[romaniv]

The article has solid observations, but I would correct one important thing. It's not AI confidence, it's AI psychosis.

A lot of people I know are forced to use AI at work. They universally tell me that their coworkers generate awful PRs with bugs, nonsensical code and fake unit tests. But they also universally tell me that they are different, have special workflows and prompts that create good code.

The psychology behind this is obvious. Hype and the literal threat of being fired forces everyone to develop coping mechanisms. Bragging about your own adaptability is one of those mechanisms. Unfortunately, the scale and intensity of this and the fact that the mechanism is clearly weaponized for marketing means we're living in an increasingly deranged society.

[adam_arthur]

AI is clearly a force multiplier, both negative and positive.

The truth is there are prolific developers like Antirez who have built quality new projects at an incredible pace (Dwarfstar 4, Redis features).

But as unpopular as it is to say it, in the working world ~80% of developers pre-AI mostly just attended meetings, did a little busywork and committed small patches here and there. Probably around 20% really moved the needle and contributed the bulk of net new code.

Those 80% were constrained in the volume they could output pre-AI, but now they are unleashed to do a large amount of net new work but many without the skills to structure it well+maintainably.

It doesn't help that most management has been pushing on LoC over quality the past year.

I truly believe most companies as they exist today are not structured for AI. The amount of technical debt that will be created at a rapid pace is basically time delayed self destruction for most codebases if you let people run amok with low contribution standards and rubber stamped approvals.

If you treat each AI output as a small well-scoped, well-tested module, which interoperate with each other through well designed APIs, you can have high confidence in quality. But majority of people are pseudo-vibe coding and creating spaghetti monster codebases, and there's really no way to stop it without strong and tight technical oversight.

[ghostpepper]

This is also the dichotomy I'm seeing. This meme sums it up https://fixupx.com/danhockenmaier/status/2021617680525172840

I like to imagine a similar dynamic happening when previous transformative technologies were invented: when the power tools like the chainsaw were invented, I wonder if there was a cohort of carpenters that dismissed them because of how much damage they could do in unskilled hands.

[oblio]

https://www.oxfordreference.com/display/10.1093/acref/978019...

[slopinthebag]

Doesn’t it seem a bit odd that such a prolific developer has only managed to produce a PR for a new Redis type, and a olama fork, despite having a 100x productivity booster machine for the last 8 months?

[lioeters]

There are several other examples that come to mind, of prolific and respected developers turning to AI-based workflow and ever since then just constantly producing mediocre content and code. The spark is gone, the taste, the genius - whatever you call it, the human quality that made them interesting in the first place. Sure they may be more "productive" superficiallly, but they all sound the same now. Any other decent dev with access to a corporate code-generation service can get basically the same result.

[therealdrag0]

False dichotomy much? The tools can be force multiplying without being “100x productivity boosting,” which I’ve never once heard claimed except from critics.

[pseudalopex]

antirez said 100x.[1]

[1] https://x.com/antirez/status/2058808933309927877

[therealdrag0]

He could be very wrong, or just excessively hyperbolic.

[slopinthebag]

Which implies his Redis PR would have taken two decades without AI.

[mstaoru]

This is a good observation. Antirez is one my favorite authors and software developers. I'm 100% sure he could have made the same Redis PR and ds4 (or something else) without LLMs in about the same time with the same quality or higher.

[throw310822]

> Those 80% were constrained in the volume they could output pre-AI, but now they are unleashed to do a large amount of net new work but many without the skills to structure it well+maintainably.

You could say that AI turned them from stupid and lazy to the famously dangerous combination of stupid and hardworking.

[Aurornis]

> They universally tell me that their coworkers generate awful PRs with bugs, nonsensical code and fake unit tests. But they also universally tell me that they are different, have special workflows and prompts that create good code

If someone told you this (their coworkers generate bad PRs, but they generate good PRs) in the age before LLMs would you have also declared it psychosis?

Having to deal with lazy coworkers who submit bad PRs has been a feature of workplaces since the dawn of programming. Programming languages are a tool and they can be used or misused by the operator.

The difference now is that a lazy developer with an LLM can become prolific with their sloppy output and it’s harder for a lazy manager to notice.

If you’re trying to imply that nobody can use LLMs to good effect then that’s just denial at this point. The way good developers use AI isn’t to prompt and then submit PRs. It’s used as a pair programming partner. The developer still writes, reads, edits, and is responsible for the code.

[mathisfun123]

> their coworkers generate bad PRs, but they generate good PRs) in the age before LLMs

You think this a clever gotcha but it's not because no one was generating PRs before LLMs.

[bwfan123]

> fact that the mechanism is clearly weaponized for marketing

There is a lot of dumb money chasing the AI dream. That money is not asking hard questions about return on investment yet. But, the narrative is starting to shift as evidenced by this article. Even Meta is questioning the value of AI Agents.

In response, we see Anthropic, OpenAI, and Microsoft creating giant teams of FDEs (forward deployed engineers). The idea is to keep the AI dream alive one way or another because once that dies, the whole thing comes crashing down. So, dumb money is going to put up a strong fight in the marketing-sphere, and articles like these are needed to counter that.

[dofm]

That's "fundamental attribution error" as much as it is any form of psychosis, though I do agree it is at work.

Also at work: the Dunning-Kruger hypothesis, Gell-Mann Amnesia and motivated reasoning.

[fithisux]

The problem is that these people are left untreated because that is what people at power can exploit.

This is social collapse.

[sfadsdfsdfad]

I am not forced to use it, but I am a peculiar type of dev. Very lazy. Can go for weeks without producing anything of value while I think of stuff, delete stuff and generally just cogitate. I would work for 3-5 hours if I could but I have to sit in a chair for 8 hours to please the pointy haired people.

My colleagues are absolute work horses. They work for 8-10 hours minimum. Always responding to emails, putting out fires. Always typing furiously, producing lots of code. You are waiting for it, I know it and yes, it's all shit and all those fires are of their own making.

I picked up stuff in a month or two that took my colleagues years and I'm not a genius, as you can undoubtedly tell. I am just exceedingly lazy and I honestly think it's pathological. It got me in a lot of trouble in the past, but now, with the way AI is going I'm having a blast.

I have some proclivity for architecture and directing large complex projects. I hate typing code, but I love pointing out how it should be done. This shit is a life saver but for the love of God if you are not supremely lazy and averse to code, don't use it. My productive colleagues + AI = goddamn disaster.