So now oil filters become more efficient as they load up basically because of the "magnetic" attraction of already captured wear metal particles attracting other metal particles inside the filter ...

He should realize that most oil filters won't catch a lot of the ferrous wear particles because they are well below 10u, and a used oil analysis can only see around 5u and less sized particles. Maybe he should have Donaldson do an ISO 4548-12 test with ferrous test dust that meets the ISO test dust size requirements (which has a lot of particles below 20u) and see the test results. I'd say the filter is still going to lose efficiency as it loads up.
And if oil filters did behave like he claims, then why doesn't he mention the filters affect on the ferrous ppm levels in used oil analysis since per his claim the filter should be skewing that data the longer the filter is ran. Per his claim, all ferrous particles should be caught in the filter due to "magnetic" forces ... like a built in "Filter Mag". If the filter is ran long enough, maybe the Fe ppm would head towards zero, lol. In the second video he even clearly says the filter gets more efficient right from the beginning with use up to the point it clogs and goes into bypass (8:26 and after in the 2nd video). That's totally opposite of the efficiency loss "hockey stick" shaped curve seen in ISO 4548-12, and the study that Purolator/M+H did.
As others have already mentioned, he has zero test data to prove his claim. And now it's due to the attractive forces of the metal particles going on inside the oil filter, where as in the other video his claim was that "cake layer" theory makes air, oil and fuel filters to all get more efficient with loading due to pore blockage, lol.
Note that the 1st video was posted 6 months ago, but the 2nd video was posted just 2 weeks ago, so he hasn't changed his claim on oil filter efficiency vs loading. Why don't these hosts ask him what test date backs up this claims?