For Learners
Of course this is hard.
It's also doable.
Every adult who has seriously tried to learn a language knows: there is no shortcut, no trick, no app that does it for you. Anyone who tells you otherwise is selling something Shima is not selling.
What we can tell you is this: the effort is real, the method is right, and the payoff is real. This page is an honest account of what you're getting into — what learning a language actually requires, what Shima does and doesn't do, and what you'll need to bring. Read it once. Come back when you need it.
The language belongs to you. The app is just the infrastructure.
The real situation
You understand more than you can say. That's not a you problem.
If you've studied a language — in class, with an app, through music and TV and years of effort — you probably know the feeling. You understand the question. You know roughly what you want to say. And then there's this terrible pause while your brain scrambles for words that feel just out of reach.
This isn't a sign that you haven't learned enough. It's a sign that you've learned a different thing than speaking. Understanding a language and producing it automatically are two genuinely distinct skills, built in different parts of the brain, developed by different kinds of practice. You can get very good at one without developing the other. Most language learning — apps, classes, textbooks — builds comprehension. Almost none of it builds production.
Shima builds production. Specifically: the ability to retrieve and produce whole phrases quickly enough that you can think about what you're saying, not how to say it.
"Knowing a grammar rule and being able to use it under pressure are not the same thing. They are not even the same kind of knowledge."
DeKeyser, Skill Acquisition Theory — the research Shima is built on
When a native speaker says could you wait a moment, they don't assemble that phrase from parts. They retrieve it whole — a single mental unit, stored and ready. That's what fluency actually is: a library of stored multi-word chunks, retrieved fast enough that your conscious mind is free to deal with meaning. Building that library is what Shima is for.
It takes repetition. Real repetition, over real time. There is no version of this that doesn't require effort. But effort aimed at the right thing produces results that effort aimed at the wrong thing never will, no matter how long you try.
What you're working toward
The moment a hard sentence suddenly flows
There is a specific thing that happens when a sentence you've been drilling starts to automate. You've said it dozens of times and it's still felt like work — you can feel yourself reaching for the next word. And then one day, without quite noticing when it changed, it just comes. The sentence arrives whole. You don't construct it; you retrieve it.
This is a real neurological shift — conscious assembly giving way to automatic retrieval. When it happens for a sentence you care about, one you've actually needed and stumbled over, it feels like something clicking into place.
Shima's engine notices this before you do. When your response time drops — not gradually, but suddenly — the sentence begins to recede. You'll drill it less. It's becoming yours. Other sentences surface in its place.
What to watch for
When a sentence stops feeling hard, that's not luck. That's the thing working.
Notice it. Let yourself feel it. The first automatic sentence is the proof of concept. The second feels different than the first. By the tenth, you've stopped counting.
This is what production fluency actually feels like. Not "I memorised that." More like: I have that now. You can use it tired, mid-conversation, under pressure. That's the goal.
How your attention works
Ten present minutes beats forty-five distracted ones
Attention is finite and has a shape across the day. There are windows where your brain is genuinely ready to produce language under pressure, and windows where it isn't. Drilling during the wrong window is not just inefficient — it trains a slightly worse version of the habit.
You don't need a schedule. You need honest self-knowledge about when you're actually present versus when you're going through the motions. Most people who practise consistently don't practise on a pristine calendar. They practise in the pockets — commute, coffee queue, ten minutes before a meeting. That's real practice if you're genuinely there for it.
Shima has no session boundaries, no minimum time, no completion state. There is no "done." You drill, you stop, you come back. The stream is always mid-sentence. Nothing resets. Nothing is lost.
Permission to stop
Leaving mid-sentence is not abandoning your practice. It is the design.
The Zeigarnik effect — the psychological pull of an unfinished task — is what brings you back. The stream stays open so the pull stays real. You can't fall behind. You can't fail a stream. The deck is always there.
When to practise
Your day already has the right moments in it
You don't need to add time to your life. You need to redirect some of the time that's already leaking. The moments most people spend scrolling — commute, queue, the gap between one thing and the next — are exactly the right length and cognitive load for production drilling.
Different parts of the day suit different modes:
Sleep matters more than it sounds. Your brain consolidates procedural memory — the kind Shima builds — primarily during sleep. The gap between sessions isn't downtime. It's when the learning gets stored.
| When | What's happening | Best mode |
|---|---|---|
| Morning commute | Waking up. Alert but not yet overloaded. Movement helps encoding. | Walk mode |
| Mid-morning | Focus window. Best cognitive performance for most people. Seated drilling here compounds fastest. | Seated drill |
| Lunch / early afternoon | Post-meal dip. Attention softer. Ambient exposure works; demanding drilling less effective. | Resting card |
| Afternoon gap | Second wind for many people. Good for short, focused sessions. | Seated drill |
| Evening commute | Decompressing. Movement + language is a good combination here. | Walk mode |
| Before sleep | Memory consolidates during sleep. A short session here — genuinely present, not exhausted — is well-timed. | Seated drill |
Waking up. Alert but not yet overloaded. Movement helps encoding.
Focus window. Best cognitive performance for most people. Seated drilling here compounds fastest.
Post-meal dip. Attention softer. Ambient exposure works; demanding drilling less effective.
Second wind for many people. Good for short, focused sessions.
Decompressing. Movement + language is a good combination here.
Memory consolidates during sleep. A short session here — genuinely present, not exhausted — is well-timed.
