Wildroot

Queries may use an external AI service. Details

Wildroot Browser

TodayStory on Mastodon

LLM

Trending on Mastodon: 6 people posted about it in the last 3 hours this edition covers.

Posts
31
People
25
Time
–
Mostly
AI

#AI llms software

All 31 posts of this story, best first: each person’s best post, then their other posts. The title is the name people mention most, or the opening line of the best post.

  1. @futurebird@sauropods.win

    On the day of Anthropic's IPO a group of people gathered to protest at their offices. These people are concerned that Anthropic is implementing "recursive self improvement" recklessly. And, I'm annoyed at them. Since most people don't pay much attention to the details of technical news about software... their complaint blends in with those who think LLMs are being implemented recklessly for less exotic reasons. It almost feels like deliberate misinformation designed to muddle the debate.

    OpinionPopular today
  2. I was sharing with my wife my (current) preferred approach to AI produced code by engineers. It's borrowed from her PhD - thesis defence. If an engineer can defend their code in a thesis defence, I have no concern about how it was produced. What is thesis defence for software changes? They should be able to explain, without using an llm: * What is the objective for the code change * Why did they choose this approach * What other approaches they considered * What unrelated/unexpected things they found along the way * What are the limitations of chosen approach * What is the wider/longer term impact of choosing this approach

    Show more of this post

    * Who else did they collaborate with, or who else understands their chosen approach to be able to answer these questions if needed Notice I don't want them to explain their work. I can read the code and understand it. I want to understand how well researched this approach is, and how well they understand it. I also want to ensure that they have put in at least as much effort (to produce the code or to prepare for the theses defence) as the reviewer will put in. #ai #softwareengineering

    Opinion
  3. @agowa338@chaos.social

    Today I tasked #Qwen with auditing the code it wrote. And it actually found a bunch of issues. (obviously not sure how much it missed). But I also don't know if it is funny or sad that the very same model that wrote the code can outline a bunch of security issues, including: * Token leak * Path traversal * Thread-safety * Caching issues * Missing limits on response sizes * XML injection in the nix module it wrote .... #AI #LLM

    Opinion
  4. @matdevdug@c.im

    As lukewarm as I am on #LLM, it has been interesting talking to friends about the death of #SAAS at their jobs. If a vendor raises their prices, there is now a real risk that a customer will simply use an LLM to copy the six features they want from that software and stop paying them all together. One took a Jira clone, added the two features they needed that the OSS version didn’t have and then bam end of contract. Took less than a week of work.

    Opinion
  5. The claimed solution to the Riemann Hypothesis by AI agents is peak LLM, in that it's added nothing to our understanding. Oh, and it used a symbolic reasoning system to test its working along the way. Kind of like saying an LLM beat a top player at chess using Stockfish to check each move.

    Opinion
  6. @masha@hoosier.social

    now imagine being a lawyer, where most of your job is based on legally tried and true boilerplate, except supervisors are asking you to rewrite boilerplate by plugging it into a model as a "template". and then they get frustrated when the output isn't close to identical. and the legal tech companies that piloted fill in form technology that makes "go-by" use more efficient are all using ai. *rests head softly on work laptop* my life would be so much easier if I also didn't understand LLMs

    Opinion
  7. @jyasskin@hachyderm.io

    CEOs go off the deep end when they stop listening to their line workers, who actually know the underlying reality of the business. #LLMs offer everyone the chance to lose touch with reality in the same way, with workers that are programmed never to unionize or even to raise concerns unless you remember to explicitly prompt them to do so.

    Opinion
  8. @marocddot@hachyderm.io

    Reading someone's code is like reading their thought processes. It's communication. It's knowing when to put in comments and when not to. It's putting fact and correctness into context. It's weaving abstractions that are enlightening (and even sometimes frustrating). Reading LLM produced code is dull and barren.

    Opinion
  9. The fact blind people are turning to LLMs to make screen reader friendly apps/software should tell you a whole lot about how the tech industry is/was unwilling to take accessibility in development seriously. This is a result of the tech industry at large. I'll continue to speak out against the harm these LLM companies continue to do, however, because I care about intersectionality. I'd love to use non-vibe coded accessible software. Unfortunately, that's not possible. On the popular NO AI places, the majority aren't even keyboard accessible. I am stuck and wish I could choose. To make the hypocrisy call out easier for you, sightlessscribbles.com/tools/ #AI

    OpinionPopular today
  10. @vollink@wandering.shop

    Most of what I hate about LLMs/"AI" is the humans that overestimate its abilities and push it to make decisions. Humans need to make those decisions, always. The rest of my hate is the training thievery and mega-data-center environmental impact, but even that is due to humans that overestimate its utility.

    Opinion
  11. And I'm looking forward to going to my first software conference in a year where I'll be giving a talk about making software tools be more AI friendly, sitting in on other AI related talks, and probably commiserating with other devs about the centrality of AI in our profession.

  12. @robincafolla@mastodon.ie

    Has any tech made the world so much worse, so quickly, as LLMs? I can't think of anything else that has ruined art, sped up ecological destruction, made your job worse, suppressed wages, made it impossible to buy anything online (including games), increased the number of cyber attacks, destroyed our privacy, crashed the economy, and made fascists richer.

