To some people.
Look at it this way, you could spend money on an expensive paint job to make your car look "cool", (subjective), or distinct, or you could do the same thing with solar panels, AND actually get something useful.
His point was that even with 100% efficiency, it's not worth it. Maybe there is a very specific circumstance in which it might be useful, but you're adding a good bit of complexity and so it depends on your goal. If you really just want to prove a point, and you have money burning a hole in your pocket, sure, go ahead. Realistically, if the goal is to further the progress of an EV, increase range, decrease environmental impacts, etc., then there are better avenues for that. It really just comes down to what your goal is.
To provide perspective, 40% efficiency is what NASA is capable of. And if it doesn't make sense at 100% efficiency, you're not going to get much at 40%.
Let's do some basic math, just to get a ballpark idea of what is possible. Let's say that the usable roof of a car for PV panels is 1 square meter. And let's assume you can get 20% efficiency (this is likely a stretch considering it isn't a flat surface). Every square meter gets about 1000 watts of solar energy at high noon on a clear day. Using the 20% efficiency rate, this means you can convert it to about 200 watts of power. If you figure 12 hours of sunlight, then on average with a standard distribution, you'll get 6 hours of full power. In this case, 200 watts times 6 hours is 1.2 kwh. If you have 10% efficiency (I think more likely for this surface), you'll get 0.6 kwh.
If the Model 3 comes with a 60 kwh battery, this would add 1% per day, on an average day. In the winters, it might be even less. And that 1% assumes no AC/DC conversion losses. We know that the Tesla Model S has lower efficiency when charging from a standard 110v plug (80% I believe) versus higher output plugs where you can get 95% efficiency. Given that this is even slower than 110v outlets, your effiency would drop more, call it to 60%. So your 1% per day could become more like 0.6% or 1.29 miles per day (0.6% times 215 rated miles).
While I have no doubt my math is simplified and wrong in many places, I don't believe there is any economic sense in putting a solar panel on your BEV unless you just want to do it as an experiment, because based on my math, you would only be adding about 1.29 miles per day of range to a Tesla Model 3. Maybe this is different for some, but in general, if your goal is to have clean energy, stick more solar panels in the middle of a field and connect them to a grid. Putting them on your car is way too indirect to help the environment or extend range. Honestly, for range purposes, you're better off using a 110V outlet for an hour per day.
Some data points came from this website:
US Solar Insolation Maps