Artificial intelligence is a commodity, but understanding is a superpower

The debate about intelligence versus wisdom is as old as history, but artificial intelligence has transformed it into an intensely practical question. The cheaper professional knowledge becomes, the more precious it is to know how to use it. It’s becoming ever clearer that the most valuable thing is not just the power to do things but wielding that power effectively. Formulating and comprehending aims in the context of complex systems and uniting the burgeoning sprawl of content with clarity of strategic vision: These are stars of the new game.

This is nowhere truer than in software development, where content is executable. Here we have a bizarre paradox where it’s a known fact that more lines of code means more surface area of maintenance, where established practice shows that more output!= better outcomes. And yet the current fad is that people who understand software—software developers—will soon be replaced by AI.

Honestly, I think the inverse might be closer to the truth. The heart of a developer’s skill set is the ability to move between ideas, goals, and implementation in software. As it turns out, this corner of the universe is currently growing by orders of magnitude. While anyone can now use human language to generate working code from AI, each time they do, developers have a bit more territory to roam.

Maybe the generated code is of high quality, meets the requirements, and integrates with the overall project intent and infrastructure. Maybe it’s easy to understand and maintain; maybe it isn’t. Code that is well-thought-out and delivered implies comprehension of both the goals and the underlying system. And you know what you call a person who does that? A software developer.

AI can’t deliver that sort of code because AI doesn’t understand anything. To take in and absorb the importance of things is a purely human function. It is also hard work. It is becoming ever more rare, at just the moment when it is becoming even more necessary.

Intention as the middle ground of enterprise innovation

The middle ground of enterprise innovation is where strategic goals are connected to business and development activity. This middle ground of bridging intention is where purpose meets technique. It is something AI cannot do without human guidance. It can only assist.

As a developer and a human being, you want to push yourself as much as possible to incorporate the intention of things into your practice. By insisting on understanding a project’s intention and uniting it with your own understanding of the particulars of implementation, you become far more valuable. AI then makes it easier to magnify your intentions into automated activity.

We can speculate that AI will get better at this middle ground in the future, but it will never actually have intention. It will only ever move under human direction. Resist becoming just a connector or interpreter of intention to implementation. Keep on working to develop and contribute your own unique understanding. Implementation can be automated, but the unique qualities of understanding cannot.

Why LLMs will not replace higher-level languages

If you listen to some AI enthusiasts, it might seem that AI’s ability to mass produce code to meet requirements makes understanding the intention of that code less important. I’d say it makes it less necessary up front. There may even come a time when AI’s natural language interface is something like what fourth-generation languages are today. I can see a possible future where languages like JavaScript and Python are a layer below the AI interface, akin to how C is today. But if that is the analogy we’re using, then it seems clear we will always need people who deeply understand that layer, just as today we still need people who understand C, assembly machine code, and chip wafers.

But I don’t really see large language models wholesale replacing higher-level language programming anytime soon, if ever. The current generation puts immense pressure on the humans involved to ensure minimal change and conciseness. These things, as we well know from experience, are utterly essential. Getting the job done versus getting it done gracefully is not a trivial distinction when it comes to programming.

You can abstract the way a loop is implemented. You might be able to use the LLM instead of a for loop or a forEach function. But somewhere in there, it’s still implemented as a loop or function. Someone still needs to understand the concept of iteration as it relates to data and the system where it operates.

Finger pointing at the moon

It is notable that the Gartner Hype Cycle in June 2024 had generative AI on the downslope, rushing quickly toward the slough of disillusionment. When something has so much excitement and potential around it, it’s tough to stand aside and see it clearly, but that’s exactly what we as developers need to do. We have the perspective to really understand what AI can and can’t do and use it in the best ways. The more we do that, the better the results will be for all of us.

The fact is that AI is trained on the average output of humanity, so you’re going to get average results. Extraordinary genius and everyday excellence in human achievement result from tying the intangibles, intention, and the spirit of creativity into the minutiae of medium and technique. AI can only give a sampled approximation of things, not the living core. That core is the source of genius.

In Zen terms, and the philosophy of Bruce Lee, AI is all finger pointing at the moon, but no moon.

Intelligence versus wisdom

I learned how to play Dungeons & Dragons years ago, and the distinction between the intelligence and wisdom attributes was explained to me as “knowing it’s raining” versus “knowing to come in out of the rain.” Another way to render this is knowing how to do versus what should be done.

For me, intelligence moves toward reduction, whereas wisdom moves toward integration. Intelligence sees the parts and wisdom sees the whole. We need both, obviously, and embracing both is what I’m advocating for here. The fact that intelligence can be artificially produced calls for exercising both intelligence and wisdom more effectively, but especially wisdom. We will need all the wisdom we can get to manage the expanding volume of content created without understanding.

As a developer, your unique perspective and how you bring varied elements together are irreplaceable. At whatever level and in whatever context you find it, understanding is the most precious thing. Understand deeply, both the how and the why, and everything else will flow from that.