103 | J. Kenji López-Alt on Cooking As and With Science
What this covers
Kenji López-Alt, an MIT-trained chef and author of The Food Lab, joins Sean Carroll to explore how science illuminates cooking without constraining it. The conversation weaves together chemistry, physics, and biology with psychology and culture, arguing that understanding the mechanisms behind cooking—heat transfer, the [Maillard reaction](jargon:Maillard reaction), [starch retrogradation](jargon:starch retrogradation)—liberates rather than limits a cook's choices. López-Alt distinguishes between following recipes as turn-by-turn directions and learning the underlying principles that function as a map: the former gets you to one destination reliably; the latter lets you adapt to your ingredients, tools, and intentions. This framework extends to sous vide, pizza steel conduction, and convection ovens, where precise understanding of energy transfer produces outcomes previously impossible.
The conversation moves beyond mechanics into Taste as Brain Synthesis—the recognition that flavor is constructed from sensory input plus memory, mood, and cultural context rather than residing solely in the tongue. López-Alt challenges common cooking myths, showing that many persist because techniques optimized for restaurants get applied uncritically to home kitchens, and that blind taste tests miss how people actually eat, surrounded by visual cues, stories, and personal association. He explores Flavor Synergy and Flavor Balance Across a Meal, explaining why Thai cuisine balances sweet, salty, sour, and hot across courses rather than dishes, and how Acquired Taste shifts preference as bitterness receptors train the palate. The discussion includes practical ground—salt and acid as the keystones of seasoning, why stew-the-next-day tastes less different than folklore suggests, and how to learn a new dish by understanding its history and testing multiple recipes. Throughout, López-Alt resists prescription while his audience gravitates toward it, a tension he connects to personality and Tightness vs Looseness.
López-Alt argues that understanding the science behind cooking is liberating rather than constraining: technique lets you express culinary intent, but taste itself is constructed in the brain from culture, memory, and context, so there is no single 'correct' way to cook or universal pairing.
- Science is a process of understanding mechanisms (like heat transfer and the Maillard reaction) that frees a cook to adapt rather than follow rigid recipes
- Taste is synthesized in the brain from sensory input plus memory, mood, and cultural bias, so blind tests are limited and personal/contextual value is legitimate
- Cooking myths persist because techniques developed in one context (restaurants, authority figures) get applied uncritically to different situations
Understanding the chemistry and physics of cooking—heat transfer, Maillard reactions, starch retrogradation—enables cooks to adapt techniques to their own goals.
- Bread staling is not mainly drying but starch retrograding from a gelled form into a crystalline structure as it cools; with wheat flour this recrystallization is partly reversible (reheating softens bread), but corn starch undergoes irreversible retrogradation, which is why a corn tortilla is best fresh and never regains its elasticity after cooling—and why tacos in Mexico, where tortillas are made fresh or kept hot, are better than reheated supermarket tortillas in the US.
“staling is not just about drying, it's actually starch retrograding from a gelled form into a crystalline structure, so it sort of re-crystallizes as it cools down. With wheat flour, that recrystallization is reversible, well, partly reversible.”
- Any food containing protein and carbohydrate that is heated at high heat undergoes the Maillard reaction—a complex, not-fully-mapped cascade of browning reactions—which adds hundreds of new chemical compounds, giving more sweetness, savoriness, and flavor complexity; the complexity is not merely subjective because there literally are more chemical compounds present.
“Both steak and asparagus and bread and anything that you heat up with really high heat that contains protein and carbohydrate will undergo the Maillard reaction, which is the classic browning reaction”
- A pizza steel produces a crispier crust than a pizza stone because steel has a higher volumetric heat capacity (holds more energy per unit thickness/mass) and is far more conductive, so at the same oven temperature it transfers heat to the pizza base at a much higher rate via conduction.
“steel is much denser, so it holds more energy. It has a higher volume metric heat capacity, so it holds more energy per unit, per thickness per unit of mass than stone does. It is also much more conductive”
- Sous vide was enabled by higher capability rather than new scientific understanding of heat transfer—the technology to precisely control a water bath (and now air, via combi ovens accurate to a tenth of a degree)—so it is fundamentally a control advance, not a discovery about how heat works.
“it's inspired by higher capability. We now have control. It's a technological breakthrough. I don't think it's a breakthrough in scientific understanding of heat transfer.”
- Hot air right around food transfers its energy and then cools and stagnates, so a convection oven—essentially an oven with a heatproof fan—continuously cycles fresh hot air past the food, increasing the efficiency of heat transfer, the same principle a traditional pizza oven achieves through a cyclone of air drawn in the mouth, over the fire, and up the chimney.
“By fanning the air around and constantly cycling it, you create convection currents, and that increases the efficiency of heat transfer.”
- Most chemical reactions in cooking with heat are irreversible—there is an arrow of time in cooking—though some are reversible; you can push the food in one direction but generally cannot undo the transformation.
