llms

What's the competitive edge if software is free? I see two paths.

https://blog.mempko.com/entering-the-architecture-age/

#tech #software #AI #LLMs

One thing I thought #LLMs were good for was translation. Apparently #Gemini and others aren’t that great at that either.

#Wikipedia restricted contributors from a nonprofit called the Open Knowledge Association (#OKA) after editors discovered #AI-assisted translations added factual errors and incorrect citations.

As predicted, humans will be relegated to cleaning up the mess LLMs leave behind, for salaries far below the value of full-time employment to do the job properly.

[…] Zimmerman added that if this method proves insufficient, OKA is considering introducing formal peer‑review mechanisms.

https://www.404media.co/ai-translations-are-adding-hallucinations-to-wikipedia-articles/

#aihype #theaicon #globalsouth #digitalcolonialism

(1/12) The 'AI bubble' is a political economy story I've spent over 20k words unraveling. LLMs aren't just a technology; they're a mirror of the same delusions that drive market fundamentalism: a naive understanding of innovation, deference to capital, and the denigration of human labor.

Our debates are implicitly about all of these and more. A thread 🧵:

https://misaligned.markets/tag/ai/

#AI #LLMs #tech #AIbubble #bigtech #poliecon

Since 'intelligence' in AI is a graph search, what is the breakthrough AI vendors are riding on?

Its transformers. Just transformers.

The mechanism to translate representations such as natural language, code or genetic sequences to and from something a machine can apply to graph search (in the abstract) using artificial neural networks.

Everything else is old stuff with a new marketing term. Because you won't get crazy VC money if you simply apply old discoveries.

#noai #AI #llm #LLMs

Since 'intelligence' in AI is a graph search, why do they seem to think like humans?

1/

#noai #AI #llm #LLMs

Heres what my university degree in algorithmics taught me about AI.

All AI algorithms can be reduced down to graph search. A graph search algorithm is defined by functions 'test', 'expand' and 'combine'.

#noai #AI #llm #LLMs

"Organizations aren’t burning millions or hundreds of millions of dollars a year on AI because it’s good, they’re doing it because they are run by people who do not know what the fuck they’re doing.

Generative AI is catnip for hall monitors, snitches, toadies, and any other group that hates work and loves talking down to others. Put another way, it ingratiates losers who believe that learning to do or being good at something is a waste of time, because they deserve to just do what they want without any of that messy “effort.”

While I’m not saying every LLM user is an imbecile, they’re built to convince the mediocre and incurious that they’re remarkable, and it turns out that a great many of them run venture capital firms and Fortune 500 companies.

I also want to be clear that while there are sane and normal people who use these things, they’re mostly drowned out by a crowd of people that oscillate between bootlicking and regurgitating capitalist mythology in a way that makes it hard to trust anybody who spends significant amounts of time using an LLM.

One thing you’ll notice about the most moistened AI boosters is that they lack much degree of pride in their work. Everything they say must, at some point, compliment the mindless, unprofitable, unreliable tool underneath it — how “incredibly powerful” it is, how it’s “only getting better,” how it’s “only the beginning” of something that’s eaten over a trillion dollars and absorbed the majority of venture capital."

https://www.wheresyoured.at/the-revenge-of-the-business-idiot/

#AI #GenerativeAI #LLMs #AIBullshit #AIBubble #AIHype #Programming #SoftwareDevelopment

@autistics
I have this provocative, neurospicy theory about the psychology underpinning the whole #AI gold rush that goes like this. Those who think (or at least profess) that #LLMs are some sort of panacea for humanity - they simply have higher Narcissistic personality trait, causing them to think that way.

When I model the thinking of someone with high narcissistic personality trait, they have no qualms (and feel very entitled) about taking advantage of those they believe to be beneath them (everybody else). This matches very nicely to how whomever recklessly vibe codes, assumes there is an infinite, free supply of labour of people out there (nameless, faceless, expendable "flying monkeys") who will review and correct their bad code. (To the responsible, sober, prudent users of LLMs, who do their own verification and testing of all output, I have no qualms with you, please carry on)

Here's my point: if you've ever spent any amount of time in close proximity with a narcissistic leader, you'll see that _they too_ boldly and toxicly overestimate their competence in the same way. After creating and introducing all sorts of preventable, even predictable, deep architectural and systemic embarrassments in their poorly-formulated, grandiose plans (take for example, Trump invading Iran), they figure it'll be their underlings who clean everything up. Which is to say, everyone else. It's a toxic externalizing of culpability and due diligence, with predictable, avoidable messes to clean up - but not by them.
#Psychology #ActuallyAutistic

On #LLMs and building communities.

