> As if the people who are working in those divisions can't adapt to something new that the company wants to grow into.
Timing "we need this new team" with "We don't need this team" is a challenge. It is rare that both occur at the same time, or even the existing team's skill set matches with the new challenge.
Shifting SREs to do machine learning is set up for failure.
All tech companies I've been at have had this challenge of churn.
As an engineer, the deeper the experience in a field, the more business value you can create (all else being equal). So when you switch speciality, value creation is going to take a hit. Some skills carry across directly, others take months/years to acquire.
This is why companies can simultaneously be hiring teams and firing teams at the same time.
If the team is doing something critical for the business strategy, they need to be on-point straight off the bat. That normally is achieved by buying start-ups or doing deep partnerships. (MSFT + OpenAI comes to mind).
Timing "we need this new team" with "We don't need this team" is a challenge. It is rare that both occur at the same time, or even the existing team's skill set matches with the new challenge.
Shifting SREs to do machine learning is set up for failure.