Here is where engineers get their reputation for being "difficult to deal with", "not a team player", etc. Pointing out all of the alternatives and their consequences is often interpreted as making things hard.
Option 1: Use design as-is.
This is possibly the quickest path to a goo-sensor based inclinometer because the hardware and software already exist. This is assuming that the posted code is correct. One of the cons is that not all hobbyist device programmers support the PIC16C54. Another is the OTP aspect mentioned earlier.
Option 2: Replace PIC with a flash-based A/D PIC.
A more elegant but more difficult solution. The PIC16F818 includes the A/D converter, but you will need to write the code to run the A/D converter. Since it is flash based it can be reprogrammed without an eraser. Most of the '16C54 code will drop in. This option is only valid for someone who wants to master the PIC instruction set.
Option 3: Lose the goo-sensor.
They work great on the bench. Interpreting a signal from a sloshing vial of goo in a moving vehicle is another story. I'd hate to see all of your effort result in something that doesn't do the job.