The familiar structure of a modern recipe—title, ingredient list with precise measurements, and step-by-step instructions—feels timeless, yet it's a relatively recent invention. And now, this format may be evolving again as AI enables more flexible ways of representing recipes.
For most of history, cooking knowledge traveled through communities by demonstration. As society became more mobile in the 19th century, cookbooks became popular. But early recipes served more as memory aids than self-contained formulas. Consider this recipe for "Rusks" from 1864:
"One pound of flour, small piece of butter big as an egg, one egg, quarter pound white sugar, gill of milk, two great spoonfuls of yeast."
Note what's missing: no precise measurements, no cooking temperature or time, and no guidance on technique. Recipe writers in the mid-19th century often pointed to familiar objects or ratios ("as big as one egg") to estimate how much of a certain item was needed.
Fannie Farmer's Boston Cooking-School Cook Book, published in 1896, popularized recipes with standardized measurements and detailed instructions that anyone could prepare. (Or try, at least.)
As recipes became more standardized, curating a deeply personal collection of recipes became a new form of self-expression. A recipe box in the mid-20th century contained cards bearing handwritten notes and splatters of oil that told stories beyond the recipes themselves.
The internet triggered another transformation. Recipes moved online, evolving from simple text to blog posts with photos, and eventually to YouTube and TikTok videos. This reintroduced elements of demonstration and community that were lost in written recipes. We can once again see techniques performed and chat in the comments section.
But with infinite recipes at our fingertips, we face a paradox of choice, where abundance can lead to less satisfaction. Today's home cooks must develop skills in information filtering and adaptation that previous generations never needed.
We’ve gone from playing instruments to making mixtapes to shuffling Spotify.
I see evidence that AI may transform the way cooking content is developed, stored, and shared, hopefully solving some of the current challenges and creating new opportunities.
- AI can analyze cooking patterns across diverse recipes, helping cooks see underlying principles behind the lists of ingredients and instructions. These extracted patterns can help cooks understand the 'why' behind techniques, supporting confident improvisation rather than rote following.
- Personalization can bridge the gap between standardized recipes and an individual's needs. AI can help cooks adapt recipes to match a nutritional profile, fit a budget, or fulfill a desire to use what's local and in season.
By understanding how ingredients and techniques interact, these systems can suggest substitutions, adaptations, and variations that preserve the essence of dishes while allowing for individual expression
My project, Flavorful, explores these possibilities through an AI-powered system based on flexible frameworks that adapt to a user's shopping list, pantry, and personal preferences. Flavorful's tools help cooks understand culinary principles and confidently adapt them to their own needs—making cooking once again a personal act of creation that honors its earliest roots.