The underlying truth behind all such discussions is that too many people focus on the mechanics of content hosting and delivery but have nothing special to actually share. So they will spend all their time engineering the perfect blog site from scratch or moving from platform to platform looking for an ideological fit while no one really cares about consuming any of their output.
My advice is always to forget about the details and just write. Set up a medium account, or substack, or blogspot, or get a VPS and install Wordpress or Ghost, or just tweet out your article line by line. If there is value in what you write then the readers will be there. And they will follow you if you decide to move your work elsewhere in the future.
I've been running a blog on a Perl/flatfile static site generator (with MySQL bolted on later for search) since 1999. Made a few tweaks over the years, but the core of the same. Had lots of people over the years suggest better ways to run a blog, but none of them were getting the traffic I got (well, used to get), and none of their options are nearly as easy to use.
If you do use a third party, please regularly download backups of your work. One of the reasons my blog was so popular was because it also has an archive of articles that I shamelessly copied from other sites (with attribution), most of which have since gone defunct, their content otherwise lost.
That last one's good advice. There's nothing guaranteeing that a given YouTube video or article, or even website is going to stay up. If I really like a video, I'll try to download it just in case it gets pulled down - either by the author or by the DMCA/platform. I haven't figured out how I'd like to rehost them but I'm sure there's something out there.
> ”If there is value in what you write then the readers will be there”
I don’t think that’s true anymore. Of course, it depends on how you define “value”, but I think having something “special” to share not always attract readers. Controversial takes, click bait, hate bait, exaggerated praising to a specific niche, all of this works better than “value” in your writing to get readers. Even if you stick to “value”, you have to do it for months or years to get readers. So consistency seems to be more effective than said value too.
I think that’s similar to SEO. 5 to 10 years ago the common advice was that you should just do white hat SEO because Google would always catch you because they were so smart, thousands of the smartest engineers on earth against your silly tricks. Google would always won. But, now, it seems clear to me that black and grey hat SEO beats Google most of the time.
In this case Google kind of painted themselves into a corner. Their ad based model was always counter to producing a good product. As time went on Google profits were more and more aligned with black hat SEO practices. They have now found themselves in a position where in order to fight the SEO spammers they would have to take a huge loss in revenue. They sort of just gave up at that point.
I get the point of results counter to common wisdom though. The pendulum is always swinging so sometimes one side is right and sometimes its the other but that doesn't make it wrong. Life requires timing along with critical thinking.
Yep, completely agree with this. I wrote about how I created my blog workflow [1] so that I can write from my iPhone and easily publish notes. And it’s a lot of work that on Substack you get for free.
Also I can’t say that I enjoyed tinkering with Go templates, finicky file syncing etc. just for publishing some Markdown text in a pretty package.
If I had to do it again, I would choose Substack, and keep my Markdown files backed up just like I do now. In the event of having to switch, I can always just publish those files elsewhere.
And if Substack also supports custom domains, I would also use that to avoid dead links.
There are cases where your own site is needed however: for unique presentations like those on ciechanow.ski, for sites where the visual look is a defining feature of what is presented etc.
Heh thanks ^_^ that is the effect I was looking to convey. You have to try it, it’s a really satisfying way of writing. Feels like handwriting but works 10 times faster with less hand cramping
Hosting is the easy part. Monetization is the hard part. Substack writers indirectly promote Substack. With a static site, you are on your own. Although, I admit I do not fully understand the concern. Does substack establish any exclusivity? Can I not simply repost my content if I decide the leave or the site goes away
My blog is about the creation of a coding language that, once finished, could allow for the compilation of the text of the whole blog into a program that would erase the blog. Which is total nonsense.
My new blog forks your language so it still results in compilation of the text of the whole blog into a program that would erase the blog, but it's a quine, so as the blog deletes, it's simultaneously punched out on the cardpunch, a nice neat deck waiting to spawn the next iteration. Oh, hold on, looks like Banksy is calling me.
My new blog is a palindrom program that does what I said first when read, and that does what you said when read backwards.
