Actually, that might be evidence that there really hasn't been much progress in Systems/PL.
Covid, my time in this house, and the people I was with, gave me a chance to explore Scheme and CLOS derivatives (GOOPS). It's a joy to work with, and it helps me understand Ruby and other later developments. The different hooks it gives you into your hierarchy really map out the space that's possible.
It would be nice to lose myself in BYTE. It was influential. But it feels wrong somehow? Maybe that's on me. Maybe someday I'll write a BYTE random-page generator service.
But I tend to agree - the biggest advances seem to be in hardware: now you can run Unix on an internet-connected watch, voice recognition/translation/automatic transcription on a mobile phone, and real-time ray tracing on a game console.
C++ improvements are related to the bare language itself, in regards to tooling, no one seemed to have cared to go beyond what something like C++ Builder has been doing for the last 30 years (more or less).
Python improves on BASIC, while still missing the compilation story from most BASICs.
Diving into Xerox PARC documentation about Interlisp-D, Smalltalk, Mesa XDE and Mesa/Cedar is quite revealing in what we already had 40 years ago, and how long it has taken to bring those ideas back into mainstream.