Why ‘Monkey Work’ Makes Better Engineers

26 4 minutes read

Why ‘Monkey Work’ Makes Better Engineers

Listen to this article

I’m seeing a greater number of social media posts about how great it is that computers have freed ‘engineers’ from all of the ‘grunt/monkey/donkey work’ they used to have to do. This trend, which has been apparent throughout my career, has accelerated recently with the introduction of so-called AI.

Setting aside for now the fact that these tools are widely misused, were never designed to do the things that people are using them to do, and in the case of ‘AI’ are a confidence trick, these jobs need doing, they need doing well, and they are important for a number of reasons.

The jobs I am talking about are things like producing and checking calculations and drawings; materials takeoffs; producing schedules of valves, instruments, drives etc.; engagement with suppliers for pricing; and increasingly detailed nuts and bolts design.

The idea that these tasks are essentially beneath the dignity of the (usually more junior) engineers who tend to be assigned them, or that they are a waste of engineering resources which might be better assigned to tasks which can be billed at a higher rate is shared by both junior engineers and management. They are wrong.

Many junior engineers know less than nothing about engineering. Those who teach engineering nowadays mostly never worked as engineers. Engagement with real engineering deliverables is what teaches graduates that what they learned at university is not really engineering, and starts them on the path to learning what engineering actually is. These jobs are not ‘grunt work’ for early career engineers. They have much to learn, and this is the time-honoured way to learn it.

Having painstaking ‘donkey work’ done by would-be engineers rather than by a glorified predictive text engine like ChatGPT means that there is accountability. If it is wrong, we know who screwed up. This might add to the undesirability of the task for the ‘grunts’, but equally this provides their motivation to get it right. Attention to detail and accountability are things engineers must learn. This is how they learn them.

Details matter in engineering. Trusting software to produce schedules, isos, piping layouts, materials takeoffs, costings, etc. without a detailed 100% check of output is in my view negligent. Trusting black-box software to simply tell you the answer to an engineering problem, with no way of checking whether it is correct (other than calculating it yourself by another method) is stupid.

Companies who allow engineers to use such software, on the grounds that it saves costs, must assume that they do not really need to check the output. It is quicker for an expert to produce reliable answers to engineering questions than it is to fully check AI output, especially if the input was done by junior engineers.

Academia has a lot to answer for here. Many of those who teach ‘engineering’ are pure scientists or mathematicians, who consider academic research and purer sciences superior to the rough, dirty, mercenary world of engineering.

They are commonly rather disdainful of industry, practicality, and what they call the ‘technician level knowledge’ of engineering practitioners (who they call ‘industrialists’, rather than ‘engineers’). They know nothing about the things they disdain, and they consider themselves too good to learn. Having attempted to teach these things to academics, it is my opinion that they flatter themselves. Many lack the aptitude to learn them, even when properly motivated.

Today’s green graduates tend to believe that they are already fully-formed engineers, but there was probably never a time in history when they were more wrong. They need those few years of ‘grunt work’ now, more than they ever did.

They will learn that the struggle to produce your own model of a process design in MS Excel is not for donkeys. It teaches you things about the way the parts of the process work together that you would not have grasped if you had just asked AI to tell you the answer, or done just enough to get a simulation package to converge.

They will learn that laying process plant out in space is an essential part of process design, and that the necessary engagement with suppliers to do that properly is an opportunity to really learn how things work together. They will probably also find that bad assumptions were made at previous stages of design which need correcting, making them less likely to make those bad assumptions themselves in future.

They will learn that if you aren’t properly costing your decisions, you aren’t really engineering. Especially for beginners, it can be hard to conceive of how large the cost differences can be between options, and grasp the difference between the nice-to-have and the essential.

They will learn from their time-consuming checking that even very experienced engineers can make errors in their calculations and drawings, but they will also learn how the real pros do it.

The grunts may not appreciate this, but they are freeing up the real engineers to do the high level stuff that needs to rely upon their own struggle to master the basics.

But if the grunts already know, why would they need or want to learn? Well, I’m still learning, and I’m still doing my own grunt work a lot of the time. I’m not going to tell you that after thirty-something years of experience, I have learned that I know nothing. I know a lot more than the majority of engineers I come across about my little bit of engineering.

I see a lot of errors, a lot of sloppy work, a lot of over-reliance on unreliable sources of information of all kinds. Perhaps that is a function of doing a lot of troubleshooting and expert witness work, but I don’t think that now is the time for less grunt work.

I think we need more checking, more supervision and more mentoring of junior engineers by the remaining competent engineers. I think junior engineers need to grasp that they are not engineers yet, and if they think that engaging with real engineering is monkey work, they never will be.