Oh, hey, it seems like Marcel Duchamp has resurrected.
In college, I did my calculus homework on blank printer paper using Sharpie markers. I could only fit one integral per page, maybe two if I was lucky, so I’d turn in these 15-page stapled assignments. It was calculus all the same, and I got great grades. I think the purpose of this comment is to reinforce that the medium doesn’t matter — if you focus on the content itself, you’ll progress.
Perhaps your integrals were artistically drawn and deserved the thick strokes and extra pages, but I find it easier to write with thinner ballpoint pens.
This reminds me of an old tip[1] about bringing your own thin dry-erase markers to onsite interviews, because you would get more whiteboard space that way.
But setting up the mechanics of content hosting and delivery is what I do it for. The actual writing is just an afterthought. And after a few years of not writing you have a good excuse to set up some different hosting because you don’t have any content to migrate.
It’s like Skyrim modders making the perfect load order with thousands of mods, but go “meh” after a few hours of playing, and the thing starts all over.
In such a case you are not a writer, you’re a tinkerer/mechanic.
For instance, I might publish a clip on my own site first, then mirror it to places like YouTube, TikTok and Reddit so more people find it. Or post an article on my own site first, then repost it on Medium and Substack (with a canonical saying my own version is the original if possible).
That way you're not tied to a specific platform, and can just find another place to share/post your work if any of them go to hell.
+1 one for the POSSE technique. Assume all walled-gardened platforms will be shut down eventually and only passion projects like personal homepages remain intact. Although even that's not a given. People just get tired and let the domain expire or lose enthusiasm and let the project rot. Archive.org might have your old site, but even that could go away given enough time. Impermanence seems to be a feature of the web, despite many saying it's a bug.
https://prose.sh employs this tactic. You control all the source files and then upload them to use. You get to keep all your content and you are welcome to host your site elsewhere.
Is there an automatic way to do this? Like a SaaS where I paste my article URL, chose where to replicate, and it does automatically? Bonus points if it gets them from a feed.
https://mobile.twitter.com/mcnostrilcom/ does this very well by posting a single panel from today's comic page and a caption starting with "Today on mcnostril.com".
"Spend your time creating something" Check: making topological shapes
"Get a domain you like " Check: kleinbottle.com
"Maintain it for a long time, [perhaps as] static html" Check: since 1996
"Share your writing, photos, designs" Check: everything's self-written & self-coded
"No investors" Check.
Results: Much learning. Fun. Friendships. Kind notes from strangers. Occasional satisfaction.
It's always an honor to see your comments on HN, been a fan for years. And I just remembered I'd always intended to buy a bottle from you, going to do so right now.
Off-topic question - have you ever heard of hypercubing? As in solving Rubik's cubes in 4 dimensions (and more?) It's a growing niche that I'm involved in and I think it might fit your interests. Would love to tell you more about it if you're interested :)
(For the record, the way we solve cubes in 4 dimensions is by simulating them on a computer, but there are also "physical versions" of the 4d puzzles.)
1. You still own your work if you post it other places, just not the platform.
2. You own the platform if you host it on a VPS, but not the hardware.
3. You own the hardware if you host it at home, but still, only license the software which runs it.
4. Many more.
You want real advice? Ignore this guy, and just start. Most of you people concerned on where you should start blogging/pocasting don't even start because of all the choices and the fearmongering (like this guy).
Just write something and hit send. Not doing that is the only guarantee of failure. Everything else is fixable.
Exactly. Whether you control the underlying operating system (or network connection) or just the content at the application layer, you're still relying on someone else. If you don't have good content to share, it's moot anyway.
At the end of the day, any service trades control for convenience/access. I run a podcast hosting service that's slightly more expensive than hosting yourself with WordPress. Paying me to host your show gets you a platform that's had 99.999% uptime for years. You don't need to think about renewing your domains or certificates or installing WordPress/Apache/nginx/Ubuntu/whatever updates. You don't need to think about how to get accurate analytics or filtering out automated requests from your log data. You don't need to worry about your show being protected by a single auth factor. If your time is worth $60/hr and my service is $10/mo, is your podcast taking more than two hours a year of your time (minus hosting costs)? Some people don't want to make that tradeoff and that's fine, but it's simply a better choice to find a third party that you trust to do stuff for you.
