If Sam Altman had asked me, yes ME, about justifying DNNs as a technology. I would have told him, "my brother is Moses, and respectable Jewish person, never, ever say 'We don't know how it works' or 'It's a black box'".

Because, believe it or not, I fell into this trap about 2 years before the explosion of GenAI. I had made a facial recognition server, with liveness detection (it's the 'PuertoRECO' project in my Github) and the guy, some Turk or Arab in Germany, asked me 'so how it works? explain to me how it works!' --- and Godgang it, Arabic is enough of a 'violent-sounding language', add German, another such language, to the mix, and when the guy spoke English, I literally felt threatened.

So I told him "It's a black box, sir" and we got into an argument. I was still paid, but not fully, and I had already finished the project, so I put it up with GPL on my Github. That repo was what started my journey in 'computer science excellence'.

So, when these idiots say 'we don't know how it works', they mostly mean 'we don't know how it OPERATES'. Thank Lord English is such a lexically diverse language!

These idiot Jewtubers have taken this comment buy (I think Sam Altman?) and are running with it, but it's like running with scissors, even an idiot like me who did ML engineering for 2 years, and is only up-to-date with the academic work of 3 years ago (which is difference between Jurassic period, and the dictatorship of Nero, given how quickly the academia is stacking up!) could tell you, that's NOT what they mean, pretty lady making anti-AI slop!