The Intelligence Loop: Why Everything Smart Has a Little Circle Running Inside It
https://commons.wikimedia.org/wiki/File:Ashton_Frost_engine_governor.jpg

The Intelligence Loop: Why Everything Smart Has a Little Circle Running Inside It

As a person with a strong software engineering background I have been fascinated to explore the worlds of mechanical, electromagnetic and electronic engineering here at ePropelled and how they interact with the world of software engineering.

My job title includes the word 'intelligence' and it has become extremely clear that there is a great deal of intelligence built into everything that ePropelled builds.

What do I mean by 'intelligence' in this context?

When you strip clever machines down to their essentials, you find a surprisingly simple pattern hiding underneath: a loop.

Not a metaphorical loop—a literal cycle where a system senses, interprets, acts, and checks what happened, over and over again.

Once you notice it, you see it in everything from 19th-century steam engines to modern AI.

And the best part? This loop works across completely different types of systems: mechanical, electronic, and software.

Let’s take a tour through each.

What the Loop Looks Like

The diagram lays out the basic cycle:

The 'Intelligence Loop': Knowledge is evaluated and predicted by Intelligence; Intelligence refines Models, deriving further Knowledge; Intelligence Drives Execution; Execution updates the External and Self Contexts; the Self Context feeds into the Senses via interoception and proprioception; The External Context feeds into the Senses via external senses; Senses update Knowledge.

  1. Senses – some way of detecting what’s going on both externally and internally
  2. Knowledge – an internal picture or model of what those signals mean
  3. Intelligence – the part that decides what action to take
  4. Execution – putting that decision into the world
  5. Feedback – the results loop back so the cycle can continue

It’s the universal heartbeat of a “smart” system.

What changes from system to system is how each part is built.

1. Mechanical Loops: The Steam Engine That Knew When to Calm Down

Take the classic flyball governor, one of the earliest engineered feedback systems.

It doesn’t use electronics, code, or even explicit rules.

But it still runs the loop beautifully:

  • The spinning balls sense engine speed through centrifugal force.
  • Their physical position effectively is the system’s knowledge about whether the engine is going too fast.
  • The geometry of the linkages “decides” how much to adjust the throttle.
  • The throttle valve executes that adjustment.
  • The engine’s new speed feeds back into the governor through the same physical motion.

The system behaves intelligently not because it’s “aware,” but because the loop continuously turns motion into control.

2. Electronic Loops: Circuits That Keep Things Steady

Move ahead a century and you find the same loop running inside analog controllers and everyday electronics.

A great example is a PID controller keeping a quadcopter stable:

  • Sensors sense roll, pitch, and yaw angles.
  • Those signals feed into internal settings—its knowledge about what “stable” looks like.
  • The controller computes corrective action in real time.
  • Motors execute the corrections.
  • New sensor readings come back as feedback, closing the loop.

Even though nothing here is “thinking” in a human sense, the loop lets the system respond rapidly, smoothly, and reliably to a changing world.

3. Programmatic Loops: Software That Learns What Works

At the digital end of the spectrum, the loop becomes flexible and adaptive.

Imagine a simple reinforcement-learning agent:

  • It senses the state of a game or environment.
  • It updates its knowledge—its internal model—based on past outcomes.
  • It decides which action is likely to improve its score.
  • It executes the move.
  • The result (good or bad) becomes feedback feeding the next cycle.

The components are the same as in the mechanical and electronic loops, but now the internal “knowledge” changes over time.

The loop doesn’t just stabilise behaviour—it improves it.

Why This Loop Matters

Because no matter how different these systems look on the outside, they all succeed for the same reason:

they continuously connect sensing, understanding, action, and feedback.

That’s the secret behind:

  • the steadiness of a steam engine,
  • the smooth flight of a quadcopter, and
  • the learning ability of modern software agents.

Smart behaviour comes from the cycle—not the material.

Once you see the loop, systems that seemed unrelated suddenly share the same underlying rhythm.

And that rhythm is what unlocks for ePropelled the 'intelligence' of the solutions both for ourselves and our customers.

Love this Robert! A laymans terms explanation of what is happening in the technology around us?

Like
Reply

Nice. I agree it is "loops everywhere". The dark side is there too of course, chaos and oscillation. Tuning the loops seems to be my main task these days as I wrestle with Claude and co.

Like
Reply

To view or add a comment, sign in

More articles by Robert Boothby

Others also viewed

Explore content categories