What you describe are perhaps some implementation inefficiencies. Many ABS systems operate the actuators in a binary fashion, and they are not designed for variable brake pressure, only for pressure/no-pressure. There's nothing preventing a better design, only status quo.
Whatever you're claiming to be sensing so well, a computer can certainly sense better -- with proper sensors, that is. Most cars do not have a 6DOF inertial platform, even though that would be a good starting point for any decently-performing stability/traction augmentation system.
A computer-controlled antilock system, with servo actuators (vs. binary on-off valves) can, and will, in stable enough conditions, control wheel slip down to 1-2% accuracy or so. It will maintain that wheel slip way faster than you or I can. It needs an inertial platform, or at least a longitudinal acceleration sensor, to do that. The wheel speed sensors are not really reliable for estimating the vehicle speed once the wheels start locking up. You need inertial reference for that. The wheel speed sensors are only useful to compare the individual wheel's speed to that of the car, given that you already know car's speed!
You claim that ESC "uses wheel spin speeds to measure not only slippage, but vehicle travel direction" -- sorry but ESC typically uses a lateral accelerometer combined with estimate of vehicle's speed, and with steering column angle.
As for cruise control: those are purposefully designed to be soft. It's rather easy to have cruise control that will keep your speed to better than 1% under all reasonable conditions. It will need input from an inclinometer (inertial reference!). I have had a Volvo 940 wagon with a rather sloppy cruise control that I replaced with a custom controller with inclinometer, and a beefy electrical model servo to replace the original vacuum-controlled actuator. After model identification work was done (I settled on multiple FIR models), I took it for a spin in some rather hilly terrain and you could hardly see the speedometer needle move. On typical "flat" roads, it felt rock solid, and the measured speed would be within a +/- 0.5% band around the setpoint. Oh, and it worked down to 10mph, not to the silly factory 25mph limit.
The biggest point with "solid" cruise control is that other cars aren't solid at all, so on highway it may be advantageous to have piss-poor PID-based stock cruise control -- it will maintain inter-car distance much better if the car in front of you is on cruise control, too.