One thing that bothers me a bit is that it is hard to tell how decentralized these technologies really are. Especially when you go into direction where commodity hardware becomes useless and you need specialty equipment to do meaningful mining, there's a risk that the power falls into few hands.
This is probably not an issue for bitcoin, where the incentive to cheat would be to just gain monetary profits. In some other blockchains, serving different purposes things might be different. Governments with their deep pockets might be interested in building up capacity to be able to perform covert ops.
You can short Bitcoin, Ethereum and many other cryptocurrencies on a number of exchanges right now.
The network hashrate is 5 million TH/s and, believe it or not, there aren't a million of these ~$500 devices you mention up for sale. You would have to produce them yourself, with fierce competition.
Many have looked at the numbers and thought "I can do that", and when their mining product finally ships, the difficulty has increased by a factor of 10x. Leaving them with 5% network hashrate instead of 50%, because it turned out they weren't the only ones thinking "I can do that".
You could own your data.
There would be competition to provide the best ui. The best experience.
Anti-features wouldn't exist.
One company wouldn't control how a billion people communicate with each other.
Your grandmother would have to run her own node!
Your bandwidth at home would determine if your profile was reachable. If something you post goes viral, just imagine the fun your family will have as their available bandwidth
slows to almost nothing!
There might be competition on providing a nice "experience" but you would either pay for it, have it be ad-supported or it'd turn in to a Linux-desktop shitshow.
Imagine how much more fun spam filtering could be if it didn't have a dedicated team but it would just be you (or your grandmother) going up against the legions of determined spammers.
Answered by others.
> There might be competition on providing a nice "experience" but you would either pay for it,
... yes please
> have it be ad-supported
... our only choice with the current Facebook
> or it'd turn in to a Linux-desktop shitshow.
... and this is were you earned my downvote. Stop spreading FUD. Please! The Linux desktop is quite usable to the point were many of us prefer it over Windows or Mac.
> Imagine how much more fun spam filtering could be if it didn't have a dedicated team but it would just be you (or your grandmother) going up against the legions of determined spammers.
One word: whitelists
Another word: shared
Also, I would argue that in practice email is very much centralized, with a few large providers that manage all the complexity for their users (such as spam blocking and maintaining servers).
We need to look at websites as service providers/data carriers akin to telcos. The government has issued a slew of regulations to prevent artificially strong network lock-in with telcos, mandating wireless number portability and certain levels of device compatibility.
We need to recognize that Facebook is not a content creator as such, but a neutral carrier who transmits the data generated by its users, and that the users need their freedom and mobility to be respected if we're going to have a free and competitive market in cyberspace. There is no good reason that Facebook or Google should have such excessive unnatural ownership and access blockage rights.
n.b.: I say this as someone who is very conservative politically. Intellectual property is a government-granted monopoly that we have let run far afield.
I don't resent property ownership in the slightest, nor do I resent profit-making. I do resent an oligopoly exploiting regulatory capture and public technical ignorance (cf. Clarke's Law) to pass laws that allow them to arbitrarily and brutally crush innovative and competitive entrepreneurs.
The government has given Facebook offensive weapons against competitors. They either need to respond by giving the consumer defensive weapons against corporations whose zeal for intellectual property seeks to block or subsume the consumer's access to the marketplace, or they need to remove their interference all together.
Imagine if she had to have her own USPS mailbox, or worse, if she had to setup and manage a globally visible server on an incredibly complex circuit-switched telephony network. POTS telephone service used to be more complicated, but it quickly became very easy to setup and use because a lot of the complexity was successfully hidden behind various goods and services.
The same should be true for basic network, but true network software was only developed for professional or niche use cases because NAT severely damaged the network. Imagine if the telephone network was almost always "party lines" in residential areas. Entire categories of telephone-related goods and services wouldn't have happened.
> as their available bandwidth slows to almost nothing!
That's why there used to be a lot of research into different types of multicast
routing. We even see
> pay for it
Just like POTS.
> you would either ... or it'd turn
You're trying to apply examples from our current, highly centralized network services, when a truly decentralized Facebook would be very different, both as a service and the environment in which the service exists.
 where "server" is anything that listens for requests (in this case, the "called party")
> Imagine if Facebook was decentralized. You could own your data.
