akrm

this is enough for me
opinions of my own

[akrm @cachebag]

does anyone know how to get LLMs to stop fucking ending all of their output with "One caveat", "A few things worth noting", and then proceeding to talk about irrelevant things? I've genuinely done all the grifter AI shit like SKILL.md and shit and nothing works.

[akrm @cachebag]

@denbvk What an insanely stupid way to say "it's too hard for me"

[akrm @cachebag]

@dhh This is addiction.

[TechToby @techtoby__]

The best Software Engineers in the US right now are telling you they don’t review code and have 10, 20, 30 agents running at once

Meanwhile your British 40+ middle manager drops in to standup a couple times a month “yeah I’ve tried AI, not that great would never pay for it”

[akrm @cachebag]

@skeptrune "Software" is not solved because writing it was never the issue to begin with.

It's so crazy how people can be shameless daft on this app.

[carol @cloisvillle]

NO WAY TOM WELLING ACIDDENTELY POSTED HIS DICK ON HIS STORY

[akrm @cachebag]

@ThePrimeagen Ugh. Why are you doing this to me Prime. You were the chosen one.

[akrm @cachebag]

@notloganhogg @mgill25 That doesn't make sense. You don't `git blame` merge commits. That would be silly way to place reasoning for a bug or defect.

The merger is responsible yes, but the author owns that code. Thus they should be responsible for triage and/or post-mortems.

[Manish Gill @mgill25]

The single most important AI guidance I gave to my team was - does not matter who wrote the code, you own it in production.

"I didn't do anything, I asked agents..." is something that would get any regular, non-popular engineer immediately thrown out of an incident call.

[akrm @cachebag]

@HiTopFilms I'd rather see a near perfect Jason Todd story that takes 10 years than a rushed, slopfest that was forced out in 2 years because some strangers on the internet don't understand what it takes to produce a feature length film.

[jame @kloogans]

alright

[akrm @cachebag]

@dhh @crertel it's so baffling to me how someone once so revered for his contribution to open source and the software industry in general, can now be so fucking daft and out of touch.

[akrm @cachebag]

@C99Chad @mitchellh You guys do know it is basically impossible to make a deterministic system that detects whether or not something is AI generated.

You can really only tell off almost arbitrary heuristics (em dashes, style of closing sentences, transitions, etc.)

[akrm @cachebag]

@Raynerdtech macOS and Linux can sandbox an untrusted program so it isn't allowed to touch your fs.

[akrm @cachebag]

@rfleury SunOS when?

[akrm @cachebag]

I love when I ask an LLM to tell me about a function’s call stack and it decides to run my entire linting + test suite

[akrm @cachebag]

@kuhron3 This. Nobody actually cares that he's negotiating. In any other circumstance, this is fine....

The issue the motherfucker is in no position to negotiate after the season he had.

[akrm @cachebag]

@BumYlin Ngl, if we're gonna welcome "vibecoding" (I fucking hate that term) into any space, gaming mods is completely fine with me.

Let people have fun. Barring that they don't create malware and try to spread it to users.

[Ryan Fleury @rfleury]

My position, which has been the same since the beginning, is: I will never pay for a subscription service to program (nor rely on some Internet service), and I will never turn my brain off and stop caring about the details I care about today.

I am not inherently attached to typing every character, and I think the idea of human language processing and accelerating the production of programming language semantics is exciting. I would only ever do this using local models on my own hardware, and I would employ such tools very carefully (likely using them in my own tooling), and likely not in the ways that many people are employing them today.

If I can accomplish the same work I would’ve been doing by hand more quickly, that’s pure win. But if I accelerate yet lose attention to detail, or the ability to think through those details, that’s not the same work anymore, and so there is some loss. That loss is potentially very important.

[akrm @cachebag]

@gmkurtzer My post explicitly noted that I do not believe AI can become conscious. I don't even think there's a chance.

I was entertaining the original author's post...