Hacker News Reader

Reaction: Sounds like Apple's method encourages managers to select and hire one person, who is a very good fit for his team's actual current need. Vs. Meta's encourages managers to cast a net, haul in a load of fish, and hope that one of 'em is a passable fit for his team's current need. But many keep a few extra fish regardless - since he'll have more needs in the future, and fishing is a crap shoot.

Or the reverse take, at Meta, ICs find the projects on teams that excite them and they are the ones empowered to decide what is a good fit for themselves.

That's likely still true at Apple, but you gotta start somewhere.

Anyone know what internal mobility is like at Apple? At Meta it's pretty easy to move teams once you hit the 1 year mark.

Apple is the same.

> "Also, engineering managers absolutely have endless work to assign to new hires. So given the headcount, they will definitely make use of it."

This seems like the crux of the problem. Of course engineering managers always want more headcount - I am one and I would be very happy to get more headcount. My backlog is ten miles long and keeps growing.

But I also understand that if I head enough engineers to actually take care of all the things I want, I will have likely overhired, and am chasing increasingly marginal returns on investment. Some things should not make it off the backlog, because honestly the ROI likely isn't worth it.

"Optimal" staffing at the team level is almost certainly not optimal for either the company or the product. A team that doesn't have to constantly drop things because of lack of resources is likely building a lot of stuff that is poorly-justified.

In a healthy company the desire to do everything is tempered by financial reality and sound judgment - it sounds like both got tossed out the airlock at many companies.

LOL, this is also very true.

I do question how you manage your backlog though, if there are things on it that should never be implemented, maybe you should clean it out? :-p

Could make an internal stakeholder unhappy. Much easier just to let it sit in the backlog forever.

Sometimes I come back to a very detailed bug report I wrote about something I found in another team's system. It has been fun to see it grow to almost 4 years old now. Although I am not happy it gets pushed back every release, I would certainly be less happy if it was just closed as "this will never be worth it".

I had the impression Google and Amazon use the same model as Meta. But in working in the field since the 1990s I have never encountered or worked at any company that used their model. I think their model is likely superior in terms of hiring the most people in the shortest time. With the shortcoming that it might make it easy to overhire, and it might not fill the job with the best person as often.

Lots of small to massive very successful companies don't need to use their model of hiring a pool of people and then only later figuring out which role they will fill. If you're only hiring someone when you have identified a specific role and identified the correct person for that role it might take longer to hire someone, but you won't overhire as easily and you're probably more likely to have a higher hit rate in terms of getting the right person.

Google is very similar. You go through a pipeline to send you to the hiring committee, then once you are cleared and leveled, you can now interview with managers who can read your scores etc.

The funny thing is my personal experience was the opposite. I have a lot of experience in embedded software. The recruiters I interacted with at Google knew about my specialization and targeted positions that were embedded-related (like, at waymo or their other projects). However, the actual phone screen interview was with a web programmer that had very little C++ background and seemed pretty uncomfortable with me using that to solve their quiz. I ended up being weirded out by that and not moving forward. I figured I'd have to go back and practice a lot of stuff I hadn't touched in a while to actually impress their interview panel.

Meta was different. I interviewed with someone from the reality labs team and it was super specific. Our discussion and the software quizzing they did was right up my alley. Like, they specifically had a reason to want to hire me, and given what they described, I had a reason to be very interested in the project they were hiring for. It was a big shock, because the last time I spoke to someone hiring for FB they seemed to have no idea what they were doing lol. I still didn't move forward! But I was really tempted to.

I had a similar experience when I talked to someone at Apple. Unfortunately that time they wanted someone even more specialized than I was (they wanted someone to work on things to manage multiple JTAG devices - that was already innately familiar with doing that sorta thing at a hardware level). Still, 10/10, I'd apply to Apple again. I'd also apply to Meta again... although I think they basically killed the entire reality labs thing. I was so impressed by Meta. Still am.

I wasn't trying to make it about me.. I am not in embedded software, though I've dabbled in it and that would have been a nice path to take instead.

But we all specialize in something and I don't really think I've gone down a path where any of those big companies would really ever be a good place for me now. I think whatever path you go down in terms of specialization would certainly influence your interactions with their recruiting.

I have been working/interviewing since the late 90s.. when I interviewed with various companies is over a 25 year period at this point, so they were not the same companies they are today. I interviewed at Apple in the late 90s when they were a mess, and I thought they were a mess. I didn't own a Mac, they were too expensive for me as a student, so I'm sure that didn't help. I turned down a job at MS out of college because they put a delay in the process and I'd gone elsewhere by the time they contacted me. MS would have been a better place for me to go to start. Google & Amazon I interviewed at > 10 years ago when public/industry opinion of them was much higher. By the time FB/Meta really got going I declined.. I have not had a strong opinion of them. However, of all my friends who have worked at the FAANG companies the ones who work at Meta seem the happiest, with the ones who worked Amazon having had the most unhappy experience and Google seemingly in the middle.

Ahhh I remember when Google was the shining light of the software world.

I'm not sure I'll ever get over my anti-MS bias. I respect them a lot for some things, but I had a couple of real shitty experiences with their university recruiting team about 10 years ago, and I don't like the way they make everything to be managed by others. That they sell consumer software feels like a bit of an anomaly sometimes.

Have to question your life choices if you are the "manage multiple JTAG devices in hardware" expert. (In HW, we are all that person.)

> Also, engineering managers absolutely have endless work to assign to new hires

Assuming there isn't a sharp dropoff of the ROI on the work juuuuust past the point where the current team can tackle it, that means that the work in the backlog is valuable and worth paying for. By chopping people they are saying actually, this work doesn't need to be done. Whether that work needs to be done or is valuable has not magically changed. That means the company was wrong all along about whether the work in the backlog was worth doing, and had all along had too many people on the team -- despite painfully fighting for headcount to get that work done, and despite careful calculus about which teams to spend budget on, all those managers were basically categorically wrong all along. How can we reconcile all that?

This doesn't seem a plausible hypothesis. there are tons of companies which adopt the Apple model, mind included in this model. actually you have to argue to make the case for why your role is important which makes qualifying for head count even more unlikely