show string "hi" block into on start. Watch the simulator on the left scroll "hi" — that's your micro:bit, in your browser, no cables required. Drop one radio.send number block in and watch a second simulator appear. That's the IoT we're building today. Then complete the pre-survey.radio group 7 every 2 seconds, aggregator receives and displays it, and your hand over the sensor proves the whole thing works. The key move here is using an LLM to generate the code, then toggling between Blocks and JavaScript to understand what it produced, pasting in suggestions, and learning how to prompt your way past what it gets wrong. Then you're on your own — swap the temperature sensor for light, acceleration, sound, or compass, build your own sensor/aggregator pair, and save two .hex files ready to flash to real hardware.PING, the sensor replies ACK, and a dead node warning fires if silence runs too long — then the MakeCode Data panel gets wired up so incoming readings plot as a live graph instead of just scrolling on the LED. After the demo, brainstorm new applications as a group: what becomes possible when your nodes talk back, and when you can see trends instead of just numbers? Then open build time — use the LLM to take your earlier pair further, whether that's adding thresholds, alerts, bidirectional commands, or a cleaner aggregation strategy. Document what you prompted, what it got right, and what you had to fix.radio.sendValue("id", 1) vs "id", 2). Update the Aggregator to track per-sender averages and flag senders that drop offline.radio set group — watch the failure mode. Apply Check the Machine: Task → Before (what I expected) → After (what I got) → Takeaway. Preview for when your kit arrives: the simulator's temperature is idealized; a real micro:bit's temperature sensor reads the CPU die, which runs 3–8°C hotter than ambient air. That's a real calibration problem — a perfect Fortify hook for your students on day one with hardware.