    Opinion
  13. @osma@sigmoid.social

    It's just stupid to use huge LLMs for everything. A tiny (2B) decision model was as good as Claude Sonnet 4.6 for labeling tickets, saving 98% on costs and giving much faster responses. Obviously it needed a lot less computing resources too. strandsagents.com/blog/amazon-…

    Opinion
  14. Question to my bubble: I myself avoid LLMs due to being unethical. But as Mill maintainer, sometimes I accept PRs that are at least co-coded by LLMs. I review them, request changes, and typically cut the commit messages (with a squash commit) to some useful and digestible size. That also means, I'm the one who finally adds these LLM output to our code base. What do think? Is that wrong? Does it spoil Mill for some of you? #LLM #AskFedi

  15. @_noelamac_@spore.social

    increasingly i hear people who initially show some signs of understanding that LLM's make mistakes but then they say things like "therefore i prompt my ai to ensure it doesn't hallucinate" or "to double check its results" or "provide me only with information which is right"... *sigh* it not only shows these people's lack understanding of how these models work but that they attribute ai the capability of actually being able to know when it's wrong. #ai #llm #fuck_ai

    Opinion
  16. @epmatsw@hachyderm.io

    Spent my Friday having fable get a small LLM running on my childhood N64. A blistering 0.5 t/s!

  17. If this makes you feel shut out from a community that doesn't like and so doesn't use AI then you can either decide AI use is more important to you than being part of that community and keep using it or you can decide being part of that community is worth more than the small gains you get from using AI. Those are your choices Trying to "what if" me personally down the slippery slope is not a choice you get, actually. I am simply not interested in moving the goalposts, sorry

    Opinion
  18. @mhoye@cosocial.ca

    Honestly this is the reason crypto was forced to matter and why LLMs are being forced to matter: if they’re out of the picture, twelve year old computers are fine for human use, totally up to any task that isn’t purpose built to be a computation sponge.

    OpinionPopular todayprobably Tech
  19. Got another one.

  20. @timbray@cosocial.ca

    Another reason why you might want to do your LLM-ing on your own hardware. My spidey senses are all screaming at me that that’s a big part of the bubble-rubble future.

    Opinion
  21. @yantor3d@mas.to

    all those ads all those bots all those trackers all those LLMs exist somewhere on a hard drive which means they can be destroyed

    Opinion
  22. @nathandyer@hachyderm.io

    This is totally tenable, things are great.

    Opinion
  23. @benfulton@fosstodon.org

    This is as insane as it is predictable. I wonder if you could develop a system where two single people join a chatroom with an LLM as a sort of icebreaker or therapist.

    Opinion

More posts from the same people

  1. And I'm not coming at this from a "I'll never use AI, it's evil" stance. I use AI at work a fair bit. But I use it in a way that I still own, deeply understand, and am able to explain and modify anything that I produce - whether I built it with or without AI. Around me I see work being pushed where the author can't answer a few questions about it, it is clearly beyond their usual ability, and has no chance of being maintainable workout AI. This is just lazy folks using AI lazily. I can handle this (at a cost). They don't understand it but I can, and I can make it simpler and maintainable. What really has me reconsidering my approach is the Software Factory approach I am seeing being adopted by some of the smarter people I know. They are actively making a choice that they don't need to know how the system being produced works in details. As long as it produces desired behaviour/output, it is all that matters. #ai #softwareengineering

    Opinion
  2. Unlike the slop that I see lazy engineers produce with AI, the software factory model makes sense to me. It also feels something that could be durable and is systemic enough to be widely adopted by organisations (while the slop coding approach is largely individual driven). Which is making me question the future of my profession. If a software factory model is the future, then most of my skills from this career are redundant. On the other hand, the skills I carried over from my previous career are suddenly a lot more valuable - systems thinking, cross functional communication and coordination, discovery, product thinking, delivery, etc

    Opinion
  3. @futurebird@sauropods.win

    But looking into it? I think some of the people who fear "recursive self improvement" are sincere. I just can't understand how they could fixate on the one kind of harm this new software hasn't demonstrated an ability to do. What about the people who have been harmed psychologically by chatbots? The financial malpractice? The buggy software? The ruined internet? Why can we focus on concrete harm and not... this fantasy. Listen.

    Opinion
  4. @matdevdug@c.im

    Tech conferences post the rise of LLMs have really transformed into emotional support groups. It’s a group of people who are both arguing in favor of their replacements and then promising the “human element still matters”. Like the idea that I’m going to make a “dark factory” of software and somehow at the end will keep my job is laughable. #LLM

    OpinionPopular today
  5. Sitting idle drinking tea and thinking about the thing rocking my profession. I'm kind of down to this question: > Is it important that a human is able to understand, modify and explain the behaviour of a system/software without the aid of an LLM? I'm on the side of yes, in a profession that's increasingly behaving like it's not. I've always needed to understand why before making a choice. I'm not yet getting a strong answer to why we're moving from yes to no. #ai #llm #softwareengineering

    Opinion
  6. This is talking about visual clasifiers but... No shit? The fact it takes scientists to tell us that an LLM isn't a brain, nor recognizes like one, is just 100 levels of sad. I can't believe actual research time was dedicated to this but here we are, I guess. The Tech Bros hate science anyway so, shrug? sciencedaily.com/releases/2026… #FuckAI

    Opinion