Stephen Krashen is a linguist who's research has yielded the understanding that language is only truly learned, or 'acquired' as he says, through the brain's parsing of input.
If when you try to speak another language it is first filtered in your brain, if you need to take the lag-time to translate from your native to the target language, then your brain hasn't truly acquired the language.
A language is acquired at least in some capacity when you can formulate thoughts in the target language without having to first formulate the thought in your own native language.
And you can only attain this state of having acquired a language whenever your brain has had enough input in the target language.
When you learn on Duolingo, you're learning to become literate
, you're learning how to read and write, when you've not even become fluent in the target language.
People can speak their language perfectly fine without being literate they're whole damn lives, and people in the past certainly learned how to speak other languages without any reading, writing, or even common linguistic ground.
Now in the modern era it sucks to be illiterate, so anyone learning should still learn to be literate in their target language, but they should understand that literacy does not necessarily equate to fluency.
American language classes teach us nothing but literacy, and we're never exposed to any content of people actually speaking Spanish or French, thus even when burgers actually have language class they still never become fluent.
The same sort of thing goes for the Japanese, their classes focus wholly on English fluency but they never absorb any content in English or have someone really speak complete English expressions to them.
Now for the Scandies on the other hand, most content does not get localized into their languages because it's not profitable, and their languages are already very close to English, so not only do they have excellent literacy classes, they also grow up constantly exposed to English, and that's why I'd honestly consider them to be native speakers at this point.
They actually get the INPUT needed to acquire the English language, and lots of genZ nerds around the world spend so much of their time on English YouTube while also having classes, so they get the input they need as well and that's why so many have pristine English.
So classes help, but INPUT is the adhesive which brings everything together; the brain is a muscle and language is a muscular action.
Stephen Krashen's Input Hypothesis has a lot of scientific data behind it, and that's my summary of it, but you can look into it as well if it sounds interesting.