This week, a Reakh AI tutorial went viral on Juejin (掘金): a physics teacher wrote zero lines of code, described what they wanted in plain language, and within minutes produced an interactive convex-lens imaging courseware with drag-and-drop controls and tunable parameters. We see the real signal here not as AI entering the classroom, but as the barrier to "building interactive tools" being pulled down to a level any ordinary person can use.

What this is

Reakh AI (reakh.com) is a tool that generates interactive content from natural language. Teachers don't need programming or web-drawing skills — they describe in everyday language "what students should observe, what they should change, what they should verify," and the AI produces a demo page that runs in a browser with drag interactions, parameter adjustments, and real-time feedback. The convex-lens imaging demo in the tutorial showcases: dragging objects to change object distance, adjusting focal length and object height, five preset scenarios (camera, projector, magnifying glass, equal-size image, focal point), autoplay, and randomized question generation. The full workflow supports preview, conversational editing, AI self-checking of knowledge accuracy, and outputs a shareable online link ready for classroom use.

Industry view

Supporters argue this kind of tool's value lies in decoupling the quality ceiling of courseware from a teacher's software skills — what used to require professional developers to bring a key high school physics teacher's ideas to life can now produce a prototype in under an hour.But we see at least two risks that can't be ignored. First, AI-generated physics simulations easily break at edge cases: the tutorial author themselves hit visual artifacts and out-of-bounds color blocks when objects approached the focal point, only fixing them through multiple rounds of dialogue. A single sign error in a physics conclusion misleads students — qualified teachers must do the final review. Second, once "teacher contributes ideas, AI produces courseware" runs smoothly, it easily slides into "teachers only describe, AI decides how to teach" — long-term, this is outsourcing pedagogical judgment to models.

Impact on regular people

For enterprise IT: school and tutoring-industry budgets may shift from "employing an engineer who knows courseware development" toward "subscribing to AI tools + training teachers who can use AI," but in the short term it's more likely a complement than a replacement.For individual careers: similar to how teachers can produce courseware with one sentence, operations, product, and sales teams can use comparable tools to turn workflows, rules, and data into clickable demos (interactive sample demonstrations). This means "being able to build interactive things with AI" is becoming a new universal office skill — those who don't will increasingly be at a disadvantage.For consumer markets: the most immediate change for parents and students will be more free or low-priced interactive courseware; but whether the content is correct and aligned with curriculum standards can only be vetted by qualified teachers — tools won't solve this automatically.