I don't get this. If they are using power series, aren't they in practice still getting an approximation of the result anyway? Why would this be a better way of approximating the solution of a polynomial equation?
What Wildberger is suggesting is that, rather than taking an nth root (solution to x^n = A where A is a fraction) as a "fundamental" operation, what if we took power series with "hyper-Catalan" coefficients as fundamental operators? (This is where I get a bit fuzzy because I haven't read and understood his work.)
Galois proved that you can't have a general algorithm for solving polynomials of degree >= 5 if all you can use are +,-,*,/, and nth roots. But what if you can use a different operation besides nth roots? That's what Wildberger is proposing and apparently it works for higher degrees.
Stepping back a bit, this is very much in line with Kronecker's notion that God made the natural numbers and all else is man's handiwork. There's no avoiding infinite series for computing non-rational roots of equations, but it is possible to choose series that are easier to work with.
So setting aside the new method's practical implications, replacing an infinitely accurate approximation with a different infinitely accurate approximation doesn't feel any different.
Maybe it's a gut reaction because power series can seem so "nice" to them in their experience.
Maybe if someone explained Computable Topology to them, then they could be more accepting? But if their judgement comes from the gut, instead of intellectual integrity or reason, then I'm not sure it would be worth trying it.