[dudeinhawaii]

I'm very PRO AI and a daily user but let's be real, where is the productivity gain? It's at the edges, it's in greenfield, we've seen a handful of "things written with AI" be launched successfully and the majority of the time these are shovels for other AI developers. Large corporations are not releasing even 5x higher velocity or quality. Ditto for small corporations. If the claimed multipliers were broadly real at the org level, we would be seeing obvious outward signals like noticeably faster release cycles in mainstream products, more credible challengers shipping mature alternatives, and more rapid catching-up in open-source software projects. I am not seeing that yet, is anyone else? It feels like the 'invisible productivity' version of Gross Domestic Product claims, the claims don't seem to match real-world.

And those are the EASY things for AI to "put out of work".

HARD systems like government legacy systems are not something you can slap 200 unit tests on and say "agent re-write this in Rust". They're hard because of the complex interconnects, myriad patches, workarounds, institutional data codified both in the codebase and outside of it. Bugs that stay bugs because they next 20 versions used that bug in some weird way. I started my career in that realm. I call bullshit on AI taking any jobs here if it can't even accelerate the pace of _quality_ releases of OS and video games.

I'm not saying this won't happen eventually but that eventually is doing a heavy lift. I am skeptical of the 6-12 month timelines for broad job displacement that I see mentioned.

AIs (LLMs) are useful in this subtle way like "Google Search" and _not_ like "the internet". It's a very specific set of text heavy information domains that are potentially augmented. That's great but it's not the end of all work or even the end of all lucrative technical work.

It's a stellar tool for smart engineers to do _even_ more, and yet, the smart ones know you have to babysit and double-check -- so it's not remotely a replacement.

This is without even opening the can of worms that is AI Agency.

[sjsizjhaha]

> the smart ones know you have to babysit and double-check -- so it's not remotely a replacement.

This has been the big one for me. Of all the best devs I know, only one is really big on AI, but he’s also the one that spent years in a relational database without knowing what isolation levels are. He was absolutely a “greenfield” kinda dev - which he was great at, tons of ideas but not big on the small stuff.

The rest use them judiciously and see a lot of value, but reading and understanding code is still the bottleneck. And reviewing massive PRs is not efficient.