> The law California (and other states) passed doesn't define what content has to be blocked for which ages

No, but it's a framework that would allow other laws to do so. Because...

> it's not as if they had no idea the children endlessly posting selfies and posting "six seven" on their service weren't adults.

...you can make statements like that which sound like common sense, but it would be incredibly hard to regulate based on "if you know, you know" (or "you should have known"/"you had to have known"). The law has to provide (guarantee) a way for them to know in order to actually require them to take action based on it.

> As a parent, I might think that my 16 year old should be allowed to look up information on STDs but the websites that collect my child's age could decide they can't

This is a different problem. It sounds like you're essentially wanting to guarantee access to certain things, not just for your own 16-year-old, but for everyone else's, too (because if it was just yours, you could look it up for/with them if necessary). It'd be difficult to compel businesses to provide services to audiences they don't want to. But again, that's a separate problem that doesn't necessarily conflict with the rest of the system.

> No, but it's a framework that would allow other laws to do so.

I worry that's it's the start of a lot of "other laws" which will limit the ability for children and adult's to maintain even pseudo-anonymity online.

> The law has to provide (guarantee) a way for them to know in order to actually require them to take action based on it.

That sounds like an argument for even stronger proof of age than what the law calls for. Online platforms should do what nearly every other publisher does and provide a rating for their content. Netflix doesn't need to know how old I am. They provide a "kids" profile populated with their own curated content if that's the kind of thing I want and for everything else they provide ratings (PG, R, TV-14, etc.) It would be easy enough to push a rating to clients, they could even use HTTP headers for it. If lawmakers really felt the need to interfere in all of our operating systems it could require some means to collect and act on those ratings.

> It'd be difficult to compel businesses to provide services to audiences they don't want to.

This is the norm. It's what every business does apart from those who demand ID for every transaction. It's useful for businesses to give people their opinion or intention for who they're targeting, but it's entirely inappropriate for every website and online service to force their opinion onto others. They aren't qualified to know what's appropriate for a specific child and platforms like facebook have repeatedly demonstrated that they absolutely can't be trusted to put our children's interests above their own.