Hands-free, eyes-free
Walk mode is for when you're already moving
Walk mode is hands-free, eyes-free production drilling through AirPods. The sentence plays in your ear. You say it. That's it. No screen, no tapping, no looking. Designed for the commute, the morning walk, the run, the grocery store. Anywhere your hands and eyes are busy but your mouth isn't.
It feels different from seated drilling. Your attention is partly distributed — navigating, moving, aware of your surroundings. That's fine. There's early evidence that movement during language drilling may enhance encoding. Walk mode was built for convenience; it may be doing more than that.
Expect more variation in your response times. Some sentences will feel harder to retrieve on the move. That variation isn't a problem — slightly more effortful retrieval may actually improve how robustly the learning sticks.
When to use it: commute, walk, exercise, household tasks. Not when you need to concentrate on something else. Walk mode works because you're genuinely present to it, not because you've technically ticked a session.
The real competition for your time
Same pocket, same phone, completely different direction
You don't need more time. You need to redirect time that's already leaking. The moments most people spend scrolling — commute, queue, the waiting-room minutes — are exactly the right length for production drilling.
Scrolling and drilling both live on your phone, both fit in a two-minute gap, both are things you do alone. The difference is what they leave behind. One leaves you roughly where you started. The other compounds.
The resting card on your home screen is part of this. It's not a notification. It's not guilt. It's the next sentence in your stream, sitting quietly on your home screen, available any time you reach for your phone. You tap or you don't. It's an invitation, not a demand.
Scrolling
Passive consumption
Designed to keep you opening
Leaves nothing behind
Optimised for the platform
Zero cost to stop
Shima
Active production
Designed for you to eventually leave
Each session compounds
Optimised for your outcomes
Always mid-sentence — pulls you back honestly
What consistent practice actually looks like
A real week. Not a perfect one.
Consistency doesn't mean a clean calendar. It doesn't mean the same time every day, or a minimum number of minutes, or an unbroken streak. Those are fantasies that make you feel bad when life gets in the way. Life always gets in the way.
Real consistency looks like presence when you're there. A week where you drilled on three commutes, twice before bed, once while making dinner, and not at all on Wednesday — that's a good week. The stream holds all of it. Nothing resets. Nothing is lost.
A real week — not the ideal week
This is not a failure week. This is what learning a language while living a life looks like. The stream held all of it. Next week builds on this one.
How the learning compounds
Bounded practice produces unbounded recognition
An island is 50–100 sentences in one domain. The train station. The clinic. The factory floor. The ward. It feels small when you start — almost artificially small. That's intentional.
When you drill 80 sentences in a single domain, something happens that doesn't happen with general vocabulary study. The sentences start to reinforce each other. The patterns repeat. The vocabulary recurs in different forms. Your brain is building not just isolated phrases but a web of related language — and the web, once built, starts catching things that were never in the island.
You'll start noticing station vocabulary in contexts you didn't drill. You'll catch grammar patterns you recognise from the island appearing in conversations, in signs, in things people say to you. The island doesn't stay an island. It becomes a lens.
The flywheel is: each island you complete makes the next one faster. You've learned how to learn in this language. You've built the first layer of the formulaic library. Each subsequent island adds to it faster because more of the new material is building on something you already have.
The other side of the relationship
Shima handles the method. You bring the practice.
Shima decides what you drill and when. It adjusts without asking. It notices what's automating before you do and moves on. That's its job.
Your job is to show up with genuine attention when you do show up. Not every day. Not for a set amount of time. But present when you're there — actually attempting the sentence, not pressing through the cards. Only present practice builds automaticity.
Show up irregularly. That's fine.
There is no streak to protect. Wednesday off, nothing Saturday, four sessions in a row Tuesday to Friday — the stream holds all of it without judgement. Just come back.
Show up distracted. That's not fine.
Half-present drilling is almost valueless and slightly counterproductive. If you're going through the motions, stop. Come back when you're there. Ten present minutes is worth more than forty distracted ones.
Stop when it becomes mechanical.
You'll feel the shift — when the session stops feeling like effort and starts feeling like clicking through. That's the signal. Stop. The next session will start fresh. Stopping well is part of the practice.
A realistic shape
What to expect, honestly
The early sessions will feel effortful. Most sentences will feel difficult. Response times will be slow. This is not a sign that it isn't working — it is what the beginning of proceduralisation looks like. The difficulty is the process, not an obstacle to it.
The first automatic sentence usually arrives sooner than learners expect. It won't feel dramatic. A sentence you've been struggling with will suddenly just come. You might not notice it the first time. You'll notice it when it happens again, and again.
The first island — 50–100 sentences drilled to the point where most of them feel owned — takes most learners between six and twelve weeks of consistent practice. That's not six weeks of daily full sessions. That's six weeks of being in the app regularly, at whatever pace your life allows.
By the end of the first island, you'll be a different kind of learner than you were at the start. Not because of what you know, but because of what you can do. The second island will confirm it. The third will feel like momentum.
Months, not weeks, for meaningful fluency. The language is a long project. Shima is one tool in that project — the drilling layer that makes everything else, class and conversation and daily life, actually work. When you have the language, you won't need the app. That's the point.
The invitation
Learn the language because you can.
Not because an app gamified you into it. Not because a streak made you anxious. Because you decided the language was worth the effort, and you found a method that makes the effort count.
The language belongs to you. The class, the conversation, the daily life you're preparing for — those are yours. Shima is just the infrastructure that gets you ready for them. When you don't need it anymore, that's the success condition, not the failure one.
The stream is always mid-sentence. It's ready when you are.