What does an engineer actually do

What does a Software Engineer actually do?

If you’re a software engineer worried about AI taking your job, you might want to consider what it is that you actually do. What is the reason you’re payed to do what you do.

In hindsight, a program is just math and we have been working on proving programs for decades. It really should have been no wonder that the act of programming has been mechanized. We have been mechanizing math since before the days of Babage and Ada Lovelace conceptualized a general purpose math machine.

Did we really think we were beyond replacement? Writing code is not what I do. I have never liked the term “coder” even though to call myself a Software Engineer feels pretensous.

There is nothing wrong with writing code but do not be fooled into thinking it is the only way to get a high quality result.

I’m an artist, you give me a tuba and I’ll get you something out of it.

The issue we have today is more about imagination. How can we use this tech? Content generation is a lazy idea. Using it to build web-pages is a web developer thinking that they are the user they are designing for. An AI help desk is also lazy. This needs to be a case by case issue. Look at SendCutSend. The site is a metal laser cutting service. They offer quick turn around of custom orders, made to order. They have an AI part generator. You give it a prompt and it gives you a part based on the prompt. It isn’t perfect. I asked it to give me a 12 tooth gear that is 22mm across. It didn’t get the geometry of the teeth right, but it got all the dimensions correct and it gave me something that looked like a gear. This wasn’t edge based, but it was fast and it matches the business. This is an example of a good solution.

Remote MCPs (Model Context Protocol, a standard that lets an AI model call out to external tools and data sources over a defined interface, rather than every app rolling its own integration) are another untapped resource. MCPs offer user authenticated access. This means the content can be controlled or tailored to the user as well as other services. The mistake to avoid is imagining the machine doing what a human is doing today. Instead, imagine what is cost prohibitive to have a human do today. This is why services like Grammarly work so well and have gained wide spread adoption. People would prefer to have an AI check their spelling and grammar before they hit send because people don’t want to look like an idiot. Everyone wants a copy editor, but no one can afford it.

Someone who is unimaginative cannot see what products can be built that they do not use. Lazy people only see obvious one size fits all solutions. In game design you want someone who has a breadth of knowledge and experience more than a depth of knowledge. This will be more valuable in general product development moving forward.