And then you die, and you stop paying for your own web host, and it all goes away anyway, more easily than if you had posted on Medium or whatever (because if you're still alive you could keep migrating/syndicating your data to other services if they go under).
That basically happened with James Mathe. He was the owner of Minion Games, and he had an incredible tome of board game publishing on Kickstarter advice on his blog.
He died suddenly one day in 2019, and his self-hosted site went under. You can still read it to a certain extent thanks to archive.org[1], but as thankful as I am for archive.org, they don't have the best user experience.
Meanwhile if he had posted it on blogspot or on BoardGameGeek (which supports user blogs), it'd still be fully available today.
I think it's important to have a self-hosting solution, but I think you also need an alternative. The POSSE solution mentioned above would be better than only self-hosting.
Ideally these websites would have better data preservation and archiving processes, though, and too many just don't have anything in place. Because you're just not going to be able to convince everyone to host their own.
I'd settle for automatic conversion to text-only archives, that could be good enough while making the required space much more miniscule.
Counterpoint - this is a bit like saying "sell the product from your own website, you'll get 100% of the sales and own the customer!"
Totally true, but good luck getting the customer to your site. Way easier to put your product on Amazon, eBay, Etsy, wherever people are already shopping AS WELL - and take a hit on profit (at least initially) and figure out how you build off that.
Leverage all the social media and blog sites and substacks and whatever that you can. Build a base and then figure it out from there.
I agree. I happen to sell the same things on my (pretty good!) Shopify store as well as on Amazon. Amazon outsells about 2000:1. The reason why is simple -- it's traffic. If you can get traffic on your own, say with a giant social following or TV presence, then owning your platform probably does make sense. Otherwise go to where your customers are. It's like the old retailers' saying: "Location, Location, Location."
This was basically the only option in the early 90s before geocities etc... So most of us who were on the web in the 90s did exactly that. Difference was, when you turned your computer off for the night, your site was down
If you were really fancy you had a dedicated T1 line, but those were for rich people only
I've been on my own domain for well over a decade and my CI process hasn't changed from "drag and drop a .html file written in notepad onto a BSD private-colo server via WinSCP"
Literally never fails and my kb size pages don't crush anything.
Say this to a non developer. This just isn't realistic for the majority of people, nor do they care, or should they. What we need is interoperability between services so that backups and transfers are much easier. That isn't the case just yet. If we moved to a more API first world then it would be very much possible. With the advent of GPT related stuff it's clear we're going that direction and the interface for everything is changing too. So the concept of what is your website might fundamentally change too.
Well, there used to be FTP for this (esp. with a GUI client): just move your files from one window to another, no APIs required. Or, if this was still too technical (or too little), you could setup a folder for sync.
I don't think that it's obvious that that will ever be the case. Service providers have poor incentives to make your content movable, much less your following.
> You’ll own everything, and it’ll be yours forever. No investors, no market changes, and not a single external factor will sweep away all your work there.
The internet is ephemeral. Blogging platform user accounts, domain names, VPSes, IP addresses, home servers, and software like Wordpress, Apache/Nginx, or Linux distributions don't last forever and require frequent maintenance and updates. If you passed away tomorrow, it's unlikely your content would be accessible online for more than a few years from now.
If you want your work to survive as long as possible, print your texts on acid-free materials and store them in a plastic bag within a Thermos-like container to protect against temperature, humidity, and light changes, as well as water and chemicals. Distribute copies globally for redundancy against disasters and fires, and there's a reasonable chance your work will be around for at least a thousand years.
However, we blog and put our work online not for ownership or preservation, but to get it out there. The internet is a communication tool, and it has short-attention-span readers anyways. The goal is to disseminate your work rather than own or preserve it. So does it matter if it's on a platform that will die in a decade?
