The sourcing sheet
Every "is expensive X worth it" verdict, as a buying decision. This is the single most restaurant-useful thing in the entire three-creator corpus: it tells you exactly where a guest can taste your spend and where they cannot.
| Input | Call | The verdict, and the cutoff |
|---|---|---|
| Eggs | BUY CHEAP | Caged, cage-free, free-range and pasture-raised were indistinguishable in blind taste tests (17¢ vs 67¢ per egg). Pay up only for ethics or yolk colour on a plated dish. ▶ |
| Chicken | SPLIT | $3/lb vs $9/lb: both bland plain, so no flavour premium. But water-chilling adds 5-12% absorbed water that expels during cooking and stops fried coating adhering. Buy air-chilled for the fried lane, cheap for saucy dishes. ▶ |
| Steak grade | SPEND (on fat) | Choice vs Prime was easy to tell apart; grass-fed vs grain-finished was inconsistent. Total fat content drives flavour, not the feed label. ▶ |
| Cheesesteak beef | CHUCK | Scored blind: ribeye $23/lb = 8.1, chuck roast $9.99/lb = 7.9, pre-shaved = 7.1. Ribeye is fractionally better at over double the cost. ▶ |
| Ground beef | NO PRICE LADDER | Dressed burgers from wagyu-labelled vs regular ground beef were not reliably distinguishable blind, and the label is unregulated. A separate three-tier test ($4.64/lb 73/27, $11.99/lb premium blend, $9.99/lb 93/7 lean) found no clear price-quality hierarchy — budget beef browns and tastes great, and lean is a calorie play, not a flavour upgrade. Buy 20-25% fat regular. ▶ ▶ |
| Parmesan | 36-40 MONTH | Sweet spot is 36-40 month (~$3/lb over standard 24). 60 and 100-month are novelty. In a cooked sauce all five ages were hard to tell apart blind, so use cheap aged in the pot and real DOP only where it's raw on the plate. ▶ |
| Canned tomatoes | MID-TIER | DOP San Marzano at $6.30 was matched by $4.69 options. The real rule is negative: avoid cans with calcium chloride or packed in juice rather than purée. ▶ |
| Dried pasta | BUY CHEAP | Pasta is nearly flavourless on pure taste. ~80% of quality comes from shape, salted water and doneness. Pay extra only for bronze-cut texture; spend the savings on sauce. ▶ |
| Feta | HOT vs COLD | Sheep/goat PDO feta stays creamy when heated; cow's milk turns rubbery. Buy real feta for hot dishes, cheap cow block for cold. Never pre-crumbled. ▶ |
| Butter | NO BRAND PREMIUM | Across ~15 butters, aroma differences were detectable but none was objectively better, and sautéing changed the aroma more than the brand did. Colour comes from the cows' grass diet, not quality. Worth 50¢-$1/stick only where butter is tasted raw. ▶ |
| Olive oil | MID-TIER EVOO | EVOO is stable enough to cook with, contrary to the usual advice. "Light"/refined is just expensive flavourless neutral oil. Judge by free-fatty-acid % and harvest date, not marketing. Separately: on roasted vegetables all three tiers were indistinguishable — the aroma burns off — but expensive EVOO came through clearly in a shallow-fried cutlet. ▶ ▶ |
| Garlic | NEVER JARRED | Jarred garlic was worst in every single test — pickled and muted. Fresh wins decisively raw and wherever garlic is the star; granulated or paste are fine when it is a background note. Allicin peaks ~60 seconds after cutting and dies above 140°F. ▶ |
| Soy sauce | ONE GOOD BOTTLE | Across 20+ bottles the 4-year aged Japanese one stood out clearly in fried rice — and was undetectable in sushi and in a mixed dipping sauce, where sesame oil masked it. Cheap hydrolysed soy is pure salt with no aroma but still usable. Buy the splurge bottle only for dishes it finishes. ▶ |
| Salt | SCAM ABOVE $1 | All of it is 97-99% pure sodium chloride regardless of brand, and dissolved in water the tiers were indistinguishable blind. Trace minerals are nutritionally negligible; crystal size changes only dissolution speed and how it feels as a finishing salt. Season to 1-1.5% by weight and learn one salt's density. ▶ |
| Onions | BUY CHEAPEST | White, yellow and sweet were nearly indistinguishable in blind triangle tests. "Sweet" is a mildness label, not more sugar — all sit near 6%. Only red onion and shallot read as clearly different. Buy on price for anything cooked. ▶ |
| Potatoes | BY TEXTURE, NOT TASTE | Raw, every variety tasted nearly identical (sweet potato excepted) — the difference is starch and water, so variety is a texture decision. Russet for fries and chips, Yukon Gold for mash (and worst for frying). ▶ |
