In a sense yes. It is also useful to think of it as coins returned to the network.
Something that needs to be done to fully understand the dynamics is to see it from the network’s side.
In the beginning the network has 3.6 billion coins (85% of 2^32) that it can create to pay farmers as this amount drops then farmers rewards are less successful when a coin create attempt is made. This means that the network can keep paying farmers rewards till the farmers give up. No coins need to be returned to the network. You could think it like bitcoin where the block mining rewards reduce as the number of bitcoin increases (halves every set number of blocks mined), but this is very much an imperfect comparison.
When people pay for uploading then coins are returned to the pool of coins that can be created.
The real effect of coins being returned to the network is that farmers see the coin creation attempts succeed more often. But there is NO requirement that coins are returned for farming rewards to continue to be paid.
Process:. When a GET is done a formula determines if a coin creation attempt is to be made. The average of this happening is once every (1 / farming rate) GETS done.
A coin creation attempt is where the address (determined by algorithm) is used to try and create a coin. So if a coin exists then no coin is created and that attempt fails.
So if 15% of coins exist then 85 in 100 attempts to create a coin will succeed. When 40% of coins exist then 60 in 100 attempts to create a coin succeeds.
From this you see that even after 10 years there will be some coins coins created, just not as fast as the first year.
Then introducing returning coins to the network as a result of paying for uploads means the coin creation attempts will be more successful.
Thus if noone pay for uploads then the rewards for farming overall will be less and less as time goes on. So uploading means better returns for farming, but is not needed for farming,
So thus the economics is NOT pay so farmers get paid. The dynamics is much more complex and only in the eventual case is pay for uploads results in farmers being paid. For the inbetween cases the economics is NOT a direct pay to upload so farmers get paid. Even seasoned people miss the point that the system would still work if uploads are never paid for in SAFEcoin but some other method is used to prevent spamming the storage.
Obviously paying for uploads means better returns for farmers later on. Has little to no effect in the early to medium system since the pool is very large. Its only later on the effects of paying for uploads is felt, even then though its just a better paying system.
I still contend that the most likely reason uploading dies off is that no one is wanting to use the SAFE network anymore. I contend that if people are using the SAFE network then uploads will be happening and since data production is increasing so will the rate of uploads.