Who wants to be Intel’s new CEO?

People game OKRs, but it doesn't matter. Planning work a quarter out for your team is very valuable, no matter what you call it.

I was the TL of a very small team at Google. My manager and nearby senior engineers were very good at forcing me to plan my team's work for the quarter in the form of OKRs, and the results were quite surprising. I would say something like "I'll add that to next quarter's OKRs" and people trusted that it was actually going to get done. There were no weekly status meetings, or nag emails, or anything like that... people assumed that if my team said we were going to do something, it was going to be done. And they were right, because we actually focused on exactly what we said we were going to focus on. We had to prioritize, and so we spent a little more time on features than on system stability than I would have without any plan ("just one more day to automate this part of the release that only saves 5 seconds once a month!"). It was quite good.

Meanwhile, when I first joined Google, I went to the company OKR scoring meeting. There was a KR like "have one billion active foobar users" and the score presented was .7, the minimum "passing" score. They did not have 700,000,000 users though. I distrusted company OKRs ever since. I can do multiplication; you lied.

TL;DR: planning is valuable. Making up a number to make yourself feel good is pointless. The key is to not tie self-directed and self-graded OKRs to anything important, just look at it like a reasonable todo list for the next 3 months.

Re: definitions are important

Yes and no. Some of the biggest conflicts I’ve seen started because of misunderstanding ‘basic words’. So definitions matter in the sense that they help clarify intent. But they also can make things harder if you insisting on a certain definition comes across as condescending.

Sure.

Intel has an insane treasure trove of technology that they don't seem to know what to do with because they have confused profit with market domination (related ideas, but not equivalent, and one often leads to the other in both directions).

Example: Intel obsesses over x86 in a very weird and unhealthy way. Intel has repeatedly developed new CPU architectures that did not suffer from the burden of the technical debt that has accrued.

A lot of money was sank into, for example, the XScale ARM series. What killed it off was not because there was something technologically inferior with the product (it was rather popular for the time, the Snapdragon of it's age): what killed it was upper management thinking ARM sales could harm x86 sales, instead of just focusing on making sales happen.

The result? Marvell bought it, and turned merely took the breaks off the product line and let it naturally evolve. Marvell didn't really do anything other than that; Intel intentionally strangled the project and that's why it was sold.

That made no financial or logical sense no matter how much Intel apologists try to spin it. Intel cannot change the reality that their x86 designs are not for low power (which they finally admitted when they killed off the Atom SoC line for phones and tablets, which was drank power like it was going out of style), and Intel has locked itself out of one of the largest markets in the world.

Another Example: I have two Haswell-era machines in my house. Both of them are intended to be used with CPUs that have iGPUs. They both have mini-Displayport plugs. Neither have Thunderbolt controllers.

My Ivy Bridge-era MBP? Has a Thunderbolt controller, and it is the only machine in my house that would never benefit from it. The inclusion of said Thunderbolt controller was at the behest of Apple, not because of Intel.

What could have happened? Every Intel machine since Sandy Bridge or Ivy Bridge should have had a Thunderbolt controller baked into iGPU, and Intel's sales of Thunderbolt controllers would have gone through the roof. Instead, Intel has to compete with USB 3.x, produces controllers intended for TB-over-Type-C, and does not produce USB 3.x device SoCs (essentially losing money on every sale since Intel should be doing this), only integrated host controllers for existing Intel platforms.

The job of the CEO is to promote a culture inside of management where such decisions happen regularly. As CEO, I wouldn't be the one making decisions such as, say, making onboard Thunderbolt the default: I'd be making it possible for such decisions to happen naturally, to make sure the proper behavior is rewarded and diagnose issues in the corporate policy on why Intel isn't promoting the desired outcome internally.

From everything I have seen, Intel doesn't particularly have a toxic culture internally, not in a way that compares to the posterchildren of that particular disease. What Intel has is a weird obsession with the past, and an absolutely immense treasure trove of technology that just sits there and isn't sold or marketed meaningfully.

I think people have written entire books about the missteps of Intel. The problem is, all of us here know they were missteps, but Intel management glorifies them as the right choice and what lead to Intel's greatness.

I just see a company that doesn't understand how to profit in the right ways, and instead has lead to a public perception that Intel doesn't care about customers, hides ticking time bombs in their products intentionally (Meltdown et al.), and has lead to the term "Intel Tax".

AMD doens't sell a better CPU, nor a faster one. What they sell is a CPU that is fit correctly for the market, and have a public perception of actually caring about customers, caring about the long term support of their products, and try to leverage their technology in ways that benefit both the customer AND the shareholder.

If I was tossed in as the CEO today? I'd try fix the public perception, try to figure out why Intel just sits on endless technology and isn't part of the right markets (even though they have or had technology that fits those markets), and try to figure out why good ideas (and the people that make them) are not promoted from with in.

Intel is pretty much no different than Broadcom or Qualcomm, yet Intel can't sell the stuff they need to in a way that makes sense. Yet, every cell phone has Qualcomm parts in it, tons of IoC and network and NAS stuff has Broadcom parts in it (due to their huge family of ARM and MIPS SoCs), Broadcom also sells more Ethernet controllers than Intel.

So, what the hell. If I become CEO today, I am delivering disruption to the management culture that prevents Intel from thriving the way it should.

Was it obvious back then that Flash would be dead by now, Microsoft gives up on phones, x86 fails at gaining any foothold in the phone business and intel fails at performing the next structure shrink?

Not always. I once worked for a company that had an ombudsperson who reported only to the CEO. The ombuds' job was to investigate employees' complaints without fear of being fired him/herself. In that company, the system worked well.