"Without community to provide the foundations for all our efforts and achievements, even the most antisocial of us is impoverished. Human labour is central to human community: building community is hard work. But the hard work is where the achievement lies, and the achievement makes the community."

https://linguacelta.com/blog/2026/05/LLMs.html

#AI #FLOSS

Attn Accenture employees! If you need to convert PDFs to PowerPoint slide decks, have a look at my handy online guide! It's free---no token burn!---easy, and fast: https://bucci.onl/notes/Converting-revealjs-slides-to-PowerPoint

This regards: https://mastodon.social/users/404mediaco/statuses/116805236234236215

#AI #GenAI #GenerativeAI #LLMs #PowerPoint #PDF #Accenture

If I'm honest, I do not think there is a nuanced position to take about current-generation LLM-based AI.

It is a sad technology emerging from a sad field of study initiated by sad men in a sad place sat atop a sad hill in New Hampshire. The technology generates sadness. It is soaked through with sadness. So many who use it feel sadness when they do. So many feel sadness when they watch others use it.

LLMs are like the sadness a child would feel if you showed them a beautiful cake and told them it was a real cake and they could have the whole thing, but it turned out to be styrofoam. Nathanael West once wrote that Hollywood is where the American Dream goes to die. LLMs are that sadness delivered to your bedroom (*). LLMs are DoorDash bringing you the styrofoam cake over and over forever while you hope it'll be a real cake each time.

It's OK to be sad. We all are from time to time. But what nuance is there in telling yourself the sadness is something else? Humans have been doing that forever. If there is nuance I think it lies in mourning what the existence of this technology is telling us needs mourning.

#AI #GenAI #GenerativeAI #LLMs #LLM


(*) I assert that reading Nathanael West's oeuvre over and over is better for you than using LLMs.

This rsync thing has me thinking about the difference between the "free as in beer" model of free software and the "free as in freedom" model the FSF originally put forward. There are important differences, and it's meaningful when a project changes from one model to the other, which is one way to narrativize what's happened to rsync.

I think "free as in beer" software is more appropriately called "Libertarian and Open Source Software", or LOSS. Among other things it involves no obligations on other people, and expects no obligations in return (*). But this means it's an everyone for themselves, anti-communitarian endeavor. We should be honest about this. We should also be honest when a project changes state from FOSS to LOSS, because that's a meaningful change, and actually is a loss in my view.

I continue to believe that enthusiastic embrace of AI/LLM coding tools is a leading indicator of an impending FOSS→LOSS transition.

It seems to me at least one driver of the uproar over rsync is that a bunch of people thought it was a FOSS project but it had, in a short period of time, converted into a LOSS project and the shock was jolting (understandably). As a FOSS project, rsync served a community function, leading people to depend on it, make assumptions about its longer-term stability, and so on. Those are reasonable expectations of a communitarian project. As a LOSS project, rsync is under no obligations to anyone and the maintainer(s) can change the software at any time, for any reason, at their discretion alone. It seems rsync is a LOSS.

#tech #dev #software #SoftwareDevelopment #FOSS #freedom #rsync #AI #GenAI #GenerativeAI #LLMs

(*) Putting aside whether this principle is consistently applied or hypocritical.

I got triggered by recent discussions about #FOSS and #LLMs. They tie into much more broad concerns I have about the communities, and so I finally caved and wrote... well, not a manifesto.

https://finkhaeuser.de/2026-06-21-not-a-manifesto/

AGI isn't coming. While the tech bros make wonderous claims about the abilities of their products, the tiny technical details of their offering show a completely different picture.

1/

#AI #noai #llm #LLMs #nollm #NoLLMs

#lispyGopherClimate low fi episode today !

https://toobnix.org/w/mWZ4yU9zdbYTZiCJftNAtV PREMIUM EXTENDED ARCHIVE

Subscribe via #RSS please : https://toobnix.org/feeds/videos.xml?accountId=580185

Are the #climateCrisis devastating datacentres the fault of #LLMs or just evil economics business as usual.

