The chart is perhaps not really representative in the long run. I think there is a difference between the original issuance and recurring payments. What I mean is say the 10% sold in crowdsale, it will recycle back into the network (part / all etc.). Some will farm and be used in currency itself and not recycle. So the continued issuance is farming / app devs and possibly PtP. There are of course a lot of ways to draw the chart and a million theories on how it will all in fact work and none of us (I think) are investment billionaires, therefore it all gets blurry.
So there will be a bunch of safecoin "in the network" and "cycling through the network" and of this continual issuance, it will be
10% app devs
10% app devs
However then these % (which should calculate and not be set) are showing again twisted logic as no matter what % farming is the network uses this to balance. So not enough resource, farming increases (you get more safecoin for farming) and then when resources are high (all sacrificial data exists) the safecoin paid to farmers decreases. So whether that is 1% or 100% the network does not care it knows it needs to increase or decrease this quantity.
Then (cause we want more complexity , kidding) We wonder what if safecoin is £X and X is 0 or infinite cost, then we get into ground (in the future) where the network cannot pay safecoin as a whole unit as they may be too expensive. So safecoin needs to have smaller units or make farming feel more like the actual lottery where some folks never win and a small group get mega rich. We don't want that so divisions are required. This may mean sub dividing safecoin or having farmers farm increments that can be transferred to client wallets for storage (getting into more safeGb now).
So lots to consider, as with any economy and this is us keeping it as simple as possible
So then as people look for a single angle all sorts of debates rage, people don't consider all the various stages and dev roadmap that has to happen in a timescale that will be market driven to a great extent.
So in a way the chart is almost correct from a single view, but maybe not all views or positions (can it ever I wonder?)
Hope that helps a wee bit, well from my single view here writing this