestimate the strength of the Xbox brand!


Azure, however, is a platform built atop the web platform. Google, a competitor with Azure, has complete control of the platform Azure is built on. Even if Edge or WinMo aren't successful in their own right, nobody in their right mind should hand the underpinnings of their business to a competitor.

Like, this would be if the US decided to use Russian rocket motors on their nuclear missiles because their best tech expertise was the warhead.


> Azure, however, is a platform built atop the web platform

Not necessarily, though, right? "The web" isn't just web browsers and mobile devices. Most code that I write that's run in the cloud is just "compute" and is crunching numbers.

If anything, Microsoft is relying on Intel/AMD for all of their processors, as well as any other third party hardware manufacturers. Intel could force Microsoft to change all of their Azure servers in some way for them to upgrade and they wouldn't have much of a choice. Probably why Apple and Google have started to develop their own chips and hardware.


Microsoft tried to continue to maintain their own browser for years without any market share to show for it. For whatever reason, the browser just couldn't match the level of polish in Google Chrome or Apple Safari.

This current situation is stable for Microsoft - the fact that Microsoft is gaining market share with Edge means that Microsoft could threaten to fork Chromium, thus keeping Google in check there.

...as for mobile. I agree and wish Microsoft good luck.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: