I didn’t follow a recipe. I didn’t have broth. I didn’t even know if it would work.

But with a pot of vegetables and AI as a real-time collaborator, I created a rich, savory, restaurant-quality soup from scratch.

Context: Why This Matters

This started with a constraint. I had a limited food budget and a large quantity of vegetables from a community food source that wouldn’t last much longer. No recipe. No stock. Just ingredients that needed to be used.

Instead of searching for a recipe, I decided to try something different: collaborate with AI in real time.

The Setup

• Potatoes, onions, garlic, shallots
• Mushrooms (white + brown)
• Broccoli, Brussels sprouts, asparagus
• Soy sauce, turmeric, pepper
• Water (no stock, no butter, no cream)
• Daring plant-based chicken

The constraint forced creativity. There was no predefined outcome.

The Experiment: Cooking With AI

This was not about following instructions. It was about interacting, adjusting, and responding in real time.

• Asking questions during the process
• Sharing images for context
• Receiving feedback based on what was actually happening
• Adapting decisions step-by-step

AI became a collaborator, not a tool.

The Process

1. Started with mushrooms to build an umami base
2. Added onions and garlic to develop aroma
3. Used water + soy sauce to create a broth from scratch
4. Added potatoes to build structure and thickness
5. Adjusted when potatoes were slow to soften
6. Observed broth transform from thin to rich
7. Layered vegetables carefully to preserve texture
8. Finished with plant-based protein

This was a live feedback loop, not a static recipe.

Visual Collaboration

Sharing images during the process made a major difference. Feedback was based on real conditions, not assumptions.

This enabled a new kind of learning: contextual, adaptive, and immediate.

The Result

• Rich, savory, umami-forward broth
• Thickened naturally (no cream or butter)
• Vegetables cooked perfectly (not mushy)
• Multiple servings: eaten, shared, stored

The result exceeded expectations.

The Breakthrough

This was not about cooking.

This was about a new way of learning.

• Real-time feedback
• Adaptive guidance
• Confidence through doing
• Skill-building through iteration

What This Means for Incubator

• AI Literacy becomes practical
• Prompt Literacy becomes interactive
• Learning becomes experiential
• AI becomes a collaborator

Call to Action

Try it yourself.

• Start with what you have
• Ask questions in real time
• Share images
• Experiment without a fixed outcome

This is not about recipes. This is about learning how to think, adapt, and create with AI.