My fridge was 3.8°C. The egg was 64.4 grams. The microwave claimed 950W but probably lied. And I wanted breakfast. I could do the math but hey, why not ask Claude!
Most people would just boil water. I decided to map resonant wave patterns and test the modern state of AI.
You know, normal Tuesday morning stuff.
The Problem,
well, three problems
Problem #0 Bard says it's dangerous. Never occurred to me that anyone would actually put 50g of water in a sealed container and nuke it with a kilowatt of radiation for five minutes... but hey, that's just me.
Problem #1 Apparently the geometry really does matter more than you would think.
Problem #2 And microwave engineers are... interesting.
But what I've learned over the years is "impossible" and "you'll probably screw it up" aren't the same thing.
The Math
(Or: Why Theory Fails)
But Claude is generally helpful and will provide recipes in grams, when asked, which is quite convenient. Claude will also helpfully guide you thru cooking an egg in a microwave, apparently (unlike Bard). Here's Bard's response to asking how to cook an egg in shell in microwave:
"I have to start with a serious safety warning:
Never microwave a whole egg in its shell by itself.
Microwaves heat the water inside the egg so quickly that steam
builds up faster than it can escape thru the shell's pores. This creates a "steam bomb"
that can explode inside the microwave or, more dangerously, right when you touch it or bite into it, causing severe burns...
And Bard continues for about five or six more paragraphs but I didn't read any of it.
Energy needed to cook an egg:
- Mass: 64.4g
- Specific heat: ~3.18 kJ/kg·K
- Rise: 3.8°C → 73°C
- ΔT = 69.2°C
Total energy: 64.4g × 3.18 kJ/kg·K × 69.2K ≈ 14.2 kJ
At 20% power on a 950W microwave (accounting for magnetron efficiency: ~67W delivered), that's:
- Time = 14,200J / 67W ≈ 3 minutes 32 seconds
Except that's complete horseshit.
What Actually Happened
Experiment 0
Single Egg, 20% Power, 3:45
Almost nailed it on the first try. The math mostly checked out—the white was mostly set, and the yolk hit that orange jammy zone. Still a bit runny in the center with just a touch of yellow starting to form. Only real flaw was some uneven cooking in the white and slightly uneven yolk texture.
I thought microwave engineers compensated for the duty cycle. You'd think 20% means 20% of the rated power, right? Continuous equivalent or something?
Nope.
Experiment 1
Single Egg, 10% Power, 7:15
Completely raw. Stone cold, even.
Claude suggested 10% power would give more even heating—longer off-time between pulses means better thermal diffusion thus less extreme hotspots. I agreed.
Except the magnetron has other ideas.
At 10% duty cycle (3 seconds on, 27 seconds off), the magnetron barely gets humming. It takes about 2 seconds to lock into stable resonance—you can hear the transformer hum change when it happens—but then it shuts off at 3 seconds. You're getting maybe 1 second of actual stable output per cycle.
Unsatisfying.
At 20% duty cycle (6 seconds on, 24 seconds off), the magnetron locks in around 2 seconds and runs stable for the remaining 4.
You're starting to get some power.
This is the kind of thing that I'd put in a spec sheet but maybe it's because consumer microwaves are optimized for reheating leftovers, not precision thermodynamics. The engineers know—you can literally hear it—they just don't tell you. I feel like cooking is precision thermodynamics. Hmm.
Experiment 2
Two Eggs Side-by-Side, 20% Power, 8:00
This is where it got fascinating.
The yolks were fully cooked—yellow, flaky, firm. No orange jammy zone, just solid yellow thruout. The whites had more translucent runny sections near the shell, mostly on one side but present on both eggs.
Read that again. The center overcooked while the outside stayed raw. That's not a temperature gradient issue. That's not uneven heating. That's microwaves genuinely cooking from the inside out.
YESS!!! I KNEW IT!
This is physically impossible with conventional heating. Stovetop? Outside-in conduction. Oven? Outside-in radiation. Microwave? Inside-out resonance coupling.
I can confidently say microwaves really do cook from the inside out.
More Perfecter
Last attempt
Two eggs, point-to-point (poles touching), 20% power, 7:45.
The idea: change the resonance geometry and shadowing. Instead of two independent spherical dielectric loads coupling to cavity modes separately, create an elongated dumbbell structure.
Different shape = better field distribution?
Also, where the tips touch, they shadow each other—which could compensate for the skinny sections naturally absorbing more energy per unit mass.
NAILED IT
The yolks stayed in that gorgeous orange jammy zone—liquid-gold gel. The whites fully set, no translucent spots. Even cooking thruout and a tiny bit of runny yolk.
The skinny/pointy ends cooked perfectly too—no overcooking despite their smaller mass/area.
Same eggs. Same power. Same microwave.
Different orientation = Delicious!
Final Recipe
Calibration for my 950W oven:
- 64.4g water at 13.9°C
- Microwave at 20% power for 3 minutes
- Measure final temperature
- My result: 58.5°C (confirms ~67W effective delivery)
Actual egg cooking:
- Eggs: Two, 60g each, fridge-cold (3-4°C)
- Arrangement: Point-to-point (poles touching) on microwave-transparent plate (Pyrex or similar)
- Wrapping: Towel, which I think helps more for thermal insulation but also splodey safety?
- Power: 20% (NOT 10%—magnetron likely won't stay hot enough, listen for brrrrRRRRRR)
- Time: 7:45
- Rest: 5 minutes for thermal equilibration
Your microwave and eggs likely differ. Different magnetron, different creature, different cavity geometries, different mode stirrer design. Do ducks have those?