What's your over/under for how many engineers AppAmaGooBookSoft will hire this year? Mine is 50k. I did a back of the envelope calculation for this last year but can't find it.
90% retention rates imply a 5 year survivorship of ~60%; feel free to take any of several thousand HNers out to coffee if you want to confirm what competent, worksmanlike execution means your 6th year pay package looks like.
I think 90% retention is (probably?) an underestimate. People tend to look at measurements of tenure, see a short number, and assume Google et al burn people out/have revolving doors. Instead, that's really a function of the still explosive growth we're going through (for better or worse; I say worse but who listens to me?)
(Which still serves Patrick's larger point.)
Since the number is dominated by Amazon, who has a huge logistics contingent, I'm not sure specifically about engineers. I'd say roughly half of that total is engineers that have something to do with software (not necessarily software engineers). That bumps your number to over 15% of the total engineering headcount. I'll be conservative and say 25% of the engineers are not software (with most being in Apple and Amazon). Now we're looking at 50k being 19% of the total software engineers.
You seem to be talking about hired at all levels, while I was talking about only senior hires. In that case, I am closer to you than I indicated. My (possibly wrong) impression is that the majority of hiring by these companies is new grads and juniors, since that is what their interview processes are heavily biased towards. I put the total number of senior hires at 10k - 15k.
As for total hires overall, I'll buy either your total hires or your retention rate, but not both. These companies are not growing headcount fast enough to be doing that much hiring at that retention rate in software engineers alone.
I was addressing Patrick's numbers (50k software engineers per year and >90% retention rate). Your link provides evidence that the latter number is incorrect, which was part of my point in both posts. In fact, it demonstrates that Patrick's 5-year scenario is the exception rather than the rule.
(Of course, both you and Patrick can be right if the typical 5-year trajectory involves only well-paying jobs, even it's usually more than one job.)