Nah, it really is this bad. Look at what the CPU is doing in that first IPG screenshot. You put the machine under load for long enough (and we’re not talking hours here, or even tens of minutes) and that’s what happens.
Limiting it to 45W makes it take longer to start falling apart (by about an hour on Prime95 if I recall correctly), and CB scores are higher (just got 972), but that’s still way below what this chip should be doing.
All laptops will throttle under extreme load. Nothing new there. The issue with the i9 is its supposedly throttling under the base clock speed and fairly early as well.
What do you consider extreme load? My 2012, 2.9GHz Core i7 rMBP not only maintains its base clock speed while transcoding with Handbrake but even manages to maintain a higher than base clock speed while doing so.
The authors included conversations with software authors confirming his points. I am simply naturally suspicious of empirical data, so I'll wait for something more substantive than posts here. I am not saying they are either wrong or telling falsehoods, just saying that the data set is not convincing to me at this point.