Man, I gotta say this TPSC guy doesn't understand Windows Defender Sandbox. He thinks Windows Defender Sandbox is a sandbox similar to Comodo's sandbox feature that puts unknown apps into a separate sandbox/container but that's not the case here. It puts Windows Defender itself into a container which is a completely different thing. Besides even this feature is not stable yet. WD acted weirdly the last time I enabled it so it shouldn't have been enabled for the test. But he keeps saying Sandbox and haha and blah blah blah. Better do some research before putting things like that in your video.
I have nothing against Leo personally, but I do take issue with how his tests/opinions are done.
This test in particular is a good example. There are two possible scenarios in my mind.
A. He really doesn't know how this feature works/is intended for, which is a problem in my opinion, as he is running a test and coming to a conclusion based on something he has no clue about. This is the reason why I'm not a fan of YouTube tests in general. Just because you can run a malware pack and have some knowledge on computers doesn't mean you know everything. If you are going to do a test/review, better make sure you do your homework on each product before jumping in, as it can lead to misleading results/info.
B. He does know how it works, but is running this test to just bash WD. This is also a problem, if not worse then A, as it just shows that he is indeed running tests and essentially rigging them one way or another. Which in turns makes all his test invalid, as they aren't being done with integrity and a non bias approach.
I'm leaning more towards it being A, but either one is still a problem IMO.