RE: https://hachyderm.io/@alter_kaker/117396582238438361

If LLMs live up to the loudly repeated promises of being AI and all of us intellectual workers (and soon after that manual workers too) are replaced by statistical models of our work, our best bet for leaving anything behind before starving might be publishing our private data for training as well, so the interpolating autocomplete's output resembles us infinitesimally more.

#antiAI #noAI #ai #ia #llm #death #famine #theft #consent #memory #microfiction #solarpunk #cyberpunk #mind #simulation

RE: https://neuromatch.social/@jonny/117352647701407496

Why even try to play nice in their game and go through responsible disclosure?

These companies stole years and decades of people's labor. We should be hurting them whenever we can. We should have a network of trusted anti"AI" people who will use this to burn as many token as we can get away with, anonymously of course.

#antiAI #noAI #ai #ml #llm #resistance #punishment #justice #civilDisobedience #abuse #consent #reciprocity #autocomplete

RE: https://mastodon.gamedev.place/@aeva/117413312861881267

"AI" (LLMs) are just a way for the rich to have their cake and eat it to.

Look, either they are just statistical models of labor and then they are simple copyright infringement but not sentient so the tech could be used for ethically if created consensually.

Or, there is some emergent property which makes them sentient but then using them is slavery no matter how they were created.

#consent #sentience #intelligence #philosophy #llms #ai #noAI #copyright #labor

Only 15% of devs avoid LLMs ("AI") for moral/ethical reasons. https://survey.stackoverflow.co/2026/ai/data/ai-avoid

I always wanted to know what percentage of people were actually good (= take a stand against injustice even if there's no personal gain) as opposed to bad (harm others for gain or pleasure - e.g. anti-socials) or neutral (the rest/majority).

I guess this is one data point.

#noAI #ai #ml #llm #copyright #theft #morality #ethics #foss #development #stackOverflow #survey #ip #ownership #anarchism #engineering

RE: https://mas.to/@statsguy/117274814338546184

I feel this way too but TBH it's more about being aligned to following instructions over rules rather than doing what's right over rules.

Because an AI (LLM or actual AI) perfectly aligned to morality is *terrifying*. But of course, I'd just have to be a slightly better person. To the rich who make money not from personal work but taking a cut from other people's work (the owner class), it's an existential risk.

#AI #LLM #Capitalism #AntiAI #AntiWork #alignment #inequality #ethics #consent

Devil’s Dictionary of Programming:

configurable — It’s your job to make it usable.

dsl — A domain specific language, where code is written in one language and errors are given in another.

disrupt — To overcome any legal, social, or moral barrier to profit.

This was written 13 years ago and is more relevant than ever.

Each new generation re-learns the lessons of the past one, poorly and too late.

https://programmingisterrible.com/post/65781074112/devils-dictionary-of-programming

#programming #development #unix #linux #lisp #DSL #LLM #ethics

Stop calling it AI.

No matter whether it satisfies some definition of AI or not.

When ordinary people hear AI, they have an unstated expectation of something that gives correct, complete and unbiased answers.

If it fails one condition (let alone all 3 like LLMs) it should not be called AI in front of the general public.

#AI #AntiAI #LLM #ML #alignment #InterpolatingAutocomplete #takeover #inequality #freedom #power

If your program asks the user to verify a fingerprint (of a key, certificate, ...), at least TELL THEM WHICH HASH ALGO.

Better yet, give them exact instructions how to calculate the hash so they don't have to google around.

I am so tired of spending at least 10 minutes figuring out how to get the hash of a remote host every time there's a random key change, only to find out the hashes don't even have the same length.

#CyberSecurity #InfoSec #UX #UI #Cryptography #SysAdmin #Security #DevOps

LLM ("AI") task-completion horizons keep doubling every 5-7 months.

Corollary: The point in time when LLMs ("AI") can disempower/subjugate/eradicate humanity with minimal oversight comes *before* the point when it can do so with zero oversight.

We shouldn't be worried about LLM ("AI") takeover, we should be worried about takeover by people who own LLMs ("AI") or can afford enough tokens.

#ai #InterpolatingAutocomplete #ml #llm #alignment #capitalism #inequality #power #AiSafety #antiAI

TinyFlaccid emailed me trying to convince me to use "Codex".

The email ended with this:

> We can't wait to see what you build.

Yes, TinyFlacid, I am sure you would love that so you can incorporate even more of my work in your non-free lossily compressed model of stolen text and labor and do it before I even publish it.

#ai #llm #LaborTheft #gnu #freedom #copyright #capitalism

"This is of course a general principle when somebody is taunting you. It means they want you to fight, which means they expect to benefit from you trying."

https://www.lesswrong.com/posts/5CfBDiQNg9upfipWk/only-law-can-prevent-extinction

This is of course a simplification. Violence has 3 dimensions - legality, morality and practicality - which are independent. Eliezer concluded that at this time individual small-scale violence is counterproductive so he appeals to legality and morality to dissuade...

#ai #llm #violence #EatTheRich #morality

'Our participation in the introduction of this class of stochastic systems into the hearts of our central political, social and cognitive infrastructures is limited to debating a bit about the how of “AI”, about the “ethics” and maybe “best practices“. About creating narratives to legitimize the introduction and cushion the narrative of the inevitability of “AI”.

But we don’t get to say if at all. “No” is not an option.'

https://tante.cc/2026/04/21/ai-as-a-fascist-artifact/

#AI #LLM #fascism #consent #LaborTheft

CEOs of LLM companies all think "ideas" should be copied/remixed without credit or compensation. (And that pushing _expressions_ of ideas through matmul should strip copyright.)

Ironic, because what do they *do* all day? It's not like they do any work which could not be automated by an LLM.

The difference is they also own the companies while the people doing the work own nothing. If all intellectual work is automated, they'll keep the capital and we'll... have to eat them

#ai #llm #LaborTheft

'AI is an ideological project to shift authority and autonomy away from individuals, towards centralized structures of power. Projects that claim to “democratize” AI routinely conflate “democratization” with “commodification”.'

-- Ali Alkhatib

Bonus points for teaching me a new word (commodification vs commoditization)

#ai #llm #labortheft

Since software licenses are meaningless now (e.g. malus.sh), are there any remaining ways to legally prevent LLM training?

Can employees "liberate" an employer's code using LLMs? I am certain it wouldn't fly during work hours ("time theft") but what if you access the code after work? If that wouldn't fly either, could open source projects setup a pseudo-employer-employee relationship before allowing code access to hold people liable for stealing the work?

#llm #oss #opensource #ai #theft

"I cannot believe the amount of people who think that eventually an AI will come and do their job for them while they collect the salary. This is the delusion of someone who thinks salaries come from the salary fairy. Your salary actually comes from your boss, who will see this arrangement and quickly cut out the middleman (that’s you). Whatever AI you can buy, your boss can buy too. In fact, they can probably even get a bulk discount."

https://geohot.github.io/blog/jekyll/update/2026/03/21/democracy-liability.html

#LLM #AI #freedom

Spent the past few weeks splitting RecWars into a client and server which means a massive rafactoring: `18 files changed, 2465 insertions(+), 1720 deletions(-).`

Obviously many things that worked in singleplayer are completely broken for now...

#rustlang #gamedev #foss