đ¤ Random thought. I do not mean this in a pejorative way. More anthropologically speaking, in terms of "studying people" and observing their interactions with each another & emerging tech.
â How folks choose to use AI (or not) reveals little glimpses into one's thought process and ability (willingness?) to pay attention to seemingly inconsequential details. Code nuance, so to speak.
â In that way, it is a bit like the phenomena of the Dunning-Kruger effect. That is a fascinating story (link in the comments). Two psychologists posited that without a certain baseline level of self-awareness, humans tend to over-estimate their own abilities.
â I think that tends to compound with AI. Someone can prompt their way to passable results. Which to the untrained eye (vibe-* coder, designer, businessperson) look just like the "real thing."
â Emulating versions of success they have seen others have. Without comprehension. Like me in the 1990s, thinking if I could wear Air Jordans then⌠Maybe? A half-Asian kid could play basketball at that level.
Such workflows tend to result in some comical outcomes. Those who create the vibe-work present it to others who vibe-evaluate it, and everyone continues along happily until something breaks. That inadvertently places a heavier burden on those who understand the fundamentals, to serve as a backstop against quality control.
I half-jokingly said to another principal engineer friend:
> "I miss the days when iffy devs tended to move slowly. Now they can churn out iffy stuff at ludicrous speed."
Essentially, the difference between these statements.
1ď¸âŁ "AI writes better code than me. It can keep large chunks of contextual details straight. Whereas I cannot. Ergo, AI is better at writing code than all developers."
2ď¸âŁ "AI writes better code than me. I use it to generate things that appear correct to my eyes. However, levels of human competence exist beyond my own."







