Error handling for an API etc usually make the most sense in a central location (some middleware etc) so why would I check that each and every function call was successful.
I do like go, but the error handling is hardly the best thing about it. It’s just different. Neither is wrong.