In a world of autonmous code generation and agentic workflows, where do fit in?
Table of contents
Open Table of contents
Background
Last month I made my first appearance at an AI Meetup here in New York City. It was an amazing exprence, I got to speak with engineers across a plethora of exprienced and backgrounds.
During the meetup I was approached for an interview where I was asked “Why is it an exciting time to be in AI” and my mind immediately triggered a cache hit on an artile I read a month before that.
The article was dubbed “I miss thinking hard.” (https://www.jernesto.com/articles/thinking_hard). In this article the author, Jernesto, shares their relationahop with AI tooling and how the author feels as though deep creative thinking is diminishing as these tooling improve.
My Take
On Linkedin I shared a summary on my thoughts on this, which is dumped here:
Made my first appearance in the AI space!
I want to thank Fonzi AI for hosting a phenomenal AI meetup. The conversations had with other engineers, seeing how people are actually applying these tools across different domains, made it such an eye-opening experience.
I wanted to expand on my interview response: the Builder versus the Thinker
When I look at how engineers are approaching AI-assisted development right now, two archetypes keep showing up. The Builder is problem-obsessed. They want to reach an end state, and AI is just the fastest path there. The Thinker is solution-obsessed. They care deeply about how the code is written (think architecture, maintainability, the craft of it)
To be clear neither is wrong. But in an environment where AI seems to be accelerating everything, leaning too hard into either one creates “blind spots.”
The Builder ships fast but accumulates invisible debt. The Thinker writes clean systems but can overthink themselves into slow delivery.
On my last note of my response, many Thinkers feel cheated. They spent years refining their craft, developing taste around how good software gets built. Now they’re watching that get commoditized in real time. That frustration is legit.
But here’s a reframe: the Thinker’s instincts are actually what make AI useful at a higher level. Knowing why something should be built a certain way is precisely what separates a good prompt from a bad one. That judgment doesn’t disappear, it simply moves upstream.
The engineers finding the most success in today’s landscape are developing a third instinct: staying aware of where AI is helping them move faster and where it’s quietly eroding their judgment. These people are deliberate about design decisions instead of delegating them entirely. They know that a model can write any code you ask for, but the quality of what comes back is entirely a function of the thinking that you put in.
That’s the skill worth developing. https://www.linkedin.com/posts/bryanmontalvan_build-ai-ugcPost-7435815431107399681-snAZ
Interview
For those who are curious the interview can be found here