Per what Jim has outlined,sensorless vector mode can hold very tight speed tolerances, usually better than 0.1%. Encoders are used more for positional control, so to prevent creeping at a stop or positioning for threading. Not all VFDs use sensorless vector, so something to check when purchasing a VFD if that is an important criteria.
Problems arising with motor bearings are typically at higher voltages, bigger motors and long motor cables. I have not seen any problems reported with smaller machines that do not run continuously. Some of the VFD motors do come with a carbon brush on the shaft. In general not a problem. I would assume that there may be issues in other applications of VFDs where the motors are running for long periods, such as pool pumps, refrigerator motors, newer AC compressors, etc. There are a number of ways to address this, so the motors/bearings and electronics probably address this issue in purposely built systems that need to last for many years. There are fancy "expensive" motor cable filters, also less expensive is to use a single ferrite donut and wrap each motor wire around the donut 3-4X at 120 degree spacing. This is similar to input noise filters (not an AC Reactor or DC bus chokes which also decrease the distortion electrical noise). I typically use DC bus reactors on 3 phase input VFDs used with single phase more for prolongation of the capacitors/components life.
There are inverter/vector motors, typically the main improvement is with regard to the motor wiring insulation, so it can withstand much higher and repeat voltage spikes seen with PWM VFDs. Typically they use standard bearings, a few of the WEG VFD motors I have used do have a shaft ground brush. The other features of these types of motors is that they are designed to run over a much wider RPM and maintain their performance characteristics, so they will often be rated at full Hp up to their maximum RPM (often 2X their base speed) and also constant torque below their base speed down to basically 0 RPM. In most cases, one is operating their motors in a much narrower speed range so again a non-issue. Insulattion break down in older motors can be an issue, but many non-VFD motors are dual voltage and you are running them on the lower voltage (230 VAC) so rarely is there a problem with motor insulation breakdown expect with older motors. Most VFDs have specific motor parameters that can be adjusted for different types of motors.
As far as RFI/EMI, I have had issues with , minor speed variation/noise in cables, radio interference, and reports of erratic VFD signaling behavior when individuals do not follow VFD wiring guidelines, such as running control cables near motor cables. Typically (like 99% of the time) it is not an issue. I have had some electrical; noise issues in ancillary equipment because of VFDs, so I tend to more closely follow the VFD manufactures guidance's in their manual. Primarily use of shielded cable, layout of cables/location and grounding recommendations. The attached information from Automation Direct is a great primer on techniques for installing VFDs, worth reviewing. They have also added bulk shielded cable sales by the foot so a good source of wire/cable and also components. Great product lines for this type of work, great service and reasonable pricing. Lots of helpful information.
As far as EMI/RFI filters, you need to look at the specs and the application. Typically you would want the filter just before the power input to the VFD, so it isolates some of the electrical noise. It can be a bit more complex if you get into it, but a the hobbyist level, probably not an issue. The exception may be CNC systems, where noise contamination can be a more significant issue. Mainstream DRO's I have not heard of an issue with electrical interference, but they are rigorously tested and they probably have electrical noise filtering to minimize the interference. Interference is an issue with some of the Igaging scales adapted to machine DRO's.
