Here's a thought experiment I'd like people to ponder: Suppose we had this decentralized hardware utopia where every homeowner had his own web server node ... such as a "FreedomBox"[1][2] or a hypothetical ISP smart router with builtin 16GB RAM and 10 terabytes of disk to hold Sandstorm[3] or whatever self-hosting app stack you can think of.
Even with those favorable conditions, I'd argue we'd still evolve towards Youtube centralization. I'd encourage people to think about why that counterintuitive result eventually happens. (As a hint, Youtube does things that localized technology installed in the home doesn't solve and can't solve.)
>If technology is meant to be [...] empowering to individuals, then turning everything into a service is an exact opposite of that.
To further explore if eventual mass migration from home self-hosted videos to centralized Youtube is inevitable, we have to be open-minded to the idea that thousands of people will conclude that "putting my video on Youtube instead of hosting it on my own server is what empowers me." Why would people think that opposite thought?
[1] https://freedombox.org/ [2] https://www.nytimes.com/2011/02/16/nyregion/16about.html
Are you talking about search and content discovery? Are you talking about replicating the most popular videos over many servers so that the system can efficiently serve all the demand to watch those videos?
Maybe existing, fully-distributed products don't solve these problems. But it isn't obvious to me that they can't.
It’s not even that counterintuitive with some capitalist forethought.
What happens when a startup decides to take on Facebook by offering a social network that focuses on protecting user privacy? We would want that startup to succeed, but how could it ever get there by charging a dollar a month? It would start with a small number of users, so the total dollars per month would be small. If Facebook was only charging a dollar, this startup couldn't charge more than that -- a social network that has fewer of your friends on it than Facebook and costs more to use is not a very attractive.
In other words, I think the subscription model would make it difficult for social media startups to succeed.
"80% of the work is done by 20% of the workers".
https://en.wikipedia.org/wiki/Pareto_principle
Where this comes into play with regards to your thought experiment... is how the larger the playing field, the fewer the large players are - they tend to get "concentrated".
Plenty of Facebook alternatives exist... but why go to them? No one uses them. Large chunks of friends/family/acquaintances aren't "there" - they are on Facebook.
Plenty of authors other than Steven King exist... but given a few spaces on shelves at kiosks, who's going to get those spots? You can guarantee King will get one of those few spots...
The majority of activity is going to condense down until a few spots. Why? Because that is where everyone else is. That's where the skilled "creators" are.
The question becomes... how do you make sure those spots are fair, open, balanced, secure, etc.
Data in the hands of a few companies? What happens when they all lean left - politically? What happens when one of them gets hacked? When a "foreign actor" learns how to game that system?
I've thought for the last decade that things will continue to get "worse" before they get "better". How much worse? How can we make it better?
Million dollar questions.
I don't intend to derail the conversation so, rather than typing out an explanation of obscene length, I will refer all interested parties to the following:
Pareto Principle https://betterexplained.com/articles/understanding-the-paret...
Basics of Pareto Distributions https://actuarialmodelingtopics.wordpress.com/2017/05/21/par...
https://subversion.american.edu/aisaac/notes/pareto-distribu...
Example of some limitations in the application of Pareto Distributions https://arxiv.org/pdf/physics/0504217.pdf
Once upon a time, to start a certain class of technology company, you needed to buy and provision a small data centre. There is a minimum capital cost associated with that. That cost presents a barrier to entry. Now, one spins up an AWS or Heroku instance.
Harvard economist Edward Glaeser argues the reason cities are more productive is they turn fixed costs into variable costs [1]. (The density allows for the expensive, fixed capital assets to be reliably distributed to variable payers.)
TL; DR It's complicated. Reducing all software as a service to being dis-empowering to individuals ignores the barriers to entry such models tear down.
[1] https://www.nytimes.com/2011/02/13/books/review/Silver-t.htm...
What? No. Plenty of new companies are being built off these SaaS providers. Why should a company just starting out build their own infrastructure, host their own CRM, build their own AI models. Companies would move much more slowely if they couldn't rely on these services.
If you purchase software that your company relies on, it belongs to you. Regardless if that third-party disappears. You still have use of the software until you can determine if you need to replace it or if it'll still work out for your needs.
Now renting the software as a service is a different matter. If the third-party goes under, well then you have problems. Problems you may have to solve immediately.
