In middle school, I learned the basics of thermodynamics via a little diagram of a steam engine. Arrows (red for hot, or blue for cold) showing water moving from one place to another, and simple equations on the sides, and we were meant to help us grasp basics of conservation of heat, entropy, and the relationship between pressure, volume, and temperature in gasses. 

Imagine my surprise this week when I learned that the people who developed the laws of thermodynamics learned these concepts the exact same way. The invention of the steam engine predates the theoretical concepts that explain why and how a steam engine works, by over one hundred years.¹ 

Apparently more science happens this way than most of us really imagine. Pharmaceutical research is often shaped by throwing different compounds at lab cultures and seeing what happens, and its pretty common knowledge that we are currently way out over our skis as far as generative AI – these models are hard to predict and control because even the people who build them don’t really know how they work. ²

Its worth noting that these 18th century physicists didn’t just have some gaps to fill in, they were totally in the dark. If you look at understandings of concepts like heat, pressure, and the properties of gas at the time – they were not close. In the 17th and 18th century, there was a belief that heat functioned basically as a liquid, and that it being hotter or colder was a property of its position, more similar to how water falls or rests based on where and how you are holding it. Heat as energy was nowhere near the conversation. 

Thermodynamics as a concept developed as a way of understanding how you could make steam engines run more efficiently. The Carnot cycle (which describes the way energy is lost to entropy in a thermodynamic system) was based on observation, not on theoretical knowledge. 

The theories about heat moving like water were very elegant, but completely wrong.³ Physicists learned they were wrong because they watched how the systems worked in the real world, and they realized it was much messier and more complex than they had ever imagined. 

We often meet innovators who guard their ideas like treasure, who can explain in great detail how their programs and projects should work, how they are theologically informed, evidence-based, and pedagogically robust. At least on paper, but they haven’t implemented much of anything yet, because it’s not the exact right timing or team or funding. The fact is that there is a lot that can only be learned once you try something. There are plenty of elegant ideas that are right, but there’s plenty of elegant ideas that are completely wrong, and sometimes you have to just build the engine and see how it runs. 

This week, we encourage you to try something that you’re not exactly sure how it will work yet. We encourage you to try a half-baked idea, follow your gut, and embrace messy imperfection in the service of learning something. And when your elegant ideas don’t pan out, or don’t pan out exactly the way you expected, that might be a sign that you are about to learn something pretty spectacular.

1. Kameron Sanzo, Ph.D. https://medium.com/@kameron.sanzo/what-came-first-thermodynamics-or-the-steam-engine-c0cca0996b0b ↩︎
2. MIT Technology Review https://www.technologyreview.com/2024/03/05/1089449/nobody-knows-how-ai-works/ ↩︎
3. Kameron Sanzo, Ph.D. https://medium.com/@kameron.sanzo/what-came-first-thermodynamics-or-the-steam-engine-c0cca0996b0b ↩︎