Setting up a basic newsfeed on gravitywell.xyz and this is a test

It’s interesting how short sighted and surface level some of the AI haters are. Like if you think specific projects on github are bad for using LLMs, just wait till you find out what corporation owns github and what they’re doing with all the data on it.

While they’re hating on RSYNC for using LLMs, I’m questioning why they are okay with any projects still hosting on github at all. Literally everything on github is going to be used to train LLMs and github is one of the largest AI subscription providers already. Even the non-slop on github is being used as training data.

How do these people not also see it as a symptom of a much larger problem created by those corporations that (currently) control the AI industry. LLMs are not the problem here, its the corporations who think they “own” the sum of all the data they stole from us, and that only happens if we let it (we seem to be letting it though, as indicated by the short sightedness around github)

#AI #LLMs #Rsync #Slop

an emu is just an ostrich that's found my chemical romance

Seeing the “threat” of AI taking jobs people keep point too, really begs the question, why are so many demanding to be able to keep working if there is an easier alternative? Its like when slavery was abolished, people wanted to remain as “workers” on plantations, which i guess means we haven’t really come very far since then, but thats an argument for another time.

This is why I think the way we really solve this entire mess is by forcing all these corps who think they own this stuff into bankruptcy, declare all LLMs as public domain, and anything created by them that earns money gets redistributed to fund things that actually help humanity, we put an end to out-dated copyright and intellectual property laws, and end the gate keeping of technology and medicine due to “patents”

This is partly why I think even if you don’t like AI you should just use the free AI offers out from the big corps on occasion to be wasteful. I realize there is also an environmental negative to that, but i think compared to what paid corporate accounts do its really not significant. The goal should be to bankrupt