| Canned tuna | PLAIN ALBACORE | Across 21 cans from $1 to $21, the $21 tuna-belly can was "no way worth that"; cheap tuna plus good olive oil approximates a premium oil-packed can. Albacore is the best all-purpose pick. Buy plain and season it yourself. ▶ |
| Spices (sourcing) | INTERNATIONAL GROCER | ~$72 for 18 spices / 80oz at an international grocery versus $120+ for the same at a supermarket in small jars. Buy whole where possible: whole keeps 1+ year, ground degrades in 3-6 months. ▶ |
| Vanilla | EXTRACT, NOT BEANS | Real extract (~$10-12) clearly beat imitation in frosting. A $4-5 whole bean barely registered in pancakes. Imitation is fine in heavily baked goods. ▶ |
| Truffles | BLACK ONLY | Black winter truffle worth it split across a table (~$15/head); white is not. Oil, paste and powder don't replicate fresh. Microplane, never shave — grating releases far more aroma. ▶ |
| Saffron | CHEAP, BUT DON'T SKIP | $79/oz and $141/oz saffron were not reliably distinguishable in a blind rice test, so buy the budget threads. But it is not substitutable — turmeric gives the colour and a completely different flavour, not a lesser one. Use 0.1-0.2g per dish (more turns bitter), grind the threads, and bloom in water, never oil. ▶ |
| Caviar | CHEAPEST DOMESTIC | Across six tins from $10 to $250 per 30g he could not tell $50 white sturgeon from $130 Osetra in a blind triangle test. Start at $30-60 domestic sturgeon; the premium is scarcity and branding. ▶ |
| Balsamic | CONDIMENT GRADE | Traditional DOP ($100+/100ml, 12+ years) genuinely won every test, including over ice cream — but $15-50 condiment grade is the everyday sweet spot. The cheap vinegar-based stuff was clearly worst as a garnish. ▶ |
| Coffee | $15-20 SPECIALTY | A $75 Gesha was good but not worth $50-60 over a well-sourced $15-20 bag. Light/medium/dark labels are unreliable — buy on tasting notes. Roasting is a marginal fixed cost; green sourcing is what drives both flavour and price. ▶ |
| Flour (pizza) | DOMESTIC | $20 imported Italian 00 was not worth it. Bread flour for pan/Detroit style, domestic 00 for Neapolitan. ▶ |
| Milk | NOT BY PRICE | At matched fat %, pasteurisation method was detectable blind — and his favourite was the ultra-pasteurised one, not the priciest. ▶ |
| Air fryer | SKIP | No faster and no crispier than a convection oven once time or doneness is controlled. ▶ |
Kitchen unlocks
Single-variable findings that change how the other two lanes should be executed.
🌡️ Doneness is temperature and time
165°F is instant pasteurisation, but 150°F held 3.7 minutes achieves the same log reduction — as does 155°F for 45 seconds. Chicken and turkey can be served far juicier and still be safe. White meat is best near 150°F while dark wants 175-180°F, so separate or spatchcock rather than roasting whole. ▶ eM4bNy
🍗 Thin the breast, keep the moisture
In a thick chicken breast the edges reach 205°F while the centre reads 155°F. Pounding to even thickness took moisture retention from 79.6% to 91.4%. Normalise thickness rather than trusting one probe reading. ▶ lAn4CX
🧀 The cheese sauce that survives service
Traditional cacio e pepe has only a ~20°F safe window before it breaks. A cornstarch gel (10g starch : 150g water) widens that past 200°F and makes the sauce reheatable without splitting. ▶ 10lXPz
Sodium citrate is the stronger version of the same trick. At 2-3% of the cheese weight, plus 0.5 part liquid, it emulsifies any cheese into a stable sauce with no roux, eggs or cream — including aged high-fat cheeses that otherwise split into fat, protein and water. It is the compound already in American cheese, which is why that melts smooth and cheddar alone does not. For a line this is the answer: real cheddar flavour with processed-cheese behaviour. ▶ PTbdvN
🥩 The flat iron loophole
A chuck-primal cut that is nearly filet-tender with ribeye-level beefiness — and crucially it stays tender even cooked to 165°F while most cuts toughen past 145°F. A cheap steak that forgives a distracted line cook. Also won his cheesesteak cut test. ▶ aQ0xc-
🌶️ Season after the heat
Black pepper's piperine is volatile: pepper added after cooking had a noticeably stronger bite than pepper cooked with the steak. Same logic drove his aromatics finding — herbs and garlic land hardest poured over sliced meat, not basted into it. ▶ 97Z25r
