Also agree with this. I don't think there is anything wrong with offering bounties say for percentage gains on current miners. If you don't think it's worth your time or it's not cost effective you simply don't do it. Having milestone bounties for development encourages prolonged development which is really no different then receiving a paycheck for your time. Although depending on the algo, gains may come a lot easier in some then others. First percents will yield the most amount of money for the least amount of work as optimizing code isn't linear.
This is actually approaching what we like to call a 'business model'. It's where the rest of the professionals work in the real world. There really isn't enough competition to be had in the crypto kernel development to actually make people want to try, so they don't. As I mentioned a company could really shake this up, unfortunately I don't know enough about the software to take on something like that. There is definitely plenty of money to be made here. Large chinaminers probably already have something like that.
(but yes for someone who makes quote based on last year btc price, or make up numbers... I am not entirely surprised you find that attractive
)As you say development isn't linear, meaning to get to the first 10% hashrate increase it will cost you a lot more time and effort and this won't be rewarded...
and also we don't know if the target for the full reward is realistic either... (that's why I was saying in the example I was giving, asking for 300GH/s sha256 of a 750ti... )
So yes this isn't a pretty good model and not worth my time (unless I had some time to lose which isn't the case at the moment).