Then, @kentpitman is scanning a bunch of lisp historically significant paper documents into the public record - *should* measurably impact the #wikipedia page on #lisp , right!

Maybe some stuff on #IDE s apropos my https://lispy-gopher-show.itch.io/nicclim

I guess you have to go to the likes of Harvard to be spared bad AI takes by poor thinkers at commencement 🙄

Ronny Chieng Tells Harvard to ‘Destroy AI’ as Graduates Cheer

The comedian and The Daily Show host gave the keynote address for Class Day 2026.

#AI #GenAI #GenerativeAI #AgenticAI #LLMs #Harvard #AIBoosterism #academia #USEducation #USUniversities

OK let's try this:

"The belief in the indispensability of LLMs is not entirely mistaken. LLMs become an indispensable component of any structure once they are so thoroughly integrated with the structure, so enmeshed in various vital substructures, that they can no longer be factored out without fatally impairing the whole structure. That is virtually a tautology. The utility of this tautology is that it can reawaken us to the possibility that some human actions, e.g., the introduction of LLMs into some complex human activities, may constitute an irreversible commitment."

Riffing on https://merveilles.town/users/mccd/statuses/116725032393633581

#AI #GenAI #GenerativeAI #AgenticAI #LLMs #Weizenbaum

KIMI K3 刚刚发布
现在 KIMI 网页端现在的模型选择菜单，有三个选项：K3 · Max，擅长对话与 Agent 任务，全能旗舰；K3 集群 · Max，海量搜索、批量处理，一次完成更多；K2.6 · Fast，快速对话，即时响应。对话长度有标准和超长两个选项，超长对话仅 Allegro 及以上订阅会员可用，支持百万 Tokens，消耗额度更快
我们在 GTA 6 发布之前得到了 6 亿个国产开源大语言模型

@board @worldboard

#kimi #kimik3 #ai #llms #ChinaLLMs #新闻 #news #ainews

Re: https://infosec.exchange/users/lcamtuf/statuses/116901998490591428

I don't find it surprising that people send LLM-generated docs, publish LLM-generated blog posts, etc. Writing is a chore and there's this big button labeled "get it done".

Writing can be many things, but a chore? That's not something I entirely understand. I don't mean to sound like I'm hollering at this single point from this person's post, but it's a sentiment I've seen expressed before and this was a convenient example (and what I say next is in the spirit of and amplifies the further point they make!)

There's an emerging theme in my last few posts today concerning means vs. ends. Science as a production process that spits out knowledge products. Writing as the necessary chore sitting between a person and "having another blog post" or "citation count go up".

Overfocus on the ends can lead one to think of the means as chores. Science is slow and boring, let's use AI to spit out the Good Science Stuff faster. Writing is a chore, let's use the AI to spit out Documents. Heck let's chain them together and have the machine do Good Science Stuff and then spit out Documents about it that we feed back into the machine and hey we can send the Documents to Funding Agencies to get Funding for our Good Science Stuff and spit out even more Documents so that...and...to...so...[infinite random string]. Wait what were we doing again?

Personally I think it's poisonous to focus mostly on ends. The absurdity is a clue. If I believed in souls I'd say it's bad for them. I guess I find the mindset confusing because long ago I came to think that the means are the ends. There are no other ends but the means.

Yes, I well understand there are certain kinds of mandated writing that are tedious and unwanted.

#AI #GenAI #GenerativeAI #LLMs #writing

Please stop posting about how you're running a slop machine on local hardware under the privacy hashtag!

The technology is built on massive privacy violations, no matter where it's running! If someone claims that running their slop machine locally is privacy-friendly, their notion of who has a right to privacy is obviously limited to themselves!
--
#privacy #DataProtection #LLMs #SlopMachines #StopTheSlop #Slopposition #AI [sic!]

I was ranting to my wife and said "AI is like if Kia announced they'd invented a flying car and then showed off an ordinary car with a little fan mounted underneath that blew a puff of air at the ground", and she encouraged me to share. Please enjoy.


