"most developers at SAS couldn't even tell you the asymptotic runtime of a function"
I knew many like that, and it wasn't just older folks. We hired young people who were equally incompetent.
Most older folks want to be helpful. But if they're trying to do their jobs, they have limited time.
It doesn't take much time to mentor "A" and "B" players. "C" players, on the other hand, cannot be successfully mentored. You can spend quite a lot of time and never make them into "A"s or "B"s.
And in the SAS culture, if you "don't want to mentor" the "C" players, you are likely to get blamed -- accused of "not getting along", "not being a team player", etc.
Kent Beck's article says that companies should filter out their "C" players. But SAS did not filter. We retained our "C"s along with our "B"s and "A"s.
As long as SAS had a strong competitive moat, this didn't matter. Once Open Source destroyed our moat, it mattered a lot.
With a mix of talent, the best one can do is collect all the remaining "A"s and "B"s together to build something marketable. That's what the Big German tried to do.
I don't think it will be tried again. The business goal is no longer to grow the company, but to sell it. If the current owners don't sell, their heirs will.
The problem with selling is that SAS software no longer has marketable parts. No one wants to integrate their software with ours. So it's difficult to sell to Microsoft, Google, or SalesForce, because SAS doesn't fit their business.
The only marketable feature of SAS is its profitable revenue stream. That can be sold to private equity, or to a company like Broadcom that acts like private equity. Or it could be IPO'd, though as time goes on ownership seems less serious about that.
Either way, to sell the company, SAS must maintain its most marketable feature, its profitability.
To meet that goal, SAS will continue cutting costs, laying off, buying out, and pushing out people, until the company is sold.