Acid follows the same rule. He seared flank steak plain and applied the guajillo marinade after, getting more perceptible marinade flavour without compromising the sear; his spiedie test found acid-heavy marinades mush and dilute the meat, so his hybrid marinates in mayo and hits it with vinaigrette off the heat. For a line this is better anyway: one seasoned protein, many finishes. ▶ cnRo2f ▶ qTNjm1
🧂 The umami spec, in numbers
MSG and salt are not interchangeable. His favourite blend was 75% salt to 25% MSG, dosed near 1% of weight, and a blind triangle test correctly picked the MSG bowl as more delicious. It degrades above 437°F, so it belongs in cures and finishes rather than in the sear. Crucially it is amplified by inosinate and guanylate — meat, mushroom, dried fish, aged cheese — which is the mechanism behind Guga's house umami salt beating pure MSG blind, and behind the Weissman salt-and-MSG cure. Three corpora, one spec. ▶ RCa43F
🍟 The fries lane already has its method
Boil in salted, vinegared water (acid protects pectin so they don't disintegrate; salting the boil is "non-negotiable" for seasoning), flash-fry ~60s at 375°F, freeze, then finish at 375-400°F for ~4 minutes. The freeze step is a service model, not a garnish: it holds for months and finishes from frozen. Baked versions lost clearly every time he tested them. Russets only. ▶ IP6vgP ▶ dklh4o
🥕 Restaurant vegetables, demystified
Salt drives the flavour, not spices; just 15-30g of oil per pan gets 80-90% of the restaurant result; roast hot at 450-475°F with space, finish with acid. Frozen vegetables and frozen aromatics were indistinguishable from fresh in blind tests. ▶ zKEwA_
Smaller findings worth keeping
- Rice: the packet's 1:2 ratio over-hydrates — use 2:3 for distinct grains.
- Pasta: ~1.8× the pasta's weight in water, timed from 180°F, cooked 53% faster than a full pot.
- Spices: toast whole then bloom in fat early — both beat adding raw at the end in blind tests. Bay leaf does measurably work, but needs heat.
- Karaage: starch instead of flour, plus a double fry, is the crispiness mechanism (matches the Weissman lane).
- Kettle chips: crunch comes from a longer fry at a lower effective temperature, not a hotter one.
- Sweet potato fries: too little starch to crisp — brief boil then a cornstarch slurry adds it back.
- Onions: cut pole-to-pole for mildness and even texture, orbital for maximum punch.
- Salt: teaspoon weights differ wildly between salt types — calibrate your own pinch and salt by taste.
- Dry brine: 1.5% salt by meat weight seasons all the way through; unbrined meat tasted salty only on the exterior. Wet brining matched it but costs a bucket and fridge space. Usefully, 2 hours gave no meaningful gain over 18 — this does not need to be a day-ahead prep task.
- Bias cuts: cutting on the angle roughly doubles surface area; bias-cut carrots browned visibly more than straight-cut in the same pan and time.
- Evaporated milk beat regular milk for a stovetop cheese sauce (smoother, no scorching risk), and stood in for heavy cream in butter chicken at roughly half the calories with almost no flavour loss.
- Wings, confit logic: a low 230°F first fry for 25 minutes renders the fat, then a 400°F finish crisps only the outside. Another par-cook-and-hold model, same shape as the fries freeze and the Weissman par-fry.
- Skirt steak flagged as the standout budget cut: fatty, juicy and tender as long as it is sliced against the grain.
- Eggs peel by start temperature: cold-start eggs racked up 24 peeling flaws against 9-10 for boiling-start or steamed. Steam them; it is the single cheapest prep-line fix here.
- Cheese blend: for a stovetop cheese sauce the sweet spot was 50-75% melting cheese to aged — all-Velveeta is smooth but blank, all-extra-sharp is grainy. Sodium citrate is why processed cheese melts without splitting.
- Crispy skin: air-chilled birds air-dried 4+ hours, coated with baking powder, cornstarch and salt at 1% of weight each, crisped far faster in a 450°F oven. Regular chicken gets there too, but only by cooking longer and drying out the meat.
- Bread: diastatic malt visibly out-rose the same dough without it, and steam took the crust to 195°F in 1 minute instead of 4, buying much more oven spring.
- Braises reheat better: he makes them a day ahead deliberately. Collagen dissolves at 160-205°F and the resulting gelatin holds 10× its weight in water, which is why day-two is juicier, not drier.
- Basil dies in the fridge: chilling injury sets in below 50°F. Store it on the counter in water and add it at the end; cooked-in basil read as "generic herb" against the fresh-added control.
