Modestly improve folding RFCS

Work in progress; but I would suggest to merge it already. Mostly to keep track of my personal notes publicly, and to link with the codebase available in https://github.com/o1-labs/proof-systems/.

[x] error term computation
[x] concrete example on how to fold a multi-variate polynomial
[x] how we go from the concrete example to the generic multi-variate polynomial
[ ] how we handle relations described by a list of constraints/multi-variate polynomials.
[ ] how lookups can be handled
[ ] how we can support "parallel folding" by using what we coined "a decomposable folding scheme"
[ ] how we designed the IVC circuit

o1-labs / rfcs

Modestly improve folding RFCS #49