“There is an arrow of time with cooking. And there are some things that are reversible and some things that are not, but for the most part, cooking reactions are irreversible.”
- Technically, cooking is applying heat and triggering a series of chemical and physical reactions in raw ingredients to transform them into something more desirable to eat, with edge cases like ceviche (coagulating proteins via acid without heat) and combining ingredients (like dipping fish in soy sauce) also falling under cooking.
“cooking is applying heat and chemical reactions to raw ingredients to transform them into something that is more desirable to eat. That's what cooking is.”
Taste perception is constructed in the brain from sensory, emotional, and contextual factors, not just physical sensation on the tongue.
- When eggs are color-coded so people can't distinguish them, perceived quality differences disappear; conversely, adding a drop of orange food coloring to scrambled eggs makes their deep orange color trick people into thinking they are wonderful eggs—demonstrating that the visual cue of yolk color directly shapes taste perception.
“if you want to give people the best brunch ever, you just add like a drop of orange food coloring to your scrambled eggs before you cook them... that deep orange color will make them think, 'Oh, these must be wonderful eggs.'”
- In fully blind tests nobody can distinguish a backyard chicken egg from a factory-farm egg, but people never eat eggs blind—they see the package, know the origin, picture the chicken's life and animal welfare, and see the yolk color—and all those factors genuinely affect perceived flavor, so it is fine and even beneficial that ethically-raised eggs taste better to you.
“there's nobody in the world who can tell the difference between an egg that came from a chicken in their backyard and that came from a factory farm. In a completely blind setting, you can't tell the difference. But you never eat those eggs completely blind.”
- Perception of taste is not just the physical interactions on the tongue; the brain synthesizes smell, taste, sight, sound, feel and texture together with memory, current mood, personal connection to a food, and conscious and unconscious biases—so taste is created in the brain, making purely blind taste tests extraordinarily limited because nobody eats blind in real life.
“our perception of taste is not just about what we're perceiving on our tongue... Taste is something that you synthesize in your brain.”
- How we balance and value different flavors is largely cultural and learned: children dislike bitter flavors (plausibly because many natural poisons are bitter), and novices pick the sweetest chocolate or least bitter olive oil, but experts prize balance and astringency because once bitterness receptors are stimulated they make us salivate more, bringing out other flavors—so preference shifts with experience, as when a wine market that starts with sweet wines matures toward more complex ones.
“a lot of those flavors and the way we balance flavors and the way we value those different flavors is largely cultural.”
- The more context and story a food has—personal or global, its history and production—the more interesting it becomes to eat, analogous to how a Beatles song is better in the context of the album and era; but one person's idea of a valuable story (a hand-raised named cow) is not objectively superior to another's (a lab-grown steak), so one should reserve judgment about what counts as an interesting food story.
“the more context food has, and the more of a story it has, and the more you can tell about its history and production, the more interesting it gets”
- Computer-derived 'unexpected' flavor pairings (e.g. broccoli and chocolate) are dubious because it is impossible to divorce the cultural, historical, and personal elements of flavor from the chemical interactions: the mere knowledge that a computer chose two foods changes how they taste—for better or worse—so this branch of molecular gastronomy ignores that taste is created in the head, not purely sensed.
“it's completely impossible to divorce the cultural and historical and personal elements of flavor pairings in food from the actual interactions, and the actual chemical, physical interactions that they have”
Science and intuition in cooking reinforce rather than oppose each other; understanding mechanisms gives cooks the map, not just turn-by-turn directions.
- Science and emotion in cooking are not a dichotomy or trade-off; you can have high science and high heart simultaneously, and they actually reinforce each other because better understanding of how food works lets you more easily express what you intend, just as understanding color or an instrument lets an artist express what's in their head.
“I think the two are orthogonal. You can have something that is high science and high heart without the two competing. I don't think it's a dichotomy really.”
- You don't need to understand the science of cooking to cook well, but understanding it is liberating: the more you understand a process, the more able you are to adapt to a specific situation and make ingredients 'bend to your will,' so the goal is to give cooks the map rather than just turn-by-turn directions.
“the more you understand something, the more able you are to adapt to a specific situation, and the more able you are to make your ingredients bend to your will”
- A precise recipe is like GPS turn-by-turn directions: it reliably gets you from point A to point B but you learn nothing of the surroundings, whereas teaching the underlying science and technique gives you the whole map, so you can start from different ingredients or tools, take different routes, or reach different destinations.
“A recipe is sort of like a term. It's like going onto Google and saying, 'Hey take me to the grocery store'... and that's what following a recipe precisely is.”
- At its core, science treats what we think are facts as facts only until contrary evidence shows they aren't; the whole point of science is to question facts and not accept what someone told you unless it can be proven, so doubting facts is central to what science is.
“at the core of science is that what we think are facts are only facts until contrary evidence shows that they aren't. The whole point of science is to question facts”
Flavor pairing and preference are learned and cultural; techniques optimized for one context should not be uncritically applied to others.
