alt Hacker NewsSounds good to me. I am not too dogmatic about the exact implementation of who does the work and your system seems to align incentives properly while also avoiding the too many cooks in the pot/too much dilution of the money into small projects to have an impact issues.
From past failures the 2 things I want to be addressed are: 1) Have a proper procurement agency with actual experts at the helm, they are the "customer", they hand over the bids, they measure success, they should of course listen to end users. 2) Shield the project from petty internal politics. While I understand that political interference is inevitable, especially if you get public funding. The top priority is to have a good alternative to existing software in these 3 categories I defined. Not yet another job program/kickback to politically well connected entities.
The typical reason for poor procurement outcomes is the who-watches-the-watchers-ad-infinitum problem.
At stake: huge amounts of money
Naturally, people will go to great lengths to try and obtain that, by influencing the key decision makers.
So you put in a layer above the decision makers.
So they influence the layer above the decision makers.
Repeat-repeat.
The only true defense against corruption, at scale, is competition. Because then everyone tries to knife each other in the back, and it generally zero-sums.