Well.. yeah. Hiring somebody who failed the interview is pretty dumb, obviously.
Although the author made one point that I think is inaccurate: He said it's wrong to hire a front-end engineer to do backend work. In my experience, most FE engineers are more than capable to switch to the back-end, but the opposite is rarely true.
"Availability" means being able to access the blockchain/system/network and get the data you want, when you want it.
The way most blockchain systems find out which nodes are on the network is usually through a gossip protocol, so typically your app will start off with some "seed nodes", which it contacts, then gets information about other nodes on the network.
Once you have the IPs of say 100 nodes scattered around the world, on different networks, running different OSes, etc. the chances of losing access to all those nodes is virtually zero.
Edit: You seem to be thinking about blockchains in terms of how traditional cloud/web based systems work. That's the wrong way to think about it. Blockchains have fundamentally different architecture. They do not use load balancers.
Look up DNS poisoning, spoofing, ISP manipulation, etc. There have been quite a few attacks on it.
Maybe "flaky" is a bit too strong of a term. It generally works I guess.
The main advantage blockchain offers is not really within the realm of getting the right answer to a DNS query, it's more to do with giving the owner of the domain name 100% control of the domain through cryptography and smart contracts. This also makes the process of domain transfer a lot more efficient, cutting out the need for middlemen like Godaddy, etc.
Connect to multiple nodes and this won't be such a problem.
Obviously if a single node has a disk failure or OS crashes, and you are connected to that node, the blockchain is down for you and everybody else connected to that node .... but most of the other nodes are still up and the system as a whole is up.
One "killer feature" of blockchain is availability. For example, the bitcoin network has something like 99.99999% uptime since it was invented. I think is was down for something like half a day back in 2009 due to a bug, but since then it's had basically 100% uptime.
Blockchain can be used as a rather expensive and crude database, so one possible application is a key-value database which is guaranteed to always be up.
I suspect that as blockchain technologies improve and better techniques for sharding are developed, this could become quite a common real-world application.
From reading the comments most people here don't have any idea about the research that's currently happening in the blockchain space. They're just making ignorant comments about ponzi schemes and money laundering.
One of the actual useful things it can do is DNS. The current DNS system is about 40 years old and very flaky. Blockchain can do DNS better than the current system.
Blockchain based DNS allows people to actually own their domain, instead of the current system where we basically rent domains. It also makes it impossible for domain to be seized.
"However cryptocurrency does not provide any technical answers to the inefficiencies since its entire existence is purely predicated on the appeal as a speculative investment first and not on its efficacy to transmit value."
Although the author made one point that I think is inaccurate: He said it's wrong to hire a front-end engineer to do backend work. In my experience, most FE engineers are more than capable to switch to the back-end, but the opposite is rarely true.