- There is no such thing as foods that universally go well together or a universally better way to cook a steak, hamburger, or pizza; flavor pairing results are highly personal, depending on your taste, upbringing, culture, and current environment, so if a combination sounds delicious to you it will probably be delicious to you (given good execution).
“There's no such thing as like these things universally go well together... It's like it all has do with your personal taste and how you grew up and your culture and everything”
- Cooking myths persist because techniques are passed down by authority (whether a Michelin chef or a great-grandmother) without testing, and because techniques optimized for one context—a high-stress, low-margin restaurant cooking 40 eggs at a time, or motivating a line cook—get uncritically applied to a different context like cooking two eggs for your family; you should always ask what parameters and setting a technique was developed under and whether they apply to you.
“If a 12-star Michelin chef tells you this is the way it is, then by the power vested in their authority, that must be the way it is. But you're not taking into account the fact that... Maybe that works in a restaurant setting... but it's not the same as when you're cooking two eggs at a time for your family.”
- To learn a new dish, start with its history and cultural/historical context to understand what makes it what it is, then research five or six different recipes, compare where techniques differ and why (e.g. the Spanish tortilla onion debate), place the dish in the context of your own life to decide what you want from it, and if you have time try it multiple ways—'practice and be smart.'
“First of all, I start with the history of a dish, actually... And then from there, I'll research a bunch of different recipes.”
- When designing a restaurant dish, the technical elements come from experience and operational testing (cooking 120 schnitzels a night differs entirely from one at home), but more important is making sure the dish fits the restaurant's theme and pays its cultural dues to the cuisines it borrows from, requiring rigorous research to avoid cultural exploitation—as with their German-Korean Spaetzle-Kimchi dish modeled on tteok-bokki.
“it's one thing to be able to cook one really great chicken schnitzel at home, and it's a very different thing to be able to do 120 of them a night”
Precise temperature control via sous vide and salt and acid balance are the technical foundations that unlock reliable, foolproof cooking.
- Traditional cooking applies heat far above the target serving temperature, creating a narrow window between perfectly cooked and overcooked; sous vide sets the water bath to the exact target temperature (e.g. 127°F for medium-rare steak), so the food cannot exceed that temperature, giving a large margin of error and foolproofing doneness even for a novice cook.
“with traditional cooking, there's a sort of balancing act... There's a very fine window of time between perfectly cooked and overcooked, because the whole time I'm cooking it, I'm cooking at a higher temperature than I'm aiming for”
- Holding tough cuts of meat (pork belly, beef chuck, brisket) at a low precise temperature for long periods—e.g. bacon at 145°F for 48 hours—tenderizes them while keeping moisture loss minimal, allowing tough cuts to be served at the juiciness level of a medium-rare cut, something not previously possible.
“what happens then is that the meat is tenderized, but the moisture loss is very, very minimal. And so you're able to sort of get... You're able to basically cook tougher cuts of meat... and tenderize them, but still serve them at the juiciness level that sort of a medium rare or a medium piece of meat would be”
- Getting the salt level right is the single most important element of seasoning in essentially every culture—without it food tastes bland—and after salt, getting the acid level right is most important, with many dishes benefiting from a hit of acid at the end, though not every dish needs added acid if it is served with something acidic.
“salt, I think, is in most cultures, in fact, I think in every culture, getting the salt level right is the most important element of properly seasoning a food. Without the right level of salt, things just taste bland. I think after that, getting the acid level right, is the most important thing.”
- Different flavors have synergistic effects on each other (bitterness makes you salivate and perceive sweetness more; acidity cuts richness), and cuisines like Thai often balance sweet, salty, sour and hot across a whole meal rather than within a single dish—many authentic dishes are purely bitter or purely hot—the same logic by which a hot dog pairs better with lemonade than milk, or wine's acidity cuts a rich ragù.
“if you actually travel around Thailand and eat a lot, you'll find that there are many dishes that don't contain all of those elements all together... But generally, they're eaten in meals where they have all the different elements to balance each other out.”
Tongue taste-map myths and assumptions about stew aging persist because techniques are passed down by authority without rigorous testing.
- In López-Alt's own (admittedly hard-to-control) testing, freshly made stew or chili versus chilled-and-reheated stew showed surprisingly little difference; the main difference is texture/thickening rather than flavor, and the two are hard to isolate, so the widely held belief that stew tastes better the next day is weaker than expected.
“in the test that I've done, there's actually not that much difference in a freshly made stew versus a chilled and reheated stew. The main difference I think comes down to texture, like thickening”
- The classic maps of the tongue assigning specific tastes to specific regions are largely disproven; taste receptors for all the basic tastes are distributed across the whole tongue, only slightly more concentrated in some areas, so the idea of tasting sweetness only at the tip or bitterness only at the back is false.
“There's old maps of the tongue where it's like, 'Oh, we taste sweetness here, we taste saltiness there.' Those are mostly bunk. There is a distribution of all these sensors all over our tongues”