I haven't made the analysis but maybe Airbnb works because it is actually more profitable for an entity to host 100 airbnbs then to have a regulated hotel with 100 units, in which case the analogy doesn't work because that's the crux of my argument.
In the case of decentralized storage, this is not the case, as we've seemed to have agreed. Which means this:
1- The 4 data centers will host 99% of the available storage because it's so profitable for them
2- Because the supply of storage on the network is so high, the reward given to the hosters goes down (this is how FileCoin works I'm guessing, the reward needs to come down to close to the overall cost of the suppliers, like difficulty adjustments in Bitcoin)
3- Because the costs of the data centers which now have millions of disks all under centralized electricity/infrastructure is so low, the reward becomes lower than the cost of being an individual storage supplier.
4- At this point, the individuals with "latent resources" have no incentive to put these on the network, and actually are incentivized to sell the disks instead, because at least that way they'll make money.
5- Who is incentivized to BUY these disks? See #1
That's how economies of scale work and why they will make FileCoin useless.
Your argument is "maybe not", which isn't adding much. So can we agree that if the reward you would get is actually less than the electricity+bandwidth costs of you leaving your computer on all day, then you're better off keeping the 500GB free?
Just like in Bitcoin, my free CPU cycles are better kept free because if I try to mine Bitcoin I would have to leave my computer on 24/7 and it would cost me more than the BTC rewards I'd very rarely receive?
We've literally just swapped free CPU cycles for free disk space, and we both know PoW is centralized and it's not worth individuals with a laptop to mine, but somehow the free disk space won't suffer the same fate?
Prior to AirBnB property owners left their homes unoccupied while they continued to pay property tax and upkeep on them during those times. Similarly, there are people with computers powered on all day that have disk space unused while they continue to pay for the electricity costs. If you assume these are the people who are going to make up the majority of the FileCoin network, this is effectively "found money" for those folks. (Albeit probably not much money.) This might not be the way it plays out, but if it does, your arguments about margins and electricity bills are moot: these costs are not increasing after the network is connected.
edit: Also even if FileCoin doesn't throw off a ton of ETH, presumably there will eventually be a wide array of services like the FileCoin network that you can transparently trade your liquid crypto assets into. For example, consider a person who happens to have 50TB of disk they don't need for the next month but would like to perform some low latency rendering jobs on the RenderCoin network during that time instead. They could easily just trade one for the other, increasing market efficiency for those resources. This example would be a small transaction in terms of ETH, but multiply that across the entire set of all computing resources, across all services, and it is a potentially tectonic shift in effective allocation of resources.
I leave my computer on all day anyways, I'm pretty lazy. Or rather, the cost of having to wait for it to boot is greater than the electricity. The fact of the matter is that I'm paying a large premium for a small slice of usable time and I'm not in the minority.
I haven't made the analysis but maybe Airbnb works because it is actually more profitable for an entity to host 100 airbnbs then to have a regulated hotel with 100 units, in which case the analogy doesn't work because that's the crux of my argument.
In the case of decentralized storage, this is not the case, as we've seemed to have agreed. Which means this:
1- The 4 data centers will host 99% of the available storage because it's so profitable for them
2- Because the supply of storage on the network is so high, the reward given to the hosters goes down (this is how FileCoin works I'm guessing, the reward needs to come down to close to the overall cost of the suppliers, like difficulty adjustments in Bitcoin)
3- Because the costs of the data centers which now have millions of disks all under centralized electricity/infrastructure is so low, the reward becomes lower than the cost of being an individual storage supplier.
4- At this point, the individuals with "latent resources" have no incentive to put these on the network, and actually are incentivized to sell the disks instead, because at least that way they'll make money.
5- Who is incentivized to BUY these disks? See #1
That's how economies of scale work and why they will make FileCoin useless.