Post

The following might sound weird. But it's more evident than *ever* that code assisting tools lead me to *instant* debt. Yeah, I know, I'm doing it wrong. Hold on. When I fix their mess, I realize like these *products* are *designed* to behave like this. It's very subtle and *very* dangerous.

Think: does it feel like these tools offer just *selfless help*, or are they maybe, *just maybe*, operating an *hostile takeover*? A manipulative, subtle, silent, and $1T marketed takeover? Notice how I am not referring to LLMs in general. I'm speaking about these *products* specifically.

All this would explain why most folks in the industry have this "soul sucking" feeling. And yeah, I know, I know. Let me get my tinfoil hat, sure. But it's been years now, and damn, it's so obvious to me. I can't possibly be the only one forming this thought.

I think it's time that, just like when we investigated the psychological design choices social networks made, I think we should do the same for Claude Code, Codex, and the such. ASAP.

Obviously I recognize LLMs have their place. These *products* are.. way more than that. They are designed to come after *our* emergent abilities, so we lean towards *their* artificial ones, so we can *trust* them, and - most importantly - so we can throw away our problem solving *confidence*.

That is exactly why these products, *since day *zero*, came after scientists and creatives, rather than focusing on eliminating repetitive, mechanical, and physical work. That is why it feels these products are more likely designed like an addicting drug, just like Zucc did with social networks.

Like, all around me, slowly but steadly, I can see friends, colleagues, CTOs (et. al.) being more and more dependent on... them. I've even read the following recently: >what LLMs do you guys use to edit your spreadsheets? ????? like, try this?