Earlier quoted context omitted.
I heard this idea, or variants of it, quite a lot recently. Some of the examples I've seen it tried - I've seen the people setting it up trying to fix the outcome by carefully choosing the question, then providing expert advice on options scoped by the question. Framing of the question is a powerful tool to promote the outcome you want, and avoiding ever asking certain questions is another. Not saying it doesn't have…
Did you reply to the wrong comment? The idea I just put forward does not involve questions, it involves replacing one group of people with another doing the same job. Framing questions is already a problem with legislation. You can frame "do you want to increase online surveillance" as "do you want to protect children" very successfully!
If you are saying choose people at random to be an MP for 5 years ( or whatever ), then sure that's different and it would be an interesting experiment - though that would be a pretty stressful job to pitch people into at random.
It would be interesting to see how those random 600 people would organise to get stuff done. In the current government you have specialisation - home secretary, foreign secretary etc - you wouldn't want to keep that structure and randomly allocate roles - but if you have the 600 vote on everything then you still have a bandwidth problem.