Here is a thought experiment for you to understand: imagine if Gridcoin was as big and as valuable as Bitcoin, which has about $1 million dollars worth of bitcoins mined every day. A good chunk of this million dollars would be distributed based on BOINC rankings. So many people would be interested in gaining control of the BOINC servers. They could either hack them. Or they could offer to outright purchase the domain names and entity managing them, maybe they would even hire the staff running the servers. They would give appearances of operating legitimately at first. But eventually they would interfere with the rankings for their own financial benefits, either plainly maliciously, or with excuses to appear semi-legitimate (they could say "since we run the BOINC servers, we deserve a share of the profits"). The Gridcoin community would be upset and disagree with this. Maybe they would try to abandon trusting these BOINC servers, but how? They would not all agree on a solution. This would create forks in the chain. Maybe they would try to set up a new entity to run a new set of BOINC servers. At this point the situation is a mess and is no different than Ripple/Stellar to whom this exact scenario happened: part of the Ripple community abandoned Ripple and followed Jed McCaleb's Stellar fork.
Morale of the story: absolute power corrupts absolutely. You cannot give power to a central entity (BOINC servers) to control distribution of money. This is too much trust and is bound to break at some point.
And in addition to these social problems caused by centralization, what about the technical ones? What happens when the BOINC servers are down, ie. under DoS attack? How do you resolve gridcoin transaction conflicts which could be resolved by looking up the BOINC ratings? The whole gridcoin network would be unable to operate due to a few servers being down. On the other hand, a true distributed currency like Bitcoin does not depend at all on a single server. This is why being 100% fully distributed is incredibly superior to being 90% distributed like Gridcoin. Even if it was made 99% distributed, the 1% of centralization is what will eventually hamper it.
2) I think the BOINC foundation is very trustworthy and the advantages of contrubuting to science outweigh the disadvantages. Also BOINC itself is totally independent from gridcoin and well financed on its own for the public good of citizens contributing their computing power to science, so if they would tinker with the stats they lose their credibility and this would have consequences for them, eg. loss of fundingg or another team being funded to run the project servers, the software is open source anyway.
2) You can be as honest as possible, but many people still won't trust you. This is precisely why, eg., Stellar is not trusted and not embraced more widely, despite being set up as a non-profit foundation, with a charter, a voting system, being completely transparent, etc. People and companies around the world (especially those with a tendency to have anti-USA views) may not trust BOINC (hence gridcoin) because BOINC is operated in and funded by the USA. Do you think most, say, Chinese companies would be willing to fully embrace gridcoin, knowing it relies heavily on a US-based project like BOINC? No!
I can see your viewpoint though. People who care about science and who may already be BOINC users would probably like gridcoin. But most people in the world (unfortunately) don't care that much about BOINC, and when given the choice of Bitcoin or Gridcoin, they would probably go with the former (if only because of their anti-USA views, or because Bitcoin is already more widely accepted).
You simply could not say half the Gridcoins go to BOINC and half go to FAH because what if there are more users on FAH in compared to BOINC?
Also POS has its issues like POW does: if someone controls half of the currency, then they could attack the network. At this point.. it would be very easy for somone with a couple thousands of dollars to buy half of the Gridcoin out there and perform the attack, its very expensive to buy half of the BTC mining hashrate.
BOINC has not released this system yet, it is still in the works and i look forward to seeing how it would work. Even though FLDC is centralized to an extent, every possibly solution dealing with BOINC and FAH has centralization leaks, just different degrees.
Also with BOINC, anyone can create a project to work on. Someone could potentially create a cron job application masked as lets say a protein folding aplication to harness BOINC power for not only A. Gridcoins but B. computational power used for something malicous like cracking email passwords. This bad actor would then gain two benefits.
Though after awhile BOINC may discover this and remove them, the possability for even a days worth of work could be problematic. Dont get me wrong, I love BOINC, but for this system to work, they would need to have an approved list of specific projects one could work on to receive Gridcoin.