If you want your work to be preserved for many hundreds of years and in a way that's accessible to everyone, you need to build a monument. Otherwise, you can pick one of the two: dissemination or preservation. Though of course, improving the longevity of your content on the internet can be done and it's not entirely black and white. But if we are looking at the big picture and using words like "forever", I don't think internet is a medium for preserving anything on that scale, nor will it exist in the form it does today a hundred years from now.
I've gone through various iterations of my blog/site. I started in LiveJournal back in the day, then moved to Blogger, and then tested the waters with Medium. Eventually, I concluded the same as the author: that I had to own my content in my own site, but that it was OK to use other established platforms for redistribution. And this is why I haven't jumped onto the Substack bandwagon.
I described the thought process I went through here: https://jmmv.dev/2016/01/medium-experiment-wrapup.html and am currently using Hugo to build my site. Granted, creating content via Markdown is slightly more convoluted than using a web UI, but the feeling of control and future-proofness I get cannot be matched.
"You’ll own everything, and it’ll be yours forever."
This is overstating it. You cannot "own" a domain name. Domains are essentially rented and eventually (whether on purpose or by accident) your domain name will expire.
I host my own domain on my own server in my own home. It obviously isn't the easiest option, but there are a lot of options. I could pay a bit extra to Wordpress to get them to host my blog with my own domain. Or I could host a Wordpress blog somewhere else, or even just static HTML files.
Actually I try to do both. I have started working on a static hand made homepage in addition to my blog. But whatever the approach, I still have to produce the actual content. Sometimes I feel that is the hardest part.
I'm writing on Substack now, because I really like the ecosystem at this point. But I'm also well aware that they could tank it like Medium did, or close up shop at any point. So I have a recurring task in my issue tracker to make a full backup. One of the reasons I'm willing to build on Substack at this point is because of the ability to dump the most meaningful data - subscribers and posts.
That said, the dumped data is not in a format that can be just uploaded to a different platform, or to my own site. That's not an issue for me, because I can write a little code that will transform the posts exactly how I need them for upload to a different platform, or my own site. When I end up needing to do this, I'll probably add custom styling that will look better than Substack's anyway, particularly in reference to code blocks.
I feel for nontechnical people. How do you decide where to invest your time? What do you do when the platform shifts in a bad direction, or closes entirely? AI tools will probably help a lot of people in this situation, if they were good about maintaining backups in the first place. I imagine "Can you convert my Substack archive to a format that will work on x_new_platform" would probably be a useful prompt.
Everyone tends to give this advice of owning your own work through a central location and syndicating everywhere else. In other words, building your own platform.
But generally speaking, sometimes you have to leverage platforms that have the current attention of the world as that is where you will drive engagement to your platform. It is fairly rare to just win organic search traffic especially starting out so you have to "hijack" some attention here and there where applicable.
If you follow any notable person online who writes, creates content, etc, they are not afraid of going where everyone else is going whether that means they may have to be exclusive with their creations for awhile or whether that platform allows them to retain the rights to their creations.
At the end of the day, it all helps them build an audience that they can take anywhere and there's many possibilities depending on what you'd like to do with that platform.
Substack is that popular thing right now because it combines both the content and email list ownership that you can then take anywhere were it to go under tomorrow. It is a very accessible way to "own your work" today, especially for those who don't want to bother setting up a wordpress blog or similar.
Nobody is going to read your syndicated blog link unless you tailor your content to the platform. There is a different way to present it to say LinkedIn, TikTok, HackerNews, IndieHackers, etc. People hate syndicated accounts, they bring very little value unless you’re already well established with popular ideas.
ChatGPT is a great litmus test here. Ask it to syndicate the main ideas of your content to a respective platform.
I’m not sure what you mean by “syndicated blog link” but my experience—and I’m aware it’s just a single data point—is that you don’t have to tailor your content.
You can do it, I’m not arguing against it.
And maybe that would result in more readers. I’m just saying it’s not mandatory and there are alternatives.
