One big issue with ability scores in general is how utterly useless certain scores are for certain classes, yet they are all priced the same. And this ties to species but also to character versatility in general. Many fantasy species definitely are thematically tied to certain competences, elves are dexterous, goliaths are strong etc. However in the game it rarely makes sense to lean into those competences unless you're playing a class that already would do so. Imagine a goliath wizard. A huge bloke who is enormously strong, also a decent wizard. High strength, high int. Sounds, cool right? This is what a goliath wizard should be like. So a wizard with strength 18. How useful would that be? I mean sure, it is definitely better than the usual strength 8, but worth all the points and resources it takes to get it there? Not in million years. For a wizard having high strength is a minor perk that would perhaps be handy sometimes. Probably about worth same as knowing one extra spell or something like that. And this applies to other concepts too, not just ones informed by the species; a fighter who is an intelligent warrior-scholar, a barbarian who is a charismatic leader etc.
I have seen many times people suggest tying ASIs to the class. In effect this would make the class' main abilities cheaper to get. But I actually feel that we should (somehow) do the exact opposite! The class' main abilities should be more expensive to get! Think about it! If you were playing a wizard, how many points of strength would you need to get to even consider trading a point of int for them? Or as a fighter, how many points of int or charisma you should get before you considered trading a point of dex or strength for them?