As an individual who is trying to build something, a disappearing third-party may not be detrimental if you still have the software at hand. You would not be locked down by that third-party's SaaS.
Letraset's ColorStudio cost $1,995 (in 1990 money), and Photoshop 1.0 cost around $1,000. And that didn't even buy you the upgrade, which was quite significant when layers were two point0 upgrades away.
Nah, you "own" it only as long as your rented keys are valid - what that means depend on implementation, and can mean anything from "you get to use the latest version for which you paid in perpetuity" (quite rare; notable example: Jetbrains), through "you can use only current software, as long as you pay, with forced upgrades, and only as long as we think it's not yet time to arbitrarily cut you off (and only as long as we're in business)" (the common way), to "software? what software. it's in the cloud, you have no say in anything, it's a service that works only as long as we care to provide it" (becoming the norm).
I suppose it's kind of like renting vs. buying a flat/house - the former is definitely much cheaper on a month-to-month basis, but you're subject to random landlord whims, and ultimately, you're just paying for a service - so when it ends, you're left with nothing.
As an Adobe customer, the software I use has become more “affordable” because it is easier for me to turn up $45 a month for software than the $X000 that it would cost up front. On the other hand, I am paying for software I don’t use, because to get the few programs I need I have to get the “everything” bundle.
My suspicion is that Adobe software is cheaper now because a) it has that new feeling of “affordability” (which could also be done through an instalment plan) and b) demand has gone up more than production costs have, which was already happening before they went SAAS (iirc PS CS6 was hundreds, not thousands).
Not to say that SAAS doesn’t sometimes work out the way you suggest—I don’t know enough to venture an opinion on that—but as an Adobe customer I definitely don’t feel like Creative Cloud is giving me a good deal. It feels like I am getting slowly soaked, according to some careful calculation of just how much adobe can get from me before they push me away entirely. Their CC software manager also gets a lot more cursing from me than it does appreciation.
Maybe my opinion is wrong, but if it is, I think it is a real marketing failure on their part...
I take your point but, for this one, you'd be surprised how far you can get with a decent email client and a spreadsheet.
Yes, first amendment rights are preserved in the "real" Internet through anonymous posting. And information gets shared that otherwise would not if identity were revealed.
But that begets spam, fishing, scams, and a number of bad actors that drive people towards walled gardens for communications.
I don't see how any western capitalist society couldn't fall into this trend. It's all about making more profits, no different from say buying some land with natural water sources, close it to the public, bottling the water then sell bottles. All perfectly legal and fine, until the day there's no more competition and whoever owns that land raises the prices so that some people cannot afford water anymore. That day ten people dead of thirst or shot by guards while trespassing won't make a difference when a thousand of them still can pay for that water. Still perfectly legal although not fine at all, but capitalism as we know it has no moral limitations. It's not just turning software into service being dangerous, but also what happens next.
And of course the total loss of control of our data. Lawyers using cloud services and social media for exchanging sensitive documents and discussing them may seem crazy, but I've already seen some doing exactly that. Being ignorant about the implications of today's technologies misuse can have disastrous effects.
This part is blatantly wrong. Saas enable on demand consumption and lowers tco by a great deal, i.e. indesign used to be 700$ each release, that now covers you 3 years of subscription and you don’t have to fear obsolescence.
Same with many many other software. Once you factor in obsolescence and upgrade cycles it’s very hard to keep the position of saas being more expensive.
The other point about losing control of your own stack do remain valid.
How did SaaS lead to all these things? Unless you're broadning the definition of SaaS from a revenue model and software deployment strategy to somethibg like "any software that you use that runs on remote servers". But that's the whole point of connecting computers together in a network!
The root of the problem - or at least the trunk very close to the root - is Software as a Service. It's the trend of turning software into services. Sure, it's nice for the vendors, and it's nice for corporate clients who can write off responsibility to a third party. But it also makes you no longer in control of your data, and the code that runs on it. The availability of your work becomes completely dependent on business decisions of third parties, which can - and frequently do - disappear suddenly. It's what leads to proliferation of ads, surveillance and growth strategies like dumbing down software to the point of almost complete uselessness.
If technology is meant to be - or at least is capable of being - empowering to individuals, then turning everything into a service is an exact opposite of that.