But is blockchain really an effective solution for this kind of use cases? Can't you own your data without it? Isn't it too expensive and redundant?
I can see a value in a blockchain based DNS alternative (though it would be useless without mass adoption). But Facebook killer platform in my opinion can have only one reason to use blockchain: just to use blockchain.
Our choice to analogize requests to a public HTTP server as trespass to physical property has deeply constrained the growth of the internet as a free platform that benefits its users.
The EU has new data portability restrictions going in next year, and these may help some, but they aren't quite strong enough to resolve the issue.
The same sort of competition that brought us gimps old UI? And gits various laggy gui-software?
Im sorry, but when it comes to UI- i'm for some mighty fascistic designer taking it all into his hands.
Yep. It seems to be well aligned with those who actually cared.
Often when a project gets to focused on ux it goes downhill IMO. Ref gnome spatial revelation etc.
Fine. As long as there are alternatives for the rest of us.
Example: As far as I'm aware Mac OS X is really good - but I don't like it. I really want it to be there for people who like it, but I'm happy that Windows and a number of Lknux distros exist for people like me.
We need this because internet is currently centralized and an authority in power can control the flow of information. Blockstack aims to provide decentralized internet using blockchain and that is a good thing as it makes the internet stronger.
People assume that "blockchain" is inherently good due to the buzz created by rampant speculation among Bitcoin traders, but the blockchain itself is an experimental platform that is dubiously fitted even to the narrow constraints of bitcoin, let alone forming the foundation of a "new internet".
- "running your own blockchain based messaging/mai server would be a nightmare because of spam". I think spam can actually be solved by using blockchains: given the low conversion rates in spamming, putting a small cost for sending a message can discourage spammers. Take this even further: put a variable cost to it - a higher cost for the first messages which would drop exponentially when the recipient engages in the discussion, taking it toward 0. I've never heard of this before, maybe I'll make an ICO to implement this. If you're doing it before me, I'll gladly accept some of your issued coins (my non blockchain email is in my profile)
- "blockchain apps are not scalable". Indeed, given the current algorithms they're not scalable. But it's about web scale scalable. How many apps need to be web scale level scalable and what about the rest 99% of the others? Also when we'll have a functional implementation of lightning network this might be a solved problem.
This is unrealistic. With 1MB blocks, around 2 million people can deposit/withdraw into payment channels once per week. Even using the best-performing/most-insecure layer 2 protocol ("deposit your bitcoins with me and I'll send them anywhere"), no more than 2M users would be able to join/leave the payment network per week. The Lightning Network is much more restrictive than this, since they aim for trustlessness, which further decreases performance (particularly for merchants, who want to receive large sums from a lot of people).
Unless LN aims to create a separate cryptocurrency (where you can deposit into the network but not withdraw), there's no way their trustless model can perform better than a trustful one. And if they aim to create an altcurrency, then it's not a Bitcoin payment network, and there's little reason to do it on top of a blockchain in the first place. In that case it will be like a Bitcoin side-chain that you can't withdraw from, so you're basically converting your bitcoins into this other coin, without being able to go back.
Fundamentally, the Lightning Network doesn't scale because it models global payments as anyone-to-anyone (like Bitcoin), when in fact payments flow from billions of consumers to millions of retail merchants to tens of millions of producers who pay out wages to their employees (consumers), after which the cycle repeats. Any payment network on top of Bitcoin needs to take this into account, in order to scale, since Bitcoin's throughput is so limited to begin with.
As an example, imagine you're a merchant with a $100k monthly turnover. Using the Lightning Network you, as the merchant, would need to deposit $100k into a payment channel at the beginning of the month, in order to be able to receive this amount, and this is only if the LN is highly centralized (one hop between you and the payer). For every additional node between you and the payer, an additional $100k would need to be deposited, sitting in a payment channel ready to send. This just doesn't scale.
LN is a decentralized network whose performance decreases for every node that joins it. For every additional hop your payment crosses, someone needs to deposit bitcoins into a payment channel, which both costs capital and increases the risk that this channel is exhausted and you need to touch the blockchain (expensive). A 10M-node LN would be able to send around a total of 2.1 bitcoins (remember, all nodes need to deposit bitcoins into a payment channel, and there will only ever be 21M bitcoins), and crossing 10 hops would require, on average, touching the blockchain every time 0.21 bitcoins are sent.