Instead of a Wordpress Blog or something I'd go for something static. There is a couple of software solutions out there from everything to blogs to website. This has the benefit that you can be a lot more independent. As long as you have some copy somewhere you are fine. You can back it up easily and so on. If you use WP (and MySQL) a lot more things can go wrong. If the MySQL server, your wordpress can get hacked for some silly plugin you thought you needed, the hoster of it or whatever are unavailable you might lose it, you might lose your domain which can break stuff, you might forget to pay, etc.
The idea here is something that allows you quick recreation and copying. You can just copy over static files. You can put it on your Dropbox, Google Drive, etc. and even on your external hard disk or whatever backup solution you use.
Usually the main dynamic thing for writers is actually creating posts. So if that is some offline software that just spits out static files it becomes incredibly resilient.
One might think static site generators, like Hugo, but there is also GUI applications that essentially equal what the admin of a blog or CMS see.
Agreed - my most recent personal site was built using a hand-rolled static site generator (some basic templating using Python) - development is now super easy and it runs just fine as yet another thing on my $5/month VPS.
A case this week that underlines the importance or both self hosting and also of archive.org is this weeks shutdown of 'zippyshare' which was used by some mp3 blog sites.
One particularly large one that shut down with zippyshare is 'Holland tunnel drive' which has been captured by archive.org in a couple of locations below.
Many similar sites had gone through similar isssues with shutdowns of megaupload and similar. Or being kicked off blogspot etc. The ones that survive are generally self hosted. But few older ones been archived on archive.org like this has.
My takeaway from this post is more general and somewhat different.
Everything dies, which is really fast in the case of software systems. The programming framework you have chosen, the note-taking software you use, and your friendships will go away someday. It's a sad reality, but it's the truth. So should we stop using things that will end anyway?
No! Of course not. You just have to be aware of it and be able to do something when things or people go away.
Regarding the topic of this post, use social media, blogging platforms, your VPS, etc. Just keep in mind that things die, and you should be able to recover from them. I am not suggesting any particular way for the way you share your writings; do what works best for you. However, I suggest not going to the extreme and abandoning everything that ends. Think of how you can make use of these mortal platforms and have a backup plan for when they eventually die.
Yeah, this strikes me as the end goal that Google, Bing, etc. are working towards and envisioning. Why redirect users elsewhere when we can just lock them into our AI-generated "search engine"? "Reveal original source URL(s)" will become a paid feature :)
Yes please. It's quite surprising to see a company like Netflix and Airbnb using Medium for their tech blogs. They have the world's best web engineering teams and they use Medium for their tech blogs ? Just why ? I hate their font sizes their font heights. Just looks so amateurish.
This was an interesting read for me today because, by coincidence, I spent some quality time yesterday trying to figure out how to go back to blogging, which I haven't done now in several years. My early blog was on Blogger then I switched to TypePad. Then I self-published. Then I went back to TypePad. I do wish I had always self-published because, while technically possible to consolidate content, it would take some work that I've never got around to doing.
So yesterday I was reviewing my options for rebooting my tech blogging, and went looking at the platforms. I was steering towards HashNode after reading the advice from another technical blogger to say to just focus on your content. It does look like a quality site for technical blogging. But for me a must-have is the ability to write off-line and it doesn't look like that is possible with HashNode. It has an API but it's WAY more complicated than is warranted and seem to exist just for the purpose of saying that it's technically possible to upload your content instead of using their online editor.
Another thing that happened yesterday is that after searching for over an hour, I was still unable to locate my old Blogger site where I was writing about computer vision some time around 2006. I've changed computers several times since then, and whatever notes I had did not make the transition to my newer workstations. So sad. Clearly this could happen even if you self-publish but I think it's less likely. I still have all of my original self-published blog articles, and those were from around the same time.
Anyway, yesterday I had settled on either using Blogger again, or using HashNode. But after reading this article and the HN discussions I'm not so sure. I am going to look back at my self-written SSG from ~2008, and see how hard it would be to have those same semantics in the cloud (wasn't an option back then). A static site is fine if you don't need the social and collaborative features.
