> So, where is the disconnect here?
> I can say with utmost certainty that it isn't 4000x faster
The numbers you provided do not come to 4000x faster (closer to 2400x)
> Why is actual user experience not able to keep up with benchmarks and marketing?
Benchmarks and marketing are very different things, but you seem to be holding them up as similar here.
The 5x 6x 4x numbers you describe across marketing across many years don't even refer to the same thing. You're giving numbers with no context, which implies you're mixing them and the marketing worked because the only thing you're recalling is the big number.
Often, every M-series chip is a HUGE advancement over the past in GPU. Most of the "5x" performance jumps you describe are in graphics processing, and the "Intel" they're comparing it to is often an Intel iGPU like the Iris Xe or UHD series. These were low end trash iGPUs even when Apple launched those Intel devices, so being impressed by 5x performance when the M1 came out was in part because the Intel Macs had such terrible integrated graphics.
The M1 was a giant jump in overall system responsiveness, and the M-series seems to be averaging about a 20% year over year meaningful speed increase. If you use AI/ML/GPU, the M-series yearly upgrade is even better. Otherwise, for most things it's a nice and noticeable bump but not a Intel-to-M1 jump even from M1-to-M4.