It appears that many things that I thought were normal to put into an html page are not.
Those are normal things, but some companies like to deviate, often just to be different. For example, Google uses a Unicode lightning bolt as its DOCTYPE for Amp pages.
However, saving 100 bytes on a page load can end up being actual money saved when you shove out as many pages as Twitter does in a month.
Saving bytes makes sense. But how does the rendering engine know that those lines are CSS instead of junk text? And without a head tag, how do those lines not end up as text in the body of the document? And isn't it strange that it doesn't have a title tag (though, if you aren't going to have a head section, maybe skipping the title tag makes sense too).
Those are normal things, but some companies like to deviate, often just to be different. For example, Google uses a Unicode lightning bolt as its DOCTYPE for Amp pages.
However, saving 100 bytes on a page load can end up being actual money saved when you shove out as many pages as Twitter does in a month.