Another aspect of this issue is the apparent difference in intellectual preparedness of kids today. I think this is partly due to educational environment and partly a result of selection bias (in that accounts of the past focus on the "smart" kids). When I was a teenager, rummaging around the attic at grandma's house I stumbled upon my dad's old scientific stuff. I was shocked at the sophistication of what a high sch…
Actually-educational things for kids haven't gone away entirely, though they've certainly changed shape. There are a lot of kids and teenager related products around programming and microcontroller-related hobby projects that teach and require deeper understanding of the concepts. Building a radio, on the other hand, when AM/FM radio is now 90-100 years old instead of 20-30... that wouldn't interest me nearly as much…
you can definitely kit-build an am or fm radio from a kit without any understanding of the concepts, the radioshack '200-in-1' 'science fair' kit had wiring netlists for both am and fm radios, and didn't really explain the concepts. similarly you can solder together a microcontroller board and write circuitpython on it without any understanding of the concepts
in any of these cases there are a shitload of concepts to learn
i don't know much about chemistry but i think you have measurement error, valence electrons, vacuum filtration, density, oxidation states, precipitation, ph, solubility, activity, decantation, density, enthalpy, recrystallization, pressure, solubility product, never give an acid a drink, entropy, flasks bubbling over, gibbs free energy, hplc, equilibrium, distillation, the arrhenius law, sublimation, ligands, spectroscopy, differential scanning calorimetry, ...
in analog electronics you have measurement error, voltage, current, energy, capacitance, resistance, complex impedance, inductance, frequency, power, charge, parasitics, exponential decay, diffeqs, linear time-invariant systems, lti system transfer functions, convolution, the fourier transform, filtering, and only then do you get into nonlinear systems: rectification, inductive kick, transistor action, transistor switches, the ebers-moll model, miller capacitance, schmitt triggers, oscillators, etc. iirc that's just the first two chapters of horowitz and hill and there are thirteen more chapters and sixteen appendices. to be fair some of them are digital
in programming you have the whole programming iceberg at https://suricrasia.online/iceberg/. but even if you only get into the stuff that is actually involved in blinking an led on an esp32 as a microcontroller-related hobby project you have c, a compiler, a parser, context-free grammars, finite state machines, variables, lvalues vs. rvalues, subroutines, formal vs. actual parameters, arrays, hash tables, linked lists, pattern matching, optimizing transformations, denotational semantics, undefined behavior, structs, rtl, assembly language, machine code, symbol tables, linkers, register machines, propositional logic, de morgan's theorem, synchronous logic, metastable flip-flop states when crossing clock domains, jtag, debuggers, ram, flash, tcp/ip (tcp ip arp routing checksums ethernet), wi-fi, direct-sequence spread-spectrum communication, ofdm, shannon's theorem, csma/ca vs. csma/cd, pwm, etc
there's a shitload to learn but you can do stuff without knowing more than a tiny amount of it