The web is and always has been decentralized. But people only want to use centralized services. What does that say? Centralized services have benefits that can make them strictly superior when done ethically. We just need to come up with better ways of funding them that don’t involve building a massive survellaince apparatus.
What’s easier: building a new internet, or forcing companies like Facebook to be honest and transparent?
> What’s easier: building a new internet, or forcing companies like Facebook to be honest and transparent?
Those companies have proved they can't be trusted.
So I believe the answer to that question is to build better, easy to use, decentralized platforms. I think that given a real choice, people would be happy to use open systems. Up until now they haven't been given easily usable alternatives that do all the things they want to do.
So I believe the answer to that question is to build better, easy to use, decentralized platforms. I think that given a real choice, people would be happy to use open systems. Up until now they haven't been given easily usable alternatives that do all the things they want to do.
The problem is that decentralized systems are always going to be less usable than the centralized systems they compete with. Centralization is what makes better usability possible.
> Centralization is what makes better usability possible.
I disagree with this premise.
There's less opportunity for monetary gain, so less effort has been invested in decentralized systems.
I don't think it's an actual inherent technical limitation of decentralized systems. In fact decentralized systems promise many technical benefits. Of course there are also challenges, but in my opinion, not insurmountable ones.
There's a kernel of truth in that premise, in that centralization enables better usability by making things within a centralized system consistent, and thus interoperable.
When you let people run away with their creative freedom, you get the Web. That is, nothing talks to anything else in any reasonable way.
I don't think so. Standardisation is what makes decentralisation possible. The only one thing that the likes of Facebook really do well compared to decentralised is spam detection.
And by the way - spam is a problem that those platforms don't solve themselves either as they use their communities to filter that out. So it's only a matter of time before those communities realise they might as well work outside those platforms.
Centralized services use specially taylored engagement mechanics for people to get addicted to their stimuli.
They do so because there is an economic incentive for them.
Decentralized will not do this, as long as they remain non-profit and strive to stay ethical.
This means that the average user will prefer the unethical and addictive platform. Same thing as with fast-food. Not only do decentralized systems need to be technically competitive and surpass hurdles that centralized systems do not have, but they also need users to be made aware of those issues.
I disagree. The "average" user nowadays knows or at least has heard of the problems with fast food. There have been numerous campaigns about eating healthy and as many people who have done the transformation from an unhealthy full of junk food diet to a healthy one with exercising can attest the benefits are there.
We need similar campaigns for an ethical internet. People need to understand the problems with the current options and be offered alternatives as good as possible. At the end of the day we can't force companies to behave "ethically"( whatever that means depending on your definition) so a decentralized options will be a better one.
There have been numerous campaigns about eating healthy and as many people who have done the transformation from an unhealthy full of junk food diet to a healthy one with exercising can attest the benefits are there.
And yet america is fatter than ever and getting fatter quicker than ever.
I think it's time we accept that most people are sheep that give in to even the most primitive forms of manipulation, even after being made aware of it.
Or to put it a bit more delicately, our brain is hardwired to optimise for short term reward functions and the vast majority of us are unable to seek longer term rewards even when consciously desiring them.
I think it's time we accept that most people are sheep that give in to even the most primitive forms of manipulation, even after being made aware of it.
So what? That doesn't mean that the campaigns don't work. There has been a decent number of people changing their lives for the better. The fact that there is a tremendous amount of variation on cultural and education levels in the States simply means that certain areas might lag behind in changes. That doesn't mean the changes won't eventually happen.
Humanity had built one in just a few decades. Yet, it hadn't managed to successfully enforce any particular human or group behavior in a few millenia. I'm not even really sure humanity has uniform idea about what honesty and transparency actually mean and whenever and how they should be enforced.
100% agree as far as ethics are concerned. But then you get maintenance costs( both physical and software) and support. Simply put companies can do this better than government agencies( and more on time wrt patching for example). Can you imagine a gov agency responding to a meltdown like exploit?
Most of the systems they developed weren't very good and have been out competed. Going back to that doesn't seem like a solution to anything. How many people prefer IRC to Slack?
I would think the main problem is centralized services can have benefits that make them superior even when done unethically ( or perhaps quasi-ethitically ).
What that says is that most of the content distribution networks which existed outside centralised systems required a good amount of mental workload for the operator. Back then we did not have the same culture of automated DevOps and so they perished.
I don't think centralisation per se is a defining feature of the platforms you're speaking about.
