Artificial intelligence is an amplifier of existing human intelligence.
With AI, the rungs of the intelligence ladder only get further apart, not closer together.
It doesn’t miraculously allow an unmotivated individual to become the CTO of their own tech startup with minimal effort. If the hurdles stemmed from an individual’s inability to comprehend complex problems, those hurdles will still exist, no matter how well the AI tries to explain them. Even if they somehow found themselves running that company, they wouldn’t understand what they had built, how the pieces fit together, or what to do next.
There is an assumption that AI can manage the hardest parts for us and that the answer is always one prompt away. The reality is that any productive dialogue requires both parties to follow along at a similar level of comprehension; otherwise, the discussion grinds to a halt.
Many conversations exaggerate AI’s capabilities and minimize the need for human involvement to absurd levels, largely due to a lack of understanding of how things actually get built.
“Create me an e-commerce website exactly like Amazon and don’t bother me with any of the implementation details.”
As soon as we scratch below the surface, we realise that without understanding detail, human preferences cannot even be articulated at each stage.
A non-programmer cannot hope to develop any meaningful piece of software without engaging with the details: the coding style, the architectural trade-offs, the underlying stack of frameworks and libraries, and why they may or may not be needed.
Likewise, no one can operate as a civil engineer, lawyer, or surgeon without the ability to follow the technical dialogue and critically evaluate the decisions being made.
This brings me back to comprehension and an individual’s capacity to understand what is in front of them. AI is fundamentally no different from a textbook and a good mentor. We have the entire internet for reference and an extraordinarily patient tutor in AI, but asking AI to complete a task without understanding it yourself is like skipping school entirely and expecting a degree.
Some might say, “Well, I have a great idea for a business—all I need is help to build it.” Herein lies another common misconception: good ideas are not rare.
Before Uber came along, countless people had the thought that booking taxis over the internet might be convenient. But having the concept is merely step one of a million granular decisions needed to execute a viable company.
Uber, like all enduring companies, required continuous effort, domain understanding, and meticulous attention to detail. AI will never bypass the human need to supply all three.