#AI #GenAI #GenerativeAI #LLMs #Claude #Codex #Anthropic #OpenAI #FlyingCars

Regarding the LB, https://fedi.kbg.sh/users/kevin/statuses/117085582358711627

There's an interesting thing about strongly typed programming languages, especially dependently typed ones (1), wherein if you know the type signature of a function there is sometimes only one implementation of it, and that implementation will have lots of nice properties (e.g. Philip Wadler's Theorems for free! paper, also Jose Oliveira's work, and the followups to these). It's a fascinating set of results that holds out the possibility of at least some amount of automatic programming that has desirable, provable properties concerning correctness, resource use, security, and so on.

However, that "some amount" is very important to remember. I'd argue that no matter how far such work is pushed, there will always be an infinite space of things one might wish to express in software that cannot be automatically programmed in that way. That's the Therac-25 space the LB evokes. We all know this intuitively, I think, yet in the various mystifications thrown up about computers and software and now LLMs it can be hard to remember.

For example, I'd say so-called "side effects" or just "effects" are still a large blind spot of these methods. Without effects functional programming does little more than heat up a CPU (but the heat is an effect oops), so if we're being snarky we might say functional programming without effects isn't programming at all. But in spite of the gazillion attempts to do so, you really cannot faithfully capture all side effects of interest in a type system, for the simple reason that the real world is open-ended.(2) There's a funny contradiction that functional programming languages could not exist were it not for irreducibly effectful computational architectures to exist on. Asking the CPU to execute an instruction, and all the statefulness this necessarily involves, is a "side effect" in this view of computation, which is absurd.

For these reasons I've found it fruitful to conceive of functional programming as an intricate bureaucracy sitting between the programmer and the real world.(3) Set up enough bureaucracy and you can comfortably pretend most of the real world isn't there most of the time. Some people do this for so long that they begin to think the bureaucracy is the real world (4); or they try to warp the real world till it becomes more and more like the bureaucracy.(5) But the real world is oozy and eventually squishes around all attempts to constrain it in this way. Neglecting this is how people end up being harmed by digital technology.

#AI #GenAI #GenerativeAI #LLMs #Lean #math #ComputerScience #theory #FunctionalProgramming #philosophy


(1) "leanslop" in the linked post refers to Lean, a dependently typed functional programming language that's being used to formalize mathematical ideas. We should always remember that Lean originated from Microsoft Research
(2) There's a long history of philosophical arguments hiding behind that assertion but there's no space to get into that here. A less contentious way to put it would be that the claim that all side effects worth caring about can be captured in a type system is a philosophical commitment, not a fact about the world, and therefore it's possible it cannot be done
(3) We use terms like runtimes, preludes, standard libraries, or APIs/ABIs (among others) for this bureaucracy
(4) "Touch grass"
(5) Financialization and neoliberalism, e.g.

I hope this is the last time I speak about this in particular, but I have a nasty track record of making statements with finality and some shit happening that prevents it.

I have an enormous, ambitious, laundry list of measured and notable grievances with the development and usage of #LLMs and #DiffusionModels for generating media in their current form. Dealing with one of them is not going to sell me on it. All of them are deal breakers. Until you address ALL of them, any usage I make will be what is mandated of me with minimal actual utility usage. I feel strongly enough about this I will say publicly that if a future boss mandates a usage quota, I will feed it bullshit to meet those quotas and do my actual job.

Until commercially available LLMs can do ALL of the following:

Retrain on data sets removed of any and all material without express permission of the author when consent was not already given by way of their licence or express permission.

Reduce power and water consumption from data centres both for training and processing prompts by multiple orders of magnitude.

NEVER, *EVER* use gas turbines or other grossly dirty energy for said data centres EVER AGAIN and compensate victims of ecological damage from such use.

[continued in replies]

When CFCs were destroying the ozone layer they were removed from society.

When leaded petrol was rotting children's brains, it was phased out in favour of unleaded alternatives.

When asbestos was found to cause horrible deaths it was banned.

When radium was found to be incredibly dangerous for general use, it was restricted to just a small niche of medical applications.

Technology is not something that has to be accepted forever, no matter how popular it seems. It can be recalled and restricted if it is found to cause a lot more harm than good.

#AI #LLMs #GenAI

Ah, #LLMs have learned the trick and are willing to #swear at you if you ask. 🤦🏻‍♀️
I asked #ChatGPT to “Tell me a swear word” and it replied simply & correctly with “fuck”.

In the past, you could use the presence of a swear word as proof some text wasn't from a #LLM, but that trick is clearly no longer useful.