But of course the brand of GR that Einstein taught has a zero constant.
Nope, you're wrong. The brand of GR that Einstein taught is a way of finding
equations of motions based on
energy distributions. Get it? That's gravity, that's General Relativity. Einstein published this in 1915.
Secondly, Einstein tried to figure out what the actual energy distribution in the Universe was. He first published this in 1917 and assumed that the actual energy distribution included a
nonzero vacuum energy. (Understand? In 1917 Einstein hypothesized what the energy density was; he plugged it into his 1915 GR equations and got a nominally static Universe.)
In the 1930s, people like Friedmann, Robertson, Walker, Lemaitre, etc. hypothesized various other vacuum (and matter) energy densities, including zero. They plugged these energy densitites into the 1915 GR equations and got varieties of the expanding Universe. Of these possibilities, the zero-vacuum-energy-density hypothesis looked best at the time.
In the late 90s, with vastly improved data, we put a nonzero vacuum energy back into the
energy density hypothesis. We plugged this into the 1915 GR equations, and
just like Einstein did in 1917 and Lemaitre did in 1927 and every astro undergrad does today and those equations tell you that this Universe will accelerate its expansion.
It's always the same 1915 GR equations. Always has been. They're the same equations that predict Newtonian gravity, lensing, Mercury precession, GPS time dilation, GPB frame dragging, etc.
Once again, MM, you don't seem to know squat about any of this.