Post #2541537
2026-05-14 10:02 UTC
Replies (16)
-
@doragasu@mastodon.sdf.org 2026-05-14 10:08
@lina@vt.social There goes determinism.
-
@buherator@infosec.place 2026-05-14 10:08
@lina@vt.social Proper automation has an upfront cost, and I think this is what ppl try to avoid. It's not uncommon that people wait for frontier models to get better at deciding on logical conditions instead of writing two if's...
-
@wombatpandaa@mastodon.social 2026-05-14 11:07
@lina@vt.social not to mention the accountability issues AI introduces
-
@adrian_blx@mas.to 2026-05-14 11:11
@lina@vt.social Some folks in that company forgot how to do that and are now using the slopmachine for LSC... *sigh*
-
@Profpatsch@mastodon.xyz 2026-05-14 11:52
@lina@vt.social All of these refactoring tools are proprietary or internal to the companies (and dependent on their setups) though. So you need a company with $10 million in spare to design such a tool and apply it. Saying LLMs don’t improve the situation here is completely removed from reality!
-
@gimulnautti@mastodon.green 2026-05-14 12:33
@lina@vt.social It’s a cult. The rich must get richer. It’s a cult. Their ideas must be good. It’s a cult. If not, the world must be made shittier for it to be true. It’s a cult.
-
@emc2@indieweb.social 2026-05-14 13:34
@lina@vt.social Had to sit through an AI booster talk yesterday I couldn't get out of. At one point, the speaker was like "AI can write boilerplate code for you, accessors and modifiers, and other boring stuff, change names..." I'm like "Bro, emacs has done that for decades"
-
@bovaz@misskey.social 2026-05-14 13:38
@lina@vt.social and to add: at different scales, those kind of large refactorings happened also in smaller companies and shops, so it's not like they were exclusive to big companies that could dedicate people to that. So using LLMs isn't even justifiable by bs like "giving smaller companies big-company capabilities".
-
@ahltorp@mastodon.nu 2026-05-14 14:01
@lina@vt.social We also had code-generation tools, called compilers, that took input from a highly formal language with its own task-optimised syntax and semantics, and when these so called compilers generated something that didn't exactly match what was specified in the input, we would consider this an enormous problem. In fact, compiler faults would be so uncommon that it wouldn't even be one of the error causes anyone normally thought of, but sometimes people would joke about it being the cause.
-
@oberstenzian@mastodon.social 2026-05-14 14:57
@lina@vt.social That’s what’s been most disheartening about all of this. In a scant few years we went from “nearly competent” to “monkeys throwing feces at the screen”. Now a bunch of systems don’t even work anymore and critical thinking has been thrown out the window.
-
@Ralph@hear-me.social 2026-05-14 15:05
@lina@vt.social #alttext "We're being told to use [Al] agents for broad changes across our codebase. There's no way to evaluate whether that much code is well-written or secure—especially when hundreds of other programmers in the company are doing the same," a UX designer
-
@graydon@canada.masto.host 2026-05-14 15:09
@lina@vt.social Y2K made it necessary to pay programmers more than management. Generally a lot more than management. Old programmers, often. Women programmers, a fair bit. People cutting the cheques got traumatized. Some of those people have COO and CFO titles now. They still believe axiomatically that money exists to confirm the social hierarchy, not to create efficiency through accurate prices. AI is a way to enforce the social hierarchy by appeal to the oracle. And to never pay programmers again.
-
@ferrix@mastodon.online 2026-05-14 16:45
@lina@vt.social "Properly Written Robots" would be a great name for a company that made slopless automation
-
@tschundler@leds.social 2026-05-17 19:52
@lina@vt.social I have had to review incoming AI mass code changes. So far every time I've found problems. They've "fixed" tests by masking the problem instead of fixing the problem. (Worse than when the LLM just deletes the tests and so you know clearly it's being sketchy) Or introduced concurrency issues (this is Go so very prone to concurrency dangers). Very rarely if ever happened with the older large scale refactors.
-
@peteriskrisjanis@toot.lv 2026-05-19 07:36
@lina@vt.social and when dust will settle, techies will be blamed for shoddy state. Techies will be ones cleaning up. And no exec will bear any responsibility about it.
-
@quincy@chaos.social 2026-05-26 05:01
@lina@vt.social This is what really stymies me. My coworkers (CS graduates most of them) seem to forget this too. It's like the "AI" industry is brainwashing them into forgetting that there was a time before nondeterministic, water-guzzling slop bots ... they forget that - better - tools already exist to facilitate every precise task. Except one can't "chat" with them. (But why would one want to "chat" with a computer!)