Worked example: Constraints Are Not the Enemy of Creativity
Labs reference: Worked example: Constraints Are Not the Enemy of Creativity.
Provenance: Mike Besso’s prior work, reconstructed with Zai.
Scenario: Turn an embedded-systems memory—about 64K of memory and a 150-baud, two-wire network—into a useful principle for modern AI work without romanticizing scarcity.
Gather the thought before drafting
-
Mike Human
We are going to create a 45-second elevator speech, let's start by gathering our thoughts on this topic: why constraints can make creative and engineering work better.
-
Zai AI collaborator
What personal experience gives that claim weight, and what keeps it from becoming nostalgia for scarcity?
-
Mike Human
In 1990, I worked with about 64K of memory and a 150-baud, two-wire network. Every feature, byte, and message had to earn its place. I am not saying painful hardware was good; it forced priorities into the open.
-
Zai AI collaborator
How does that connect to AI work now?
-
Mike Human
AI can make words, code, images, and iterations feel free. The costs did not disappear; they moved into review, complexity, maintenance, energy, and human attention.
-
Zai AI collaborator
We are aligned: useful constraints reveal tradeoffs, while bad constraints can simply cripple a design.
First draft
Ask for the 45-second statement
-
Mike Human
Based on what we have agreed, draft a 45-second elevator speech for people excited about AI's apparent abundance. Keep the embedded-systems detail, connect it to hidden costs, and do not praise scarcity for its own sake.
-
Zai AI collaborator
Constraints are often treated as obstacles to creativity, but useful constraints can improve a design by forcing priorities into the open. In an embedded system with 64K of memory and a 150-baud network, every feature, byte, and message had to earn its place. Modern AI tools can make output appear almost free, which encourages us to generate more before deciding what matters. The lesson is not that scarcity is always good. It is that a design without meaningful boundaries can grow without direction, while a useful constraint forces us to choose.
Refinement
Make the hidden costs visible
-
Mike Human
Add the costs that show up later. That will keep the comparison from sounding like “everything was better when hardware was painful.”
-
Zai AI collaborator
Here is the revision: Some of the best engineering designs emerge not after every constraint is removed, but after the right constraints make priorities unavoidable. In 1990, I worked with a distributed embedded system that had about 64K of memory and communicated over a 150-baud, two-wire network. Every feature, byte, and message had to earn its place. Generative AI can make words, code, images, and revisions feel almost free. But the costs have not vanished. They appear later as review effort, complexity, maintenance, energy use, and demands on human attention. Scarcity is not automatically good, but meaningful boundaries force us to decide what matters.
Refined 45-second explanation
Some of the best engineering designs emerge not after every constraint is removed, but after the right constraints make priorities unavoidable.
In 1990, I worked with a distributed embedded system that had about 64K of memory and communicated over a 150-baud, two-wire network. Every feature, byte, and message had to earn its place.
Generative AI can make words, code, images, and revisions feel almost free. But the costs have not vanished. They appear later as review effort, complexity, maintenance, energy use, and demands on human attention. Scarcity is not automatically good, but meaningful boundaries force us to decide what matters.
Timing note: about 45 seconds at a conversational pace; use the personal detail only when it serves the listener.
Why it works: It grounds an abstract principle in lived experience, makes current costs visible, and explicitly rejects the idea that scarcity is automatically virtuous.
Source: labs/forty-five-seconds-to-make-it-clear/gallery/constraints-are-not-the-enemy-of-creativity.md