Black Box
A dark unit sits on a bench with one red-guarded switch. Flip it and a core wakes up in a porthole, glowing and wobbling, with a hum that rises and falls. Gauges read in kilothrums. A green feed scrolls past in capital letters. It looks like a rocket test stand crossed with a brain scan, and none of it is real: every word is invented and nobody ever built one. But it is not random. The unit follows fixed hidden rules, the same every time, and the only way to find them is the way a scientist or a test engineer would: change one thing, watch, write down what happened, and do it again. Students record experiments, write each rule they think they found as a sentence, and the app judges that sentence against their own evidence, not against an answer key. What they make is an Operator's Manual. The test of the manual is the Dark Run, where every gauge goes dark and someone has to run the unit with nothing but the manual in front of them.
Using this with a class? Make a free class and add your students yourself: each gets a username and password, with no email or Google account needed. Or give them one join code. Using Google Classroom? Post the link there; with a free class, the link also puts each student in your class, so you see their work. mytekDev is also available from the Google Workspace Marketplace, so your school's Google admin can install it for everyone.A first lesson: Project the dark unit, flip PWR, and say nothing for two minutes. The room lists everything that changes and everything they could touch, with no guesses allowed yet. Then pairs get ten minutes to find one rule, and a rule only counts if they can show the experiment behind it.
Set up a free class Share to Google ClassroomWhat a student actually does
They flip the switch and the boot text scrolls. One lever, one gauge, a lamp that says UNSTABLE, a row of ten hold lights and a tag on the bench: Get it STABLE and keep it there. They push the lever up and the Thrum gauge climbs; too far and something goes wrong. Holding the core steady for long enough unbolts the first cover plate, and behind it is a frosted tube with a blue valve wheel and a Purge button. Four more bays follow, each one earned the same way, until the bench is full: an engine bell that glows, a web of flickering nodes with a trace like an EEG strip, rings that spin around the core under a strobe, and a dish with a waveform screen.
Then the real work starts. They open the notebook, press Start an experiment, move one control, wait, and stop. The chart recorder draws any four readouts over 30 seconds, 2 minutes or 10 minutes, with a dashed mark wherever a control changed, so a reading that moves three seconds after the control is plain to see. They write what they think they found as a sentence: when I move up the Thrummet lever, the Thrum goes up, right away. They link the experiments that back it up, and the card gets a verdict: Supported, Contradicted, Tangled if two controls moved at once, or Not enough if the test was too short, covered too little of the range, or never tried the control with the rest of the unit set differently.
Some rules are easy. Some are sneaky: a reading that answers late, a control that only works once something else is high enough, two readouts that always add up to the same number, a button that seems to do nothing. Finding those means testing across the whole range and under different conditions, which is the lesson. The cards print as an Operator's Manual with a section for each bay, a chart of the evidence under each rule, and a list of known hazards. The manual also travels as a short code. A classmate types it in and tries the Dark Run: every gauge, screen, light and sound goes dark, the porthole shutters, and they have to start the unit and get the rings to lock using only those rules. When it goes wrong, the manual was missing something, and the author goes back to fix it.
The notebook. A rule is a sentence built from dropdowns, linked to the student's own experiments. Tangled means more than one thing changed at once, so the experiment cannot say which one did it.
What they are learning while they play
Nothing here is presented as a lesson. This is what the activity is made of.
| What the student does | What it is |
|---|---|
| Lists what changes and what can be touched before touching anything | Observation before hypothesis. A scientist describes before explaining. The first minutes on the unit are deliberately bare, so there is time to notice the lamp, the hold lights and the feed before anyone has a theory to defend. |
| Moves one control, waits, and stops the recording | Controlled experiments and variables. Change one thing and hold the rest still, or you cannot say which change did it. The app enforces this: an experiment where two controls moved marks the rule Tangled. |
| Links an experiment to a rule card and gets Supported or Contradicted | Claim and evidence. A rule is a claim; the recorded experiment is the evidence. The verdict comes from the student's own data, so a rule that happens to be right still needs a fair test behind it. |
| Spots a readout that answers a few seconds after the control | Delay, or lag. Cause and effect are not always side by side in time. The chart recorder's dashed marks make a delay measurable, and a student who stops watching too soon writes the wrong rule. |
| Finds a control that only works when another reading is high enough | Thresholds and conditions. Some effects switch on above a certain level. A test run in the wrong conditions concludes the control does nothing, and the app answers that with Not enough: try it again with the rest of the unit set differently. |
| Proves a control does nothing, across its whole range | Absence of evidence. It did nothing and I did not test it there are different findings. A null result only counts when the control was tried over most of its range and under the conditions that could have mattered. |
| Treats the unit as a box with inputs and outputs | Black-box testing. Software testers and engineers study systems they cannot see inside: try an input, record the output, write down the behavior. The unit is exactly that, with no source code to peek at. |
| Writes a manual someone else runs with every gauge dark | Technical writing. Instructions are only as good as what a stranger can do with them. The Dark Run is the test, and the failures point at the sentence that needs fixing. |
| Reads the unit's feed and asks whether the words mean anything | Media literacy and technobabble. A real technical term has a definition that is not circular, can be looked up, and predicts something you can check. The unit's words fail all three on purpose, which makes it a safe place to learn the difference. |
How to tell whether it landed
Ask, and let them show you. A student who has understood it can answer these without help.