Another thing I'll mention - in the context of the POSSE principle mentioned by others - is that IFTTT is like $5/month and seems well-suited to the POSSE approach of publishing.
Even though I agree, there will always be things that we don't own.
I'm developing my blog in Svelte, and I trust that (despite being kind of a niche framework) it will last for long. But if it gets abandoned by its developers then I'll have a problem.
The only way for me to own everything on it would be to do develop as much as I can by myself, like quit using Sass and Svelte in favor of plain CSS and vanilla JS, and also self-host it instead of relying on a VPS.
And that's not even mentioning the dependencies of the API, which is made in Rust and its less mature ecosystem
I think that the path to software preservation is a long long road
Cannot recommend this enough, but also encourage folks to set up something they’ll use because you’ll be invested in making it better. At https://hec.works/tapehendge I host all the Phish tapes I can find for my personal listening. I host around ~half a terabyte of Phish music and it costs me $10 a month, so a small price to pay to really engage with my favorite band!
It's clear enough who owns a speech, and like the OP I'm in favor of hosting your own speeches (and essays). But who owns a conversation? Does this rugged individualist approach to hosting one's own work end the possibility of online discussion? Or do we just hold our nose and do the hypocritical thing and host comments - encouraging others to NOT own their work?
Discussion is different from work. As you hint, there is no good way to co-own a discussion. Either you do it on neutral grounds and someone else owns it, or one party is the host and owner.
> Imagine a billionaire buying one of the main social media services and some things not going as expected.
If you thought that Musk was going to be beneficial for Twitter in any way, you were deluding yourself about him. Tesla and SpaceX were only successful because competent people did all the thinking and spent a lot of time managing upwards.
Read most of the content here. You have a soothingly simple and brain-to-paper writing style I enjoy. Your essay called "Age of Questions" stuck with me. It's a similar frustration to one I've had with the world recently.
I liked the first half more than the second half. I think you establish the problem very effectively, but in the second half your own indignation shines brighter than inquisition.
I've yet to see an answer that satisfies me here. Many times, inquisition is indeed used as a plausibly deniable mask for ideology. Similarly, I don't think you or I will ever reasonably be free of our own personal ideologies.
If I had a good answer, the post wouldn't be interesting to me. I could rattle off one of the common answers (technocracy, degrowth, investment in education, something something social media bad) but those seem tired and wrong to me. This kind of topic is one where it's easy to find the obvious target (this person, this system, this policy, this organisation) and start shunting blame, but that's ironically falling afoul of the problem itself circularly.
There's no solution I stand behind with enough confidence to state it on a public forum like this.
I think I found a temporary shortcut/patch for myself that I can apply on the personal level.
If I see something outrageous online I bring it and discuss in person with few people and we have more nuanced conversation and I see that questions are asked and it’s not as crazy as online.
And it calms be down and relaxes and gives me hope back. Because old dynamics - of asking questions and having nuanced conversations - they still work.
That's what I do, because: too many to choose from, not enough flexibility. And hosting anything with server-side rendering like Wordpress does not bring too much benefits.
Solution: Markdown in Git -> HTML -> Firebase Hosting over own domain (easy to migrate and replace)
I mostly agree with the article. I use social media to promote what is on my website, my books, etc.
I do use Blogger though. Twice I have switched to self hosting on my domain, and on my web site using Jeckal. Both times I switched back for convenience.
Blogspot has disabled a number of backend services so that it’s no longer compatible with, e.g., MarsEdit. I had one last blog I was still updating on a blogspot site that I just moved over to its own domain because of this.
In my case most of engagement happened on social media, and I saw little benefit from my website; so eventually I stopped updating my webpage and started posting directly to social media platforms.
Come catch me at avgcorrection.name where I put a disclaimer on the main page saying that my opinions are my own and don’t reflect the opinions of my current employer nor any of my previuos employers.
My advice is always to forget about the details and just write. Set up a medium account, or substack, or blogspot, or get a VPS and install Wordpress or Ghost, or just tweet out your article line by line. If there is value in what you write then the readers will be there. And they will follow you if you decide to move your work elsewhere in the future.