Shipped my first app today without much validation, did I make a mistake?

You asked for blunt ;), so: two of your three questions are premature and the third is the only one that matters.
the free-tier question (3 lessons/month, too stingy or generous?) is a question you can only answer with users, and you have zero. tuning pricing before anyone's converted is optimising a variable that has no signal behind it yet — park it entirely.
the "did I make a mistake shipping without validation" question is also moot now. you shipped. the mistake, if there was one, is already made or already fine; either way it's behind you and re-litigating it doesn't move anything.

which leaves the real one, and SystemicThinker aimed right at it: is "bored with generic content" a pain people act on, or one they complain about and keep using Duolingo anyway? I'd sharpen it further — the tell isn't whether people find your idea appealing (everyone will nod, "learn from videos you like" sounds nice), it's whether anyone has an existing behaviour you're replacing. do they already paste transcripts into ChatGPT to make flashcards? already screenshot subtitles? if there's a manual workaround people are already doing, that's real pain and you're productising it. if there's no workaround — if people are just mildly bored but coping — then "more interesting content" is a vitamin, and vitamins don't pull switches.

So the one thing I'd do this week: find five people who currently use a language app and ask what they do outside it to stay engaged. not "would you use mine." what are they already hacking together. that answer tells you if you built a painkiller or a nicer vitamin.

Launching without validation isn't necessarily an error. It depends on what you do next.
Your first question is the right one to sit with. There's a difference between "people are tired of generic content" and "people will actively switch to something new because of it." Duolingo fatigue is real, but a lot of people complain about it and keep opening it every day anyway.
Before you touch your pricing or free tier, it might be worth finding out if anyone has actually stopped using an existing app specifically because of the generic content problem. Not someone who agrees it's annoying, but someone who left because of it. That answer tells you a lot more than any freemium structure will.

You didn't make a mistake, but you might be misdiagnosing your own pain point, and that's the thing to fix first.

"Boredom with generic content" is probably not the real driver. People don't quit Duolingo because it's boring, they quit because they don't have a strong enough reason to keep going. The ones who stick have a concrete why: a partner who speaks it, a move abroad, an exam. Boredom is the symptom they name, weak motivation is the cause. So "learn from videos you like" makes it more fun without fixing retention, because fun isn't the missing ingredient, stakes are.

But there's a sharper version hiding in your idea. The people who'd love "learn from a video you'd watch anyway" aren't bored beginners. They're intermediate learners who already consume content in the target language (K-dramas, anime, Spanish YouTubers) and want to turn passive watching into active learning. Real, specific, underserved. They've outgrown Duolingo, not quit it. That reframe changes everything: not "less boring than Duolingo" but "turn the content you already watch into lessons."

On the free tier: 3 lessons/month is too stingy to prove value. The magic moment is "I turned a video I loved into a lesson and it worked." People need to feel that a few times before paying. Let them hit the wall after they're hooked, not before.

Who did you build this for, a beginner or someone already watching content in the language? That decides your positioning, pricing, and where you find the first 100 users.

This is a real direction, but the key question isn’t whether boredom with generic language apps exists—it clearly does. The harder question is whether “learning from content you already watch” creates a repeatable habit loop, or just a novelty effect. Retention usually depends on whether the learning moment is tied back into active recall, not just passive enjoyment.

I think shipping first is the right call here. The honest truth: most language apps ARE boring and generic. The real question isn't whether the problem exists - it's whether enough people care enough to pay. You're now in the best position to find out. Track which videos resonate most, watch user drop-off patterns, and iterate. Three lessons/month is actually tight - most users will hit that in 2-3 days. Worth experimenting with the free tier size based on early usage data. Good move getting it live.