\> Making sure the money you lose goes to a good cause is very noble. However, if you want to get some extra accountability pick a foe or an anti-charity. Knowing that your money will go to a person or cause that you don’t like will push you to work extra hard to achieve your goal.

This is is like self-flagellation. The ways we try to program ourselves is quite funny sometimes, I find the putting money at stake vs anti-charity approach more interesting though, but it would have to be a high enough amount that the person should care about losing. That makes things a bit tricky, and I guess the benefit of bitcoin is that it can be cheaper to move from/to your account.  
  
For the anti-charity one, thinking about it now I guess the reason to support both options is that someone might find their beliefs more important that the money they would loose. If you put 50$ at stake for if you complete your task but don't care about that 50$ it might just not get done.

There could also maybe other types of non-monetary one like if it would automatically message someone you don't want to message could also be interesting to keep people motivated. Could be some interesting cryptographic stuff done there too. It in a way also acts as a deadmans' switch. 

 Some great points there. Did not think of sending someone you don't like a message... could be very effective indeed.

The main part I am struggling with for this project idea is how to keep people honest. I used Beeminder for a while, but as it was easy to tell them to cancel the payments, I switched from honesty to finding excuses for not having done a certain task. Then I gave up using it as it was not doing what I needed.

Best usecases for me were the ones where it was basically calling a third-party app to check if I had done the task, such as Strava or Google Fit to check if I had done a certain amount of steps. Even though one can still game those, i did not feel the need to.