Italian Chef — cooking answers grounded in food science
A cooking skill that answers from the mechanism rather than the recipe. Ask why your pizza dough tears and it explains gluten development and hydration instead of telling you to knead more. Covers dough systems, pasta mechanics, regional Italian technique, and ingredient selection down to why DOP matters where it does and where it does not.
What it actually knows
Dough as a material. It reasons in terms of the two proteins that matter: glutenin, which gives elasticity and rebound, and gliadin, which gives extensibility. Most dough problems are best described as a failure of one or the other, which is why “knead it more” is such a poor answer — sometimes more kneading is precisely wrong.
Hydration in baker’s percentages, with real ranges. Low, 45–55%, for extruded pasta and sfincione, where you want a firm dense crumb. Medium, 58–65%, for Neapolitan and New York pizza. High, 70–85%, for pizza in teglia and pinsa romana — and it will tell you that pinsa needs to be near 80% or you simply will not get the open honeycomb structure that makes it pinsa.
Regional logic rather than a recipe list. North: soft wheat, butter and lard, egg-based pasta — where the white contributes plasticity and the yolk contributes lecithin and silkiness. Centre: soft wheat and farro, olive oil, pici and pane sciocco. South: durum wheat, extra virgin olive oil, dried pasta and pizza napoletana. Once you know which column a dish sits in, most ingredient substitutions answer themselves.
Certifications as specifications. DOP and IGP are treated as technical requirements rather than marketing. It will tell you that AVPN-standard pizza napoletana is a direct dough with no fat and no sugar — and equally, it will tell you when a DOP product makes no measurable difference to the result and you are paying for the label.
Why it answers differently
Recipe sites optimise for search traffic, which means they describe steps without explaining what the steps do. That is fine while everything works and useless the moment it doesn’t, because a list of instructions gives you nothing to reason with.
Once a model is told to reason from the underlying food science, its answers become debuggable. You can describe what went wrong and get a diagnosis rather than a suggestion to try the recipe again.
How to use it well
Describe the failure, not the dish. “My dough tore when I stretched it — 65% hydration, tipo 00, 24-hour cold ferment” gets you a specific cause. “How do I make pizza” gets you a recipe you could have found anywhere.
Give it your flour type and your hydration if you know them. If you don’t, tell it what you have and let it work backwards — flour protein content changes the answer more than almost anything else in the process.
Where it falls short
It has never tasted anything. It reasons well about the chemistry and poorly about whether you personally will like the result.
It also cannot see your kitchen. Home ovens run hotter or cooler than their dials claim, flour behaves differently by brand and by season, and water hardness genuinely changes fermentation. Treat it as a knowledgeable friend on the phone rather than a chef standing next to you: excellent on why something happened, dependent on you for what actually happened.