- Show me an experiment that was marked Tangled. What would you change to make it a fair test?
- You wrote that a control does nothing. Where in its range did you test it, and with the rest of the unit set how?
- Point to the place on the chart where your control changed, and to the place where the readout answered. How long was the gap?
- Your classmate's Dark Run failed with your manual. Which rule card was missing or wrong, and how do you know?
- Here is a sentence from a science article and one from the unit's feed. Which one is real, and what test did you use to decide?
Use it in your subject
Not a technology class? Then the app is how students show what they learned in your unit. More projects for every subject.
Science
Use the unit as the first week of a scientific method unit: students plan an investigation, record it, and defend each rule with data before they ever meet a real variable in a lab.
Math
Read the chart recorder as a graph: find the delay between a control change and the readout's answer in seconds, and find two readouts whose sum stays the same while each one moves.
English
Treat the Operator's Manual as a technical writing assignment with a real reader: a classmate runs the Dark Run from it, and the revision is driven by where that reader went wrong.
Computer science
Teach black-box testing before students meet unit tests: choose inputs, predict outputs, record results, and write the behavior down without ever seeing the code.
Media literacy
Put a line from a real technical manual, a line from the unit's feed and a line of movie-style technobabble side by side. Students sort them and explain the test they used.
Practical notes
- AgesAges 10 to 15, grades 5 to 9, with middle school at the center. Younger students find the obvious rules and bring the board online; older ones go after the sneaky rules, the Dark Run and the missions.
- TimeA minute to turn it on and find the first rule. The first bay opens in a few minutes of steady hands. A satisfying session is 45 minutes; finding every rule and passing the Dark Run with someone else's manual takes several class periods.
- EquipmentA computer or Chromebook with a modern browser and a mouse or trackpad. Sound adds the hum and can be switched off. On a machine without 3D graphics the core is drawn more simply and everything else works the same. Nothing to install.
For co-ops, microschools, and classrooms
Give the whole room the same unit and compare manuals, or let each student open their own unit from the crate. Every crate unit has the same parts wired differently, carries a serial like 4K-7Q2, and is checked by a solver before it is offered, so it can be finished, and nobody can copy a neighbor's answers. A shared serial puts the whole class on the same wiring. With free accounts each unit saves with its own notebook, experiments and mission times, and the missions (Hold it, Lock, the Dark Run and Cold start) give students who finish early something hard to do next.
- Settings has Open every bay now, for showing the full board to a group without earning it first.
- The notebook shows each manual's code with a Copy button; swapping codes is how the Dark Run is done with someone else's manual.
- A project holds up to six units, each with its own notebook.
Common questions
Is any of the science in Black Box real?
No, and the app says so. The machine, its parts and its units of measure are invented, and the app never borrows a real scientific term with a wrong meaning. What is real is the method: the unit follows fixed rules, the same every time, and students find them with controlled experiments and evidence.
How does it check a student's rules without just giving the answer?
It judges evidence, not answers. A rule card is linked to experiments the student recorded, and the verdict comes from those recordings: Supported, Contradicted, Tangled when more than one control moved, or Not enough when the test was too short or did not cover enough of the range. A correct rule with a poor test still comes back Not enough.
Can students copy each other's answers?
Not if each one opens their own unit from the crate. Every crate unit has the same parts with different wiring, so one student's rules are wrong for the next unit. For whole-class discussion, give everyone the same serial or use the standard BX-1.
Does it need an account?
No. The demo above is the whole app. A free account saves the units, notebooks and manuals so the work carries over between class periods.
Still have a question? Ask us, and a person will write back.
Where to go next
Start with Black Box
Trying it costs nothing and takes about five minutes. An account is what makes the work last.
Black Box is one of the making apps on the platform. The machine is fiction and says so; the investigating is not. It works best when an adult refuses to tell students the rules and asks instead what their evidence shows.
Page last reviewed November 2026.
Google, Google Classroom and Chromebook are trademarks of Google LLC.