On the other hand, I actually like the fact that this is the case. So, instead of seeing how this show ended up as a bad thing, I see it as a good thing because frankly, I find engine programming and the type of problems that Casey has had to solve on the stream to be interesting, entertaining, and educational.
Personally I love the direction that HMH has veered off in (focusing on advanced niche topics for extended periods of time), but then I watch more for background entertainment than for any kind of learning purpose so I'm probably not in its "target audience" to the extent that Casey intends HMH to be educational. So how I feel personally is besides the point...
A catch-22 seems to have arisen whereby HMH is now largely beyond the technical reach (in the sense of being able to follow along easily, having sufficient prerequisite knowledge, etc.) of a large portion of the audience that it was presumably originally intended for (novice to early advanced). More advanced devs are probably going to tend towards skipping around to topics that are relevant/interesting/entertaining to them. And to the extent that there are any novice/intermediate developers somehow hanging on to the series through all the advanced lighting/SIMD stuff/etc., I'd argue that they're grossly misallocating time that could be spent focused on far more fundamental things. Aside from being a great reference tool and potential entertainment/idea source for more advanced devs, HMH seems to offer its greatest tangible educational benefits to a very narrow segment of advanced but not too advanced developers, while wholly cutting out intermediates and novices.