Connecting thermal pads to a plane will make them harder to (hand) solder, but it should be manageable with hot air. Be sure to be patient and give the heat time to sink into the extra thermal mass. You should probably pre-warm the underside to give yourself a head start. You can move the hot air around a little bit as you wait for things to heat up to avoid locally overheating anything. Carefully observe the components with the thermal pads to ensure they fully “settle” as they reflow, since you want to make sure the solder on the thermal pad is fully melted.
Some design rules also reduce the aperture of the window in the paste mask over the thermal pad so that the amount of paste on the underside pads is somewhat reduced - just be careful you don’t have a big ol glob of solder paste there otherwise it could squeeze out and make shorts with your other pins.
Consider adding thermal spokes to the vias, pads, and thru-holes connected to the heatsinking GND plane if those pads/holes are not intended to conduct heat, high current, or mechanical support - it will make them easier to solder by reducing the thermal conductivity to the plane. Don’t spoke anything in the thermal paths.
Side note, just eyeballing the pictures but the motor traces could be widened to the width of the pads of the driver IC - motors can pull a lot of current in locked rotor conditions. I’m sure you’re current limited by the driver and battery and it’s probably no big deal but it doesn’t hurt to have the insurance. Always good to do a once-over checking for any potential high current paths. You’re not especially space-constrained in this design so there is no reason to go with the smallest settings on everything. You can check out some PCB trace width calculators to make sure your worst-case scenarios are covered.
Sometimes you can find spec-bending similar-but-different tech implementations within closed device ecosystems that aren’t meant to interface with third party equipment. With that said I doubt this cable is an example of that.