As a comparison, racing performance. You likely can’t feel in your body the difference between a 2 minute lap time and a 1 minute 58 second lap time. However, you should obviously push for the faster time.
That's because in your example the racing time is what matters, not what you perceive. If you perceive to be faster but the timer says you are slower, what matters is what the timer says, not what you perceive.
With music, what you perceive is all that matters, not the actual signal you are hearing. That's why two technically very different signals can be exactly the same from an audible point of view even though they might look vastly different e.g. under a spectrum analyzer.
The spectrum analyzer might tell you they are different, but if you cannot discern audible differences that's what matters from a listener's point of view.
Can be the same with anything we do. There is a reason lossless isn't default but that doesn't mean people who want the “best” shouldn't choose it even if they can’t tell the difference. Should be as close to the studio recording as we can possibly get within the limitations of the hardware you are using to listen to it.
Lossless has additional advantages over lossy formats on top of audible performance, meaning that someone might choose lossless over lossy even if there are no audible differences.
Said that, sometimes lossless has disadvantages over lossy which might tip the balance towards lossy even if some minor audible degradation takes place. As example, Bluetooth bandwidth is insufficient to support lossless, so if you value the comfort of wireless Bluetooth earbuds/earphones, you have to de facto compromise on that.