I wasn't saying that you don't need input to get fluent. I am criticising the notion of 'comprehensible input' as advertised by 'Krashenites' and sold to enthusiastic language learners. CI is a grift.

The question is how to get to the point where you can get something/anything from input alone. So many 'autodidact' 'polyglots' on YouTube have a back catalog of videos telling you that they learn languages by exclusively watching and reading native materials from the beginning. Some have their own websites with the same advice. They're all lying.

The idea that input is needed is trite but it's pitched as a revelation. It's also sold as the only thing you need. Even native speakers need more than input. It takes native speakers 15+ years to learn their own language, with direct instruction from family and friends as well as grammar and vocabulary tuition in school. I'm a bit older than that and reasonably well educated and I still have to look things up and ask others to check my English. And that's after sufficient direct 'input', too.

That's a good point about 了. Nevertheless, a learner can rapidly increase the rate at which they will understand the finer subtitles of 了 with some explicit instruction. There is no benefit in guessing wrong for years until it all clicks. Or in watching videos hoping to figure out any single word when almost none of them make any sense at all.

This is likely true for the most common words in a wide range of languages. The most common words have hundreds of meanings each. Almost every other word, if not all of them, has several meanings. An example from Spanish is 'cabeza', meaning 'head'. Up until yesterday I only pictured a head on top of someone's shoulders when I read or heard that word. So when I read a sentence with 'cabeza nuclear tactical' I thought there were two options. One, someone had a tactical nuke for a head. Or two, someone got hit in the head with a tactical nuke. It took several passes for me to realise that in this context, it meant warhead. The tactical nuke itself was the 'head'. (I think!)

Rejecting explicit learning because words have multiple meanings that only make sense in context is in my view the wrong way to approach the problem. Anyone who has gone to the effort of learning their own native language may as well use it to give themselves a headstart. This works with regard to input, too, e.g. starting with international news rather than creative writing; because a big chunk of the vocab about international topics will be the same in every language (albeit written or pronounced differently). To be clear, I don't think you are advocating for ignoring some explicit instruction – only reiterating the need to do a bit of both. N which case you are not advocating for comprehensible 'input', even if you say 'input'.

I agree there are risks with freestyle speaking too early. There's fossilisation, too. Which is why it's important to practice speaking by copying from good examples. But the comprehensible input crowd advocate for not speaking at all until it just bubbles out with perfect structure and intonation. Maybe it would do that one day; but how long is a learner supposed to be willing to wait to find out that it takes a decade or never happens?