> Nature is by far the biggest and the oldest lab of them all, with an untold count of experiments ran.

That's a weak argument. With sound reasoning you can often reduce the solution space dramatically, even before running any experiments at all.

In other words, a single argument can be worth an "untold count of experiments", if you've modeled the problem space well.

In the limit (just to illustrate the point, rare in biological systems), you could derive a solution from its constraints analytically. In which case you need exactly zero experiments.

And even in complex biological systems, it's not like "everything goes", you still don't need to try everything. Informed reasoning helps, relative to "a lab of random mutation experiments". That's why humans have been taking their ecosystems by storm… for better or worse.