I agree - grow or die. “Do one thing well” might work well for open source Unix CLI tools, but not for companies with shareholders expecting dividends. I wish it wasn’t that way, but that is how it is.
What ever happened to just doing well enough to pay the bills, your employees, and a touch saved for a rainy day? Thousands of companies operate like this: restaurants, shops, trades, nothing you'd find on the NYSE of course. Why do you need to double in value every few years? Expand exponentially like a cancer? As your company grows and you wrangle more engineering talent into your directionless product, how much innovation are you stifling in the process? These engineers could be founding their own private companies instead of working for your public monstrosity, they could be working on products that are good on merit and bad on profit. We could be years ahead if corporatism didn't invade engineering and silo the best talent into contracts of secrecy, shattering open communication on technology in the process.
Take self driving cars. Quite a few companies are working on the same goal, repeating the same work. In an ideal world these engineers would all be in the same building, sharing notes, running joint experiments, and the end product would be sponsored by all the companies funding the endeavor. But instead we live in a world that values getting there first and 'beating' your colleagues in other companies, even if it takes longer than if everyone was collaborating. We value the dollar more than the technology.
We all learn about collaborating and helping people with their issues——really just teaching each other, the best way to learn imo——with capstone projects in engineering schools, then we forget it all when we enter the workforce and sign our NDAs and suddenly gag ourselves from talking about our work to other people in the field.