Cat

Software eng, cryptography nerd, self proclaimed cool office cat

If I interact with you, it's in good faith. Lift each other up

@[email protected]

[Cat @CatThatPrograms]

Can't help but agree, but it won't matter when the cost of those decisions won't matter until after you've saved money axing experienced software engineers.

[Ben Dickson @bendee983]

And the sad part is that a lot of the vibe-coders using AI coding agents like Claude Code have never heard of "big O notation" and don't know the difference between O(logN), O(N), and O(N^2).

They see a prompt doing magic and become convinced that they have developed a production-grade application.

I know I've said this like 100 times, but coding agents will only amplify the gap between experienced software engineers and people with no coding experience.

[Cat @CatThatPrograms]

As an engineer ask yourself:

"Am I an LLM proxy?".

If yes, re-evaluate how you use tooling:

Do you push back?
Accept the 1st solution?
Critique output?
Use AI powered reviews?
Use different contexts/models for review?

How do you apply your knowledge?

Don't be a proxy.

[Cat @CatThatPrograms]

(3) AI dev loops feed the same dopamine cycle that addictive game loops provide. It's why engineers are over contributing code and under contributing actual thought which is what you should be paying for...Not a proxy to an LLM service which is the path many eng are treading.

[Cat @CatThatPrograms]

(1) It's not just juniors and mid-level ENGs that are over relying on AI and AWS will find this out soon enough I imagine. It's a specific approach towards ai tooling and refusal to let your brain become mush that will protect you and your company, not "seniority".

[Lukasz Olejnik @lukOlejnik]

Amazon is holding a mandatory meeting about AI breaking its systems. The official framing is "part of normal business." The briefing note describes a trend of incidents with "high blast radius" caused by "Gen-AI assisted changes" for which "best practices and safeguards are not yet fully established." Translation to human language: we gave AI to engineers and things keep breaking?
The response for now? Junior and mid-level engineers can no longer push AI-assisted code without a senior signing off. AWS spent 13 hours recovering after its own AI coding tool, asked to make some changes, decided instead to delete and recreate the environment (the software equivalent of fixing a leaky tap by knocking down the wall). Amazon called that an "extremely limited event" (the affected tool served customers in mainland China).

[Cat @CatThatPrograms]

(2) This is not to say senior engineers may not be more prone to those mentalities, but the assumption that all senior engineers are using AI tooling correctly is a humorous one at best.

[Cat @CatThatPrograms]

Truer words could not be written. When you don't understand the code base you will generally not generate or feed the correct context, and you kick the can on deciphering output to the maintainer because you don't know the context either...

[shadcn @shadcn]

Mass-generating PRs with your agents and clawbots isn't helping open source. It's quietly burning out the people who actually maintain it.

Please stop.

[Cat @CatThatPrograms]

If you do not leverage AI output correctly, YOU move faster. Your TEAM does not. It is critical to guard rail LLM output with a stringent review process and tune context windows for tasks. If not, you are merely moving work elsewhere, not REmoving it.

[Cat @CatThatPrograms]

@Grady_Booch Sadly we are probably going to go through a big wave of non-quality software first before that quality demand hits. Gonna have to go though some catastrophic overlooking of LLM output causing some large scale problems first is my guess.

[Cat @CatThatPrograms]

At its core AI is perfect for the state of society that cares only of quarterly value and has no concern for what crests the horizon. This stands true for not only the technical implications that await but also ecological, societal, etc..

[Cat @CatThatPrograms]

@dev_maims Is it worth it to still learn art, music, or philosophy in a time where all three can be also "created" by AI? At what edge does humanity lose its meaning if we release pursuit of the arts because something else can also pursue it?

[Cat @CatThatPrograms]

@iamdevloper If you have ever watched someone do this on a 1-2b row set from BigTable or similar...You can see and feel the true pain that is inside this simple message.

[Cat @CatThatPrograms]

@fjzeit I think you have to log off for that feature.

[Cat @CatThatPrograms]

@iamdevloper How painful was it to write this

[Cat @CatThatPrograms]

@SteveOnSpeed @MetamateDaz Seeing why people pay for the check now to not need multi-posts. Anyway, knowing your deficit is valuable if you're in a position to fix it. Most people aren't so it isn't appealing.

[Cat @CatThatPrograms]

@SteveOnSpeed @MetamateDaz I agree the skill is important, however it is mentally taxing for the average person. I have assisted multiple people start to budget and at the end of it..it frequently just highlights an income problem, not an excessive spending problem.

[Cat @CatThatPrograms]

@badlogicgames Look on the bright side! You get to be a part of the last few generations of humans at the top of the bell curve for intelligence over time.

[Cat @CatThatPrograms]

Jumped back on X curious of what the current tech meta was since I've been off socials for awhile and hot damn is it spicy.

[Cat @CatThatPrograms]

@ExistentialEnso e.g. if a sum of any given two facets of knowledge means discovery then...perhaps. Would be curious of examples that are on the tip of your mind if any.

[Cat @CatThatPrograms]

@ExistentialEnso I think it's important to define "novel" here...Perhaps "latent"? Can AI tools reveal what we already know through sums of existing knowledge, would say yes. Create entirely new knowledge...Not sure. I, however, think this quickly gets into semantics as demonstrated.