​The Paradoxical Thinking That Can Upgrade Your Engineering to a New Level

Or: How to Have Your Cake and Eat It Too

There are some things in engineering that are very counterintuitive that, if not fully understood, prevent engineers from doing their best work. These perceived paradoxes can limit or misdirect your thinking such that you are not the engineer you can be. In this article, I am going to share two examples of this and then introduce a maxim that can help you with this.

Example 1 – quality is cheap
The above statement seems very counterintuitive; indeed, it is the case that you can often cut back costs at the expense of quality.
However, when a system is easy to implement, making it cheaper to realise, there is also less chance of a mistake being made, meaning quality is higher.
This was one of the reasons Toyota was so successful in the automotive industry – by focusing on making things easier to produce for higher quality, they also became cheaper. The opposite was true as well. By cutting out waste (wasted time, wasted parts, wasted steps) to reduce cost, there was less that could go wrong so quality increased.
This is not to dismiss the importance of more expensive items when needed, but often they will be cheaper in the long run anyway. It’s just that engineers can sometimes give themselves false assurity by choosing the more expensive option.

Example 2 – simple code can do more than A.I.
One would think that with more lines of code and a complex algorithm there would be more going on so it can do more.
But code that has been written explicitly can be tuned so that it achieves the exact goal desired. If a trained ML system does something wrong, then you likely have no idea what’s causing the issue. All you can do is make a few adjustments to the training system, retrain, and hope for the best.
If, on the other hand, you have explicit code, then you can understand what’s happening (or not happening), and make changes to improve performance.
This is not to be dismissive of things A.I. Comparing ELIZA to what’s on offer today shows how powerful A.I. systems have become. It is more that we can think A.I., because of its complexity, will be the better option regardless.
But still, when one considers the success in the early 1970s of much simpler systems – such as the one documented by F. T. de Dombal et al. in Computer‑Aided Diagnosis of Acute Abdominal Pain where a relatively simple program outperformed senior specialists (91% success vs 80%) – there is the potential for simpler code to do more.

What is missing in engineering thought for this to happen?

It can be found in the 1920s.

For some time, it could not be determined if light was a wave or a particle. It had properties of each – depending upon the test. However, by the 1920s, it was accepted that light was both a wave and a particle.

The “OR” was replaced with the “AND”.

And that’s the key thinking difference you need to adopt.

Stop thinking things like:

• We can have low cost or quality.
• We can have efficient code or accuracy.

Instead start thinking things like:

• What can I do to have both lower cost and high quality?
• How can I approach this so I can use simple code and increase accuracy?

​In short, stop asking the “OR” questions and start asking the “AND” questions. Make this a habit, and you will start producing better engineering outcomes – because you will not full into traps of intuition.
Everyone wants to have their cake and eat it too.
And engineers should be trying to make that happen!