[lapcat]

I am skeptical that a BigCo employer, who has thousands of engineers to choose from and countless more who could be hired, would move an engineer from embedded code in some smart home device to datacenter resource provisioning. It feels like the management equivalent of roulette.

[marssaxman]

The perspective at Google was that people would interview for Google as a whole. You weren't interviewing for any particular job, any specific group, or any specific technology; you were interviewing to become a Googler. From there you would be assigned to whatever available job seemed most suitable.

It was not an expectation that you would transfer between such dissimilar areas, but that you could - that a successful hire could be given any problem, in any part of the company, and find a way to get useful work done there. Thus there was no point wasting interview time on questions about specific tools or frameworks: what's relevant is your ability to solve engineering problems.

I did in fact know someone at Google who went from some kind of big-data map-reduce oriented thing to an embedded device project. (Long time ago, details are hazy.) It wasn't "management roulette", he just wanted to work on something different, so he did.

I myself had been doing bare-metal embedded systems firmware before I went to work at Google, but the project I got assigned to work on was Millwheel, a large-scale streaming data processor framework. shrug

[lapcat]

What I'm suggesting is that there may be a disconnect between the way companies interview, hire, and place engineers and how the companies use those engineers.

I understand that there are some employee-initiated internal transfers. Though I suspect those aren't necessarily automatically granted. What I don't see, though, is company management acting as if the engineers were all replaceable cogs who can be assigned to any project whatsoever at any time. That would be insanity, I think, because there actually is a great deal of value in experience and domain-specific knowledge and skill. Would you really want a company based on dilettantism? It sounds like — "Long time ago, details are hazy" — extreme jumps are the exception rather than the norm.

Can smart engineers learn pretty much anything given the time? Probably yes. I've learned many things over my career, and I'm sure I could learn more and different things if I were given the economic opportunity to do so. Nonetheless, I dispute the notions that generalists are somehow smarter than specialists or that BigCo assembly-line hiring identifies the smartest people. The advantage of known specialists is that they've already taken the time to learn what they are assigned to do, so the time between hire and high quality work is much shorter, maybe years shorter than throwing a smart generalist into an unfamiliar specialty.

Obviously if you're creating a new technology from scratch, there's less value in experience. But how many BigCo employees are actually doing that, as opposed to maintaining existing tech? Probably too many of the employees aspire to create new tech and don't want to maintain existing tech. Google in particular is infamous for starting new projects and then later abandoning them. (Google has the luxury of this wastefulness because it still has ye old search monopoly. Though that too seems poorly maintained, and many people complain that Google Search is worse now than in the past.)

I would say a big part of this is that BigCos such as Google are selling an ideology to current and potential employees. "We are the best of the best. Any one of us could design a whole new operating system." Yadda yadda. The ideology is part of the attraction of working there, even though the daily reality of working there doesn't necessarily match the ideal.

[marssaxman]

Oh, well! I think I agree with everything you've said here.

[adamauckland]

I don't know if you're being deliberately obtuse or argumentative.

It's not about transitioning between skillsets now, it's about future proofing.

I went from doing corporate enterprise applications on Windows NT to embedded systems when the PocketPC came out. Mobile computers didn't exist when I interviewed but because I was a generalist they had confidence that I could pick up the manuals and figure it out.

Note - I didn't work for FAANG because F and N didn't exist and A was a bookstore and G was a search engine :)

[lapcat]

> I don't know if you're being deliberately obtuse or argumentative.

I refer you to the guidelines: https://news.ycombinator.com/newsguidelines.html

> because I was a generalist they had confidence that I could pick up the manuals and figure it out

Why you though? That's the point I'm making. Of course a smaller company with limited resources may have to repurpose existing employees. But when a BigCo with a massive budget needs to do something new, they can assign existing employees with the closest expertise, hire new employees with relevant expertise, or acquire other companies. It's unclear why a generalist with no domain expertise would be preferable.

[adamauckland]

Companies do new stuff all the time which has no domain expertise, the experts are already employed in other places, hiring takes months, they might just be doing a spike to see if it's worth investing in.

When GDPR came out, some companies employed experts to deal with it but I expect most companies just gave the job to some unlucky employee.