The Hunt for Dark Breakfast

Dark-breakfast candidates and recipes

• Many commenters propose actual dishes near or in the “abyss”:
  • Egg-heavy pancakes/crepes (similar to the article’s recipe), German pancakes/Dutch baby, soufflés, choux, Salzburger Nockerln, Japanese soufflé pancakes.
  • French toast, especially very eggy challah-based or “stuffed” versions, plus savory bread puddings.
  • Waffles + omelette hybrids (e.g., “Womelette”), waffle frittatas, egg-and-cheese sandwiches, biscuits and gravy topped with eggs.
  • Custard-like dishes such as Aggakake / oeuf au lait (3 eggs, 2 cups milk, 1 cup flour).
  • South and Southeast Asian egg breads: roti telur / egg paratha, Sri Lankan egg hoppers and string hoppers.
• Some argue that these effectively “fill” much of the dark region already, so the gap may be more conceptual than real.

Expanding the breakfast space beyond the triangle

• Several people argue the egg–milk–flour triangle is too limited:
  • Missing axes: meat (bacon/sausage), potatoes, oils/fats, sugar, vegetables, cheese/yogurt, grains beyond flour (oats, porridge, muesli), fruit, fish.
  • Suggestions to treat breakfast as a higher-dimensional “latent space” or simplicial complex rather than a 2D triangle; in more dimensions, the “forbidden” zone might vanish.
  • Others note many culturally important breakfasts (vegetable-forward dishes, porridges, full fry-ups, breakfast burritos) don’t fit well into the chosen coordinates.

Math, geometry, and interfaces

• There is a technical subthread on how the visualization works:
  • Interpreting recipes as positive vectors in {egg, milk, flour}, then normalizing to sum to 1 to get barycentric coordinates on a simplex.
  • Clarifications and minor corrections about simplexes, 2-sphere vs 1-sphere, and why “negative eggs” don’t arise in this model.
• A long comment links the breakfast simplex to Embedded Constraint Graphics: using glued simplices and barycentric interpolation for UI design, facial animation, and drawing tools, with analogies to how high-dimensional embeddings work in machine learning.

Cultural and meta discussion

• Multiple commenters praise the piece as unusually creative, comparing it to Douglas Adams or xkcd and saying this is “why they read HN.”
• Some criticize omissions (French toast, maple syrup, cheese/yogurt, porridge, biscuits) or the US-centric view of breakfast.
• There is a side discussion on why breakfasts are so uniform in American restaurants and why morning meals tend to be lighter (with one explanation invoking circadian glucose dynamics).
• Many enjoy the dark-matter parody, extending it with jokes about dark breakfast cosmology and “gravitational lensing” of diners.