The contradiction worth testing
In his 43-minute steak masterclass, Chlebowski found that butter-basting added no detectable flavour over neutral oil, and that aromatics registered most strongly when poured over the meat after slicing rather than basted during cooking. ▶ AaHbfr
That collides head-on with the other two lanes. Guga's corpus has brown butter winning the basting-fat test outright and the one-teaspoon-of-brown-butter finish tested as equal to full basting. Weissman butter-bastes as standard practice.
How we resolve it: we don't, on paper. Guga and Weissman were tasting browned butter, where the flavour comes from caramelised milk solids; a neutral-oil comparison against plain melted butter is a different experiment. This goes to the kitchen as a live test — brown butter versus neutral oil, blind, on the same cut. It is the first thing the Menu Lab should actually cook.
The mayo convergence
The strongest new thread in this corpus, and the one that lands directly on our signature-move question.
Chlebowski arrives at mayonnaise as a searing and browning medium from six independent directions, never framing it as one big finding:
- On a sous-vide steak, seared 20-30 seconds a side: "probably the easiest experience I've ever had searing a steak," with no mayo aftertaste. ▶ sKB9Lf
- Remastering a grilled chicken recipe, mayo browned noticeably better than oil, which simply ran off the meat. ▶ da3AgI
- On grilled cheese it had a far higher margin of error before burning than butter: its soybean-oil base tolerates up to 495°F against butter's 300-350°F. ▶ c2JWHI
- On a spatchcocked turkey, it browns like butter with less risk of burning. ▶ vTllx1
- As a marinade it needs less volume than oil for equal browning, so it is also the lower-fat choice. ▶ v1VUg6
- And it carries spice onto lean meat that would otherwise shed its seasoning. ▶ FBWMT6
His stated mechanism: mayo is an egg, oil and water emulsion, so it clings where oil runs off, and it browns both thermally and chemically — the egg protein is Maillard fuel, not just a hot fat.
Where this connects (my synthesis, not his claim): that is the same mechanism as Guga's milk-powder sear. Milk powder works because it paints protein and lactose onto the surface, giving the Maillard reaction fuel it otherwise lacks. Mayo does it with egg protein and an oil that cannot drip away. Two creators, working independently and never referencing each other, landed on the same move: paint a protein-carrying film on the meat before it meets the heat.
That reframes the signature. The distinctive thing is not milk powder specifically, it is the surface-film category — which means it is worth testing the obvious third option nobody appears to have run: mayo and milk powder together, protein plus reducing sugar plus a clinging fat, on one cut, against each alone. That is kitchen test #2. It is also the rare technique that is invisible on a plate, which matters, because "mayo-seared" is not going on a menu.
Systems, not recipes
His other contribution is structural: he argues recipes hold cooks back, and teaches frameworks instead. For a restaurant these are prep-list architecture.
The frameworks that transfer to a kitchen
PCSV (protein, carb, sauce, vegetable) as the build unit instead of fixed dishes — this is effectively a menu matrix, letting a small prep list generate many plates. Component batching: his ramen build makes stock, tare and aromatic oil ahead so assembly is minutes, exactly parallel to the Guga sous-vide and Weissman par-fry service models. Stir-fry in five modular parts (sauce, aromatics, protein, vegetable, starch). Risotto-style pasta, cooked in just enough liquid so the starch builds the sauce in-pan, one pan and no colander.
And the grocery finding that scales to purchasing: an identical 10-item list cost $37.44 at the cheap store versus $69.84 at the expensive one, and the cheap version still made genuinely good food.
He also independently validated the par-fry service model. Testing General Tso's chicken four ways, deep-fry then freeze then deep-fry again came out crispiest, with fry-then-air-fry a close second and air-fry-only the weakest but still good. That is Weissman's par-fry-and-hold model arrived at separately, with freezing as the hold step rather than refrigeration — and it means a fried lane can bank product ahead and still plate at full crunch. ▶ aQK8z8
And the fry economics, from his rebuttal to Adam Ragusea's case against home deep-frying: a wok needs only 2-3 quarts of oil and contains splatter better than a Dutch oven, oil is reusable four to six times if skimmed rather than burnt, and each additional batch costs about 4 minutes 20 seconds. Those are the three numbers that decide whether a fried lane is affordable. ▶ v2pCvV
What this lane does NOT contribute
He is not a menu lane. There is no Chlebowski dish going on the board the way picanha or fried chicken does. His corpus is horizontal: it makes the other lanes cheaper, safer and more consistent. Treated as a menu lane he would underdeliver; treated as the sourcing and technique layer he is the highest-leverage of the three.