We just need to come up with better ways of funding them that don’t involve building a massive survellaince apparatus.
You're absolutely, completely, 100% right. This is a critical need!
Sad to say, I can think of several models that have been tried - donations, subscriptions, and merchandising come to mind. None require building a subscription apparatus! This makes them morally superior in very important ways, but all of those approaches come with distinct downsides. They all have distortionary effects on the user experience and condition the continued existence on the service on a handful of people.
Perhaps we should stop and think carefully about why surveillence-fuelled ads are so popular an approach. It's a systems that lets one monetize lurkers who are otherwise minimally interactive. Which is most users. As a result, the income base is much broader and more predictable.
We need better models. And we need to understand our models better.
The "user owned co-operative" model for these things is extremely under-explored. However I wonder if it only works for smaller communities; certainly it wouldn't give the same rapid global scaling that Facebook had.
Last week I gave a talk on the history of decentralisation, or at least my history with it. One of my examples of a successful case of decentralisation is Minecraft vs Lego Universe. Both quite similar games, but Minecraft totally crushed Lego and out-scaled it to an insane degree. As far as I can tell this is largely because Minecraft was much more decentralised (mods, user run servers). In particular Lego Universe was crippled by the cost of centralised content moderation that Mojang simply didn't have at all.
There are examples of decentralised user-owned cooperative type systems outscaling and outperforming centralised alternatives. I'm uncertain to what extent you can make sweeping generalisations ... there certainly seem to be more cases of successful centralised platforms than decentralised, but this is partly because we tend to view decentralisation as an absolute rather than as a spectrum, and partly because a few companies with very strong hiring pipelines have subsidised the creation of vast quantities of centralised software.
Great example! I would like to check out your talk if you have a recording? I'm currently working on some plugins for https://www.openmined.org to connect to Unity and Minecraft to create local minecraft servers that are fed sensor data from the real world that triggers actions within a minecraft model of the players local environment. I am fascinated by the AR layers that are possible to connect game play to real world collective action.
What’s easier: building a new internet, or forcing companies like Facebook to be honest and transparent?
Well forcing those companies to reform is simply impossible. That's because their incentives are all wrong, so they will always slither around anything you try to get them to do right.
That means that decentralization, as hard as it might be, is always vastly easier than the alternative.
That's not true. "People" don't care whether the service is centralised or decentralised. They just want the practical benefits of something that looks like a centralised service.
Comments
The web is and always has been decentralized. But people only want to use centralized services. What does that say? Centralized services have benefits that can make them strictly superior when done ethically. We just need to come up with better ways of funding them that don’t involve building a massive survellaince apparatus.
What’s easier: building a new internet, or forcing companies like Facebook to be honest and transparent?
> What’s easier: building a new internet, or forcing companies like Facebook to be honest and transparent?
Those companies have proved they can't be trusted.
So I believe the answer to that question is to build better, easy to use, decentralized platforms. I think that given a real choice, people would be happy to use open systems. Up until now they haven't been given easily usable alternatives that do all the things they want to do.
The problem is that decentralized systems are always going to be less usable than the centralized systems they compete with. Centralization is what makes better usability possible.
> Centralization is what makes better usability possible.
I disagree with this premise.
There's less opportunity for monetary gain, so less effort has been invested in decentralized systems.
I don't think it's an actual inherent technical limitation of decentralized systems. In fact decentralized systems promise many technical benefits. Of course there are also challenges, but in my opinion, not insurmountable ones.
There's a kernel of truth in that premise, in that centralization enables better usability by making things within a centralized system consistent, and thus interoperable.
When you let people run away with their creative freedom, you get the Web. That is, nothing talks to anything else in any reasonable way.
I don't think so. Standardisation is what makes decentralisation possible. The only one thing that the likes of Facebook really do well compared to decentralised is spam detection.
And by the way - spam is a problem that those platforms don't solve themselves either as they use their communities to filter that out. So it's only a matter of time before those communities realise they might as well work outside those platforms.
Centralized services use specially taylored engagement mechanics for people to get addicted to their stimuli.
They do so because there is an economic incentive for them.
Decentralized will not do this, as long as they remain non-profit and strive to stay ethical.
This means that the average user will prefer the unethical and addictive platform. Same thing as with fast-food. Not only do decentralized systems need to be technically competitive and surpass hurdles that centralized systems do not have, but they also need users to be made aware of those issues.
They are necessary but utopic.
