Important learnings from drinking wine and microwaves
I had started a post titled: “A field guide to AI Washing layoffs”, in response to the announcement that Jack Dorsey at Block is cutting almost half his workforce. But, I am happy to report that while I was dilly-dallying on whether I should even go through the effort of writing (usually a sign that I should not), I was happy to see consensus public opinion come down so firmly on my side on this, that it really just isn’t needed anymore. With characteristic English dryness, Benedict Evans had one of the most concise takes:
Here’s a pretty powerful chart to illustrate this point:
Public opinion on the Citrini story also has landed quite clearly on the side of economic sense. On that one I have found two more good counters since by previous post:
This one from Citadel, amusingly and snarkily titled the 2026 Global Intelligence Crisis.
A Chinese parody of the Citrini story (translation available one or two extra clicks deep).
So, with those major AI points laid to rest, I thought I’d write about wine and microwaves.
What the microwave can teach us about technology diffusion
I remember clearly when we got a microwave. It was a 1,000 watt beast of a ‘Combi Microwave’, meaning it also was an oven and a grill. It promised massive time savings without compromising on taste, for example cooking a whole chicken on a rotating spit in 30 minutes, instead of the usual 1.5 hours. When we got it, we put white bread with cheese in it. We literally called it “boterham met kaas uit de magnetron”, which means “bread with cheese from the microwave”. Not the catchiest name. You would put sliced cheese on a slice of bread, put it in the middle of the microwave, put it on the max 1000 watt setting for 10 seconds and watch in awe through the window as the cheese was obliterated by the invisible radiation and the bread turned into something resembling a hot towel. It wasn’t that tasty, but it was fine. And the mediocre eating experience was more than compensated by watching this futuristic piece of technology melt cheese in seconds.
But, as everyone now knows, that was about the most we’d ever get from that microwave. Yeah, you can defrost things and warm things up. We have one at home and it’s become completely normal technology. But I don’t recall ever using it to make a whole chicken. And its ability to cook chickens 3x faster than before somehow still didn’t turn a whole chicken into a viable weeknight meal.
It seems like the bottleneck really wasn’t the ability to cook things way faster. Cooking whole chickens is messy, they take up a lot of space in the fridge. There are many other dishes you can cook in 15 minutes even without a microwave. Turns out the hard part was never the cooking.
Reputational issues soon followed for the microwave. First, health concerns. How could we be sure that these invisible rays weren’t somehow in the food after microwaving it? And was it safe to even be around this device? Or maybe the waves changed something in the food, making it cause cancer. It is difficult to understand the differences between various kinds of invisible radiation, after all. And considering what it does to a cheese sandwich…..
On top of that, the microwave became tainted with a low class label. This was for time-poor, money-poor people. The microwave was great at reheating canned food, and it became associated with that use case. Foodies scoffed at the inability of the microwave to make anything crispy or brown, or tasty. The internet proclaimed “Microwave Cooking for One” to be the “saddest cookbook of all time”. People with class and money have the time and taste to SLOW COOK stuff over many hours, preferably with real burning wood.
The microwave is everywhere now. But it’s not cool, it didn’t revolutionize cooking, it didn’t save a whole lot of time, and it definitely didn’t replace any cooks.
What drinking wine can teach us about perception
The incredible documentary Sour Grapes tells the insane story of the world’s greatest wine counterfeiter of all time. Vintage french wines can be worth tens of thousands of dollars to rich collectors. Rudy Kurniawan was able to forge these kinds of wines and scam collectors for tens of millions of dollars by selling them fake wines. He would collect empty bottles, and fill them with carefully blended mixtures of cheaper wines. Then he would forge and age the labels, seals, age corks, and make the stuff look as legit as possible. His skills in the acts of forgery are as impressive as his salesmanship and ability to network his way into the highest echelons of wine collectors. But the most important thing to me, is that it is even possible to fool the entire wine establishment by blending a bunch of Napa Valley and cheap Burgundy wines in your kitchen sink. What does it say about the intrinsic value of these rare, coveted wines?
Of course, wine snobs can be absolutely insufferable. And they have been proven to be full of it over and over again.
A Stanford study showed that in a wine tasting, subjects preferred wine that was presented as $45 over a $5 wine, even if both bottles contained the exact same wine.
A statistician spent years secretly slipping duplicate wines into the California State Fair wine competition. Only 10% of judges gave the same wine a consistent score. Some rated the exact same bottle as Gold one time and No Award the next. Across 13 US competitions, 99% of gold medal wines received no award somewhere else.
In a French study, wine students were given a white wine dyed red and asked to describe it. They used red wine descriptors like "berry" and "leather." These students could not even tell red wine from white wine by taste.
Perception and narrative are critical factors in what gets preference. Sommeliers, snobs, and gullible tasters are everywhere, not only when we’re dealing with wine. A lot of what we call "expertise" or "quality" in knowledge work is actually wine tasting. It might matter less whether McKinsey generates the slides with AI or with Red Bull powered Associates, as long as real humans present the work to other humans. The value was always partially (in large part even?) the label on the bottle.