A blog post built on a premise that could fit inside a paragraph and be elaborated to death within 2000 words now gets to have multiple chapters of useless fluff where the main purpose is lost because those words were supposed to convey the life circumstances or mindset of the author that lead to the premise.
> The codebase doesn’t care how it got written. It cares whether it works, whether it can be maintained, and whether it helps the business do what it needs to do.
> That’s not a bait and switch. That’s what happens when your organization gets access to new tools and the economics shift underneath everyone’s feet.
> What we have now is a training problem. A reclassification problem. And I’m not sure what the best HR-friendly way to frame this is… but here’s a serious question:
Until just a couple years ago I would regularly read comments complaining when a website doesn't work because the hacker browses with javascript disabled, for example.
This isn't the early-adopter crowd: it's the refuses to even be a late-adopter crowd.
I go to ChatGPT for basically any annoying code snippet and even functions now. I'm done ever having to guess at map reduce syntax again, or trying to remember if slice mutates the target array.
I'm messing with with codex more and more. But I still don't trust it to design features for me. Maybe in 6 months, I will. Is it really that important to force developers NOW to get to a place they'll get to in a few months anyway, assuming the hype is real?
That's what Microsoft does. The CVEs speak for themselves.
If your code is expensive, the fact is that now someone can write it cheaper.
A.I or more accurately LLMs are currently trained on shitty open source code.
the best practice code out there is locked in some cabinets for private companies.
if you insist on 100% A.I written code - then how are you gonna train the new generation to write software well.
you will come to the singular point - where the new generation knows nothing & the LLMs themselves can't be trained further (we are almost there btw).
LLMs as better autocomplete are perfect use case. or as a rubber duck that talks back in terms of debugging. anything else is frivolous.
The leaks of proprietary code, and the many examples of known security issues, the quality issues evident is most software, and the opinions of people who work on proprietary software all suggest the opposite.
Please like and subscribe!
- Writing style does reveal how people understand problems and their approach for solving them. People that prioritize direct solutions over complex abstractions are still valuable to catch over engineered code.
- People with "good taste" in code can catch when AI generated code takes shortcuts to accomplish a certain task, this happens every day and we can't ignore it.
The state of AI code can be way better by 6 months or 1 year, or even more (we don't really know), but we're not there yet, and we can't wait until there to hire new people without considering those points.
Wrote about why I think the job description already changed, and what I'd rather see teams do about it than have that exhausting conversation on repeat.
I am still baffled about engineer's or developer's use of AI. I see fail and fail on anything other than some novelty one-shotted tool, and even then folks are bending over backwards to find the value because that solution is always a shallow representation of what it needs to be.
My projects are humming away, hand made. They have bugs, but when they show up, I more often than not know exactly what to do. My counterparts have AI refactoring wild amount of code for issue that can be solved with a few lines of adjustment.
TL;DR I feel like I live in a different reality than those who write these blog
