I agree, they're not incompetent buffoons and I won't say that they are. What we do see is that some people get stuck in the past or in old ideas, and basically throwing the same people at the same problems just provides similar outcomes.
It wasn't until Keller was flown in to AMD's CPU division that things took a radical change for the better. Similar things must occur at Intel - but when Intel had to make that move, they hired Raja and Pat G for this and on CPU they never did a full reset while Raja was destined to start with the scraps of decades of Intel IGPs. Old boys doing old tricks. Pat G was always Intel. And Raja... honestly, sure, the man knows his way around GPUs. But he was put in the wrong position, the man shouldn't get screentime, and shouldn't be doing the project management. He makes mistakes. Big ones, and they've been evident not just at Intel but also at AMD. Overselling every time, and underdelivering as a result. Even at AMD I think his hands were tied, like you pointed out, but at the same time, if he's truly an insightful engineer in a leading position, shouldn't he be lining out a different strategy to make the best chip? Vega didn't just lack in performance (per watt) but it was also a pita to source HBM. Polaris, while 'successful' in sales, didn't move GCN forward much, and RDNA1 was a bunch of nothing if you ask me, stopgap gen if anything. Was the strategy of AMD truly to execute several generations of GPU with that little progress towards matching the cutting edge? They literally plateaud at Vega performance, hard, for a long time, and up to this day the GPU strategy often seems to lack direction (at least in the PC DIY space).
Now let's look at what Pat's done for Intel as CEO. Overselling... underdelivering. I'm sure they're great engineers. And I'm sure that's what they should have kept doing. I strongly doubt Raja was treated at Intel like Keller was treated at AMD. Part of that is company culture, arrogance vs underdog mentality must have played its part here, too.
