← All posts

You don't have to quit Duolingo

· Ben Ward

Every few weeks someone asks us some version of the same question: should I stop using Duolingo and switch to this?

No. Keep the streak. This is not that kind of pitch.

Here's the honest version instead. Almost every way people learn languages is good at what it does, and almost none of them make you speak. Not because they're badly made, but because making you speak is a genuinely different job. That job is the whole reason ConvoLive exists, and it fits next to whatever you're already doing.

What your current method is actually doing

Duolingo is the best habit machine ever built for language learning. It gets you to show up every day, and showing up every day matters more than almost anything. It builds vocabulary and it makes grammar patterns familiar. What it doesn't do is put you in a conversation. Its speaking exercises show you a sentence and ask you to read it aloud. Reading aloud is pronunciation practice. It is not coming up with something to say.

Anki and flashcards are unbeatable for retention. If you want three thousand words to stay in your head, spaced repetition is how. But a flashcard hands you the prompt. Conversation makes you find the word with no card in sight, in a couple of seconds, while someone waits.

Classes give you structure, a teacher who can explain why, and other humans. But do the arithmetic on a class of ten. An hour a week, minus the teacher talking, minus everyone else's turns, leaves you maybe two or three minutes of actually speaking. And most of that is answering a question you saw coming while the person before you answered it. A teacher can't fix this. There's one of them and ten of you. It's not a quality problem, it's a time problem.

Shows, podcasts and YouTube build comprehension, which is half of every conversation you'll ever have. The other half, producing the language yourself, gets exactly zero practice from any amount of watching.

Tutors and language partners do give you real conversation, which is why they work. They also cost money per hour, or cost scheduling, or cost the awkwardness of being a stranger's unpaid homework. Most people who intend to do a weekly session do three of them.

Notice the pattern. Everything on this list builds knowledge about the language or ability to receive it. Speaking is the one skill where you produce it, out loud, under time pressure. It's not covered by accident because it can't be covered by accident. It has to be the thing itself.

Speaking is a rep, and you're not getting reps

When someone asks you a question, you have about two seconds. In those two seconds you decode what they said, decide what you think, retrieve the words, order them, and say them. That whole loop is one rep.

A serious learner with a streak, a deck, and a Tuesday class can easily go a full week on zero reps. Then a real conversation happens, the loop is slow because it's never been run, and they conclude they're bad at languages. They're not. They've just never practised the one thing they were judged on.

Where ConvoLive fits

ConvoLive is deliberately just the speaking part. You open it, you're in a situation, at a café, at the airport, meeting someone new, and you talk your way through it with an AI partner who responds to what you actually said. Stuck? You can see suggestions for what to say next, or ask how to say something in your own language, and then say it. Every lesson is a conversation, so ten minutes in the app is ten minutes of reps, not ten minutes of tapping.

It's free, so the experiment costs you nothing: keep everything you're doing, and add one conversation a day.

Your Duolingo streak teaches you words. Your deck keeps them in your head. Your class explains the grammar. ConvoLive is where you find out you can actually say it, before the moment when it counts.