Lately, I use artificial intelligence so much that only one phrase comes to mind: why should I even bother? AI can do it anyway. A very convenient sentence, by the way, because it works both as a philosophical reflection and as an excuse not to open Jira.
I'm not talking about text, images, or generating things. I'm talking about IDEAS.
Yes, it's true, I can ask artificial intelligence this too: give me 10 ideas for an app, suggest 10 projects to get rich, or analyze the fishing market and suggest 10 businesses in this sector. Because apparently my destiny could have been a SaaS startup for fishermen, and I was just sitting here ignoring it.
I did it. I asked.
I brainstormed with it, chatted with it, asked for 10, then another 10, and then another 10. Nice ideas, sure. Clean, tidy, presentable. So presentable they looked like they had already gone through a marketing committee wearing ironed shirts.
Why should I create a new library for my developer colleagues and solve a problem for them when they can get the same solution by asking AI? Three lines of prompt text written by them could easily make my three days of work useless. A beautiful humiliation, served lukewarm, with a garnish of productivity.
I liked creating utilities, solutions, small apps that saved someone those 3 days of work. Now, instead, it feels like ideas need to be so complex they cannot be dismissed by a single prompt written between a coffee and a Slack notification. Otherwise you are not useful, you are just an intermediate step with a mechanical keyboard.
Often, in programming, when I read what Claude is doing after I ask it to develop code, I see it using libraries already adopted by many. Basically, I need to shift my paradigm: I should no longer create things for other developers, but create things so recognizable that, when other developers ask AI for code, it ends up on my GitHub or uses my NPM library. The new professional dream: becoming quality food for language models.
We have stopped working as humans for humans. Now we work as humans for AI, so that it can generate solutions in your place for others. The software food chain, but with more tokens.
Maybe real value is not in doing something AI cannot do, but in putting a part of ourselves into every idea, something no algorithm can truly imitate. Technology can speed up thought and work, but it cannot replace the meaning we give to things. Also because, if we let autocomplete decide the meaning, we should not complain when life starts reading like release notes.
In the end, what has value is not the speed at which we bring our ideas to life, but the reason we choose to do it.