At big tech companies, "writing a program over a few hours that is reflective of the kind of work that the company actually does" is not really a good, representative performance measure. Often times, you will be solving problems across multiple domains, outside of your area of expertise. You have to take on the role of PM, data scientist, SWE, researcher, etc.

Internal restructuring of the company may even take you from working on backend web apis to distributed databases. It is expected that you re-acclimatize and learn quickly. Giving you a take-home test to write some CRUD app isn't necessarily sampling those same attributes.

p.s.: also not a fan of classical algorithm style interviews. Clearly, they also have a bias.