A late-to-market strategy for higher-level offerings at #aws might've seemed wise initially...
UX/DX products don't have much of an IP "moat" -- AWS can take a wait-and-see approach and then steal market share with a cheap, tightly-integrated clone
high-level products create drag, org overhead
top-of-the-stack has more risk & churn, and product mistakes are expensive for AWS because they don't deprecate anything