NO. Do not learn R2. Learn old Pochman for Corners. While it is slow, it is a lot better than screwing up the set-ups in R2. After you get good at OP for Corners, progress to Orozco/3-Style for corners. Then 3-style for edges.
Old pochman for corners is slow as hell and has too many moves. It's useless except as one way to solve parity. R2 is easy to learn and solves two faces at once. But eventually the idea is to use R, or R' or R2 with simple setups that remain in place for a cycle when convenient, then to use commutators, conjugates, and setups when convenient, and even BH when convenient. The beautiful thing is that all these methods are interchangeable if you solve in cycles, not piece by piece the way many do with R2 or Pochman, very often and unlucky cycle for one method becomes a lucky cycle for another method. Pochman is also interchangeable with these other methods to but it adds about 14 half turns plus set up turns to solve one face, (or half a cycle)! Solving two faces can take up to 30 or more half turns! There are positions where R2, R, or R' moves will solve two faces in 8 half turns. There are no such positions in Pochman.
Furthermore, 14 second solve times (that's not counting memo), are possible with R2, (especially if your setups carry over for a whole cycle). Those times are not possible with pochman.
Orozco is a good method for corners, so is BH, and just general commutators/conjugates knowledge is as well. The great thing is that you can take a freestyle approach and use these interchangeably between cycles. depending on what you use for a buffer and if you have a floating buffer.
3 style edges are fast no doubt. But there are bad cases that can be made into good cases with slicesmoves using M, M', or M2 cycles. For instance, DF-BU-BD is a nightmare for strict M2 and a nightmare for 3-style. But if you do U' (DL2 D'M) (DL2D'M') U it's an easy solve. It's a single U turn setup that is replaced after the whole cycle. There are many such cases where one way is considerably less moves than another. For instance DF-UF-RD can be tricky, but can easily be solved with M'(U'R2U) M(U'R2U).
DFR-UFR-UBR can also be a nightmare in R2. But if you modify and use R/R' instead it can be easily solved with U ((LUL'U") R (ULU'L') R') U. That same combo is something like 30 half turns with pochman and at least 20 with regular R2.
