I’m a senior full-stack developer of 15 years, and more recently, a new tech lead at an AI startup. I’m definitely not attacking AI as a concept in general.
I work with AI agents every day and all day. That’s how I develop and plan our systems. It did not start that way. I was absolutely against the use of AI during development, but a few months back, I need the assistance because I developed carpal tunnel syndrome, so that’s what I automated, just the typing and the implementation of low-level logic so that my wrists can heal. But do you know what stock AI agents do to code when not given proper guidance? Ask any real developer and they’ll tell you about vibe coding. I guarantee these are not going to be success stories.
I’m not just judging people for being lazy, because lazy people like me will innovate ways to stay lazy by inventing/optimising new shit that allows them to stay lazy. That’s a survival instinct and an evolutionary selection mechanism: minimising energy investiture while doing the same thing as everyone around you is an evolutionary advantage.
No. What I’m judging them for is delegating their critical thinking capacity to an external entity, and stunting their own cognitive growth (their literal reason for existing in the first place, their continuity mechanism to stay in the gene pool, and their sole means of improving at being long-term lazy) by being short-term lazy. Makes sense?
Now to generative AI (for the multimedia substrate):
The vast majority of people you speak of are now polluting the collective “training set” with diluted slop distilled from all art historically created thus far, because the content generation equation went from: X people creating Y novel pieces of art per year, to X models creating Y million images per day, all thanks to a handful of idiots with more greed/money than common-sense. That diluted pool is ever-expanding, growing geometrically, and burying actual novelty with each new image Susan generates and shares for her new “Katz Rule” instagram profile.
The thing is: the next model will be trained on that averaged set, and the next, and the next. With each day and each generation increasing in conformity. And that set is what we’re stuck with for new inspiration (and future models) now. Because everyone is looking at screens for inspiration, and not at mountains or rivers, or even the real stars in the sky at night because we ruined that too.
All while we’re doing the things you just mentioned.
All thanks to a few assholes with more selfishness than common-sense chasing after unlimited quarterly growth in a very limited space that’s closing around us fast.