It's that the best use of the bladesinger abilities is actually to just hang back and do the standard mage shtick (with better AC and mobility) than to rush into combat as a barely competent melee type.
It appears a number of people in this thread have failed to understand that this is the problem I'm criticising: Bladesinger is a full mage (with better AC and Concentration saves and Acrobatics and mobility) who can also rush into melee as a competent melee type.
Full mages alone are the bar for power in 5e (less so than in earlier editions, but still so). A full mage who can compete in melee (with better AC than the heaviest armoured martial) is poor design. And avoidable. The class itself develops the concept wonderfully. It is flavourful and fun: I want to see better balanced archetypes in splatbooks
because I want to see the game expand. If you didn't play 3rd edition, you missed the egregious power-creeping of its numerous splatbooks.
Power-creeping by design appears to arise from a myopic, commercially-drive belief that player purchase decisions are principally driven by munchkin behaviour. I believe that well-written, flavourful features, that do something different and valued, are more important to players than straight power options. In my experience, groups love embracing options that don't overshadow the content they've already invested in. With the value of hindsight, 5e can do better than 3e.
Finally, for those who think that balancing the game without taking the
standard character generation system into account is fine, we're not going to reach a point where I agree with you. And that is independent of whether the game is easier to balance using points buy: stop conflating the arguments. I appreciate your input, but please don't recite the argument again here. If you'd like to explore it further, I encourage you to make your own thread.