zlacker

Statements like

>I’ve directed our teams to define next-generation product families with clean and simple architectures

always make me skeptical. It's said as if teams were focusing on dirty & complicated architectures without a leader to push them in a different direction. It's a meaningless statement at the end of the day.

>>somehn+(OP)
Hypothetically what this means is that teams will now be able to design products that are designed for tomorrows markets, rather than continuing to design products that are fundamentally designed to be sold to existing customers in profitable segments. In other words, product teams don't have to satisfy existing requirements of existing customers. No more legacy consideration

In theory this this sounds great. Based on the descriptions I've heard of how Intel works internally, my confidence in this happening is about 0.12%.

>>somehn+(OP)
The most worrying aspect of this statement is the lack of vision from the CEO himself. You direct teams to execute a vision, not to come up with one.

>>somehn+(OP)
But Intel got big on its dirty&complicated architecture, didn't they? There's a reason why their processors support more than 2000 different instructions. The only relatively "clean" parts of the architecture are the base 64-bit extensions, which came from AMD, and AES-NI, where the operations were already well-defined before Intel implemented them in hardware.

So for such a statement to be coming from the new Intel CEO could indicate a radical turnaround in Intel's approach to processor development. Or, alternatively, maybe he just hasn't met the company yet.

>>somehn+(OP)
Tbh APX is a big and welcome addition. Much more useful than yet more vector instructions