Thanks for your thoughts regarding this. You touched on some important points.
Aer means air in ancient greek. Considered calling it "air tokens" first, but there already seems to be some tokens with this name on CM. The reasoning for concentration on air is the overlay architecture for how I propose implementing it in the whitepaper. Overlay protocols (which are like off-chain computation services) and other off-chain infrastructure needed for dApps would in some sense be on top of blockchain and invisible to it (like air).
The fact that it might be fairly easy to predict the best performers in fractal meetings is a feature in my mind. We want EOS token-holders to stake the most for those contributors who perform best so that BPs that these contributors select get the most votes. Roughly speaking, the goal is to make the distribution of stake for contributors follow the average distribution of Respect. Also, while weekly rankings are predictable, it is not a sure thing, and I think that can make this game more exciting for stakers (for the same reason why betting games are attractive).
If, aside from the big respect they are getting, those people receive financing from BPs, then an unparalleled plutocracy could be being promoted. Add to that current flaws in the Fractally protocol that allow for multiple dipping of respect by teams and you would end with a ultra steep Pareto distribution of wealth.
It's not obvious that BPs will fund the best contributors. BPs will want to save and if they know that contributor is already sufficiently funded, they might not agree to give more. It's a negotiation between contributors and BPs. The best strategy for a BP might be to try keeping existing votes, while privately supporting and funding contributors who are not yet being ranked highly, but have a lot of potential. Then BPs can put a large stake on this contributor before everyone else and make more AER than everyone else. The genesis fractal rule to first select a contributor who contributes the most relative to their past contribution would be very helpful here. If this would indeed be the most profitable tactic for most BPs, I think it could actually help reduce Pareto effect among contributors and help new people rise in ranks, start working on their fractal projects full-time.
Nothing in this protocol is set in stone yet. Both this mechanism and fractally process can use refinement to make the best of this. A way for the community to update this protocol is key here, as it is probably impossible to get it perfectly the first time.
I didn't get in this document how off chain infrastructure providers would pay their bills? Only receiving AERs doesn't look too attractive to someone who has to pay servers and employees in dollars.
I propose that off-chain infrastructure providers be existing BPs of the chain (those who are willing to do that). They already have a chance to get BP rewards and if they additionally provide off-chain infrastructure, fractal contributors will be incentivized to choose them, which will increase the amount of votes they get and increase their BP rewards. That in addition to AER tokens. BP rewards can be pretty big once token price rises and we have a precedent for BPs doing more than just producing blocks. This was especially true in the early days of EOS - a lot of BPs were producing useful tools and apps for the community. They were the main source of initial dApps and other tools on EOS. Later this pretty much stopped, because these value-producing BPs weren't able to keep their BP positions where they would be paid enough. The mechanism here presents a way for BPs to increase their BP rewards by contributing more than just block production.