I don’t really understand all the hate for Max here. He is simply performing some real world tests and demonstrating that the upgraded Ultra GPU cores aren’t buying you a true doubling of performance that you would expect. The CPU cores do double the performance, but the GPU cores do not, at least currently.
He clearly shows that the power being used by the SOC is being limited for whatever reason and that is what is limiting performance. His conclusion, which is his opinion, is that at this time the upgraded M1 Utlra isn’t worth the additional $1,000 that you must pay because the performance value isn’t there. This is a logical conclusion.
We have to be honest and say that his tests so far have not been too problematic, take them apart and show everyone that he has done it. Tested them and used the data to show you the "current situation", which he did. Boldly and rudely critiqued some models and suggested which ones were not very cost effective, which he did.
Well, why did his comparisons and recommendations at Ultra cause some people to resent them? I guess it's like buying a car, no one likes to see their choice criticized, that's the problem.
Let's be clear about the fact that the Ultra is not a very good deal right now, considering the price difference, especially when you put your hopes in the future "optimizing what's possible". Interestingly enough, however, the rumored M2 is always going to catch up.
Apple's M1 showed us how far the chip has come, but for performance reasons, we all know that single-core performance is still the most important metric, and that's an undeniable fact. So in theory, the M2 should be more powerful than the M1 in terms of single-core performance. Then, considering that in life or work, it is likely that more than 80% of the situations are dominated by single-core performance. In this case, the more expensive M1 (or M2) with more cores is obviously more suitable for busy people who are always catching up on "critical minutes".
For this reason, I always include single-core performance as the most important reference (and it turns out to be so, as the single-core performance of M1, M1 Pro, M1 Max, and M1 Ultra is hardly different or even the same), based on the premise that :
1. the price difference of Ultra
2. M2 possible release time
3. the ratio of time in a day that single-core performance covers in life/work
Uh .... At the moment it may be wiser to consider a medium model.