I disagree. The "average" user nowadays knows or at least has heard of the problems with fast food. There have been numerous campaigns about eating healthy and as many people who have done the transformation from an unhealthy full of junk food diet to a healthy one with exercising can attest the benefits are there.
We need similar campaigns for an ethical internet. People need to understand the problems with the current options and be offered alternatives as good as possible. At the end of the day we can't force companies to behave "ethically"( whatever that means depending on your definition) so a decentralized options will be a better one.
And yet america is fatter than ever and getting fatter quicker than ever.
I think it's time we accept that most people are sheep that give in to even the most primitive forms of manipulation, even after being made aware of it.
Or to put it a bit more delicately, our brain is hardwired to optimise for short term reward functions and the vast majority of us are unable to seek longer term rewards even when consciously desiring them.
So what? That doesn't mean that the campaigns don't work. There has been a decent number of people changing their lives for the better. The fact that there is a tremendous amount of variation on cultural and education levels in the States simply means that certain areas might lag behind in changes. That doesn't mean the changes won't eventually happen.
Building a new internet, of course.
Humanity had built one in just a few decades. Yet, it hadn't managed to successfully enforce any particular human or group behavior in a few millenia. I'm not even really sure humanity has uniform idea about what honesty and transparency actually mean and whenever and how they should be enforced.
Back in the old days of the internet, universities and government agencies developed the algorithms.
Companies just provided the hardware.
Perhaps we need to go back to that model.
100% agree as far as ethics are concerned. But then you get maintenance costs( both physical and software) and support. Simply put companies can do this better than government agencies( and more on time wrt patching for example). Can you imagine a gov agency responding to a meltdown like exploit?
Most of the systems they developed weren't very good and have been out competed. Going back to that doesn't seem like a solution to anything. How many people prefer IRC to Slack?
I would think the main problem is centralized services can have benefits that make them superior even when done unethically ( or perhaps quasi-ethitically ).
What that says is that most of the content distribution networks which existed outside centralised systems required a good amount of mental workload for the operator. Back then we did not have the same culture of automated DevOps and so they perished.
I don't think centralisation per se is a defining feature of the platforms you're speaking about.
You're absolutely, completely, 100% right. This is a critical need!
Sad to say, I can think of several models that have been tried - donations, subscriptions, and merchandising come to mind. None require building a subscription apparatus! This makes them morally superior in very important ways, but all of those approaches come with distinct downsides. They all have distortionary effects on the user experience and condition the continued existence on the service on a handful of people.
Perhaps we should stop and think carefully about why surveillence-fuelled ads are so popular an approach. It's a systems that lets one monetize lurkers who are otherwise minimally interactive. Which is most users. As a result, the income base is much broader and more predictable.
We need better models. And we need to understand our models better.
The "user owned co-operative" model for these things is extremely under-explored. However I wonder if it only works for smaller communities; certainly it wouldn't give the same rapid global scaling that Facebook had.
Maybe not?
Last week I gave a talk on the history of decentralisation, or at least my history with it. One of my examples of a successful case of decentralisation is Minecraft vs Lego Universe. Both quite similar games, but Minecraft totally crushed Lego and out-scaled it to an insane degree. As far as I can tell this is largely because Minecraft was much more decentralised (mods, user run servers). In particular Lego Universe was crippled by the cost of centralised content moderation that Mojang simply didn't have at all.
There are examples of decentralised user-owned cooperative type systems outscaling and outperforming centralised alternatives. I'm uncertain to what extent you can make sweeping generalisations ... there certainly seem to be more cases of successful centralised platforms than decentralised, but this is partly because we tend to view decentralisation as an absolute rather than as a spectrum, and partly because a few companies with very strong hiring pipelines have subsidised the creation of vast quantities of centralised software.
Great example! I would like to check out your talk if you have a recording? I'm currently working on some plugins for https://www.openmined.org to connect to Unity and Minecraft to create local minecraft servers that are fed sensor data from the real world that triggers actions within a minecraft model of the players local environment. I am fascinated by the AR layers that are possible to connect game play to real world collective action.
Well forcing those companies to reform is simply impossible. That's because their incentives are all wrong, so they will always slither around anything you try to get them to do right.
That means that decentralization, as hard as it might be, is always vastly easier than the alternative.
Building a new internet.
That's not true. "People" don't care whether the service is centralised or decentralised. They just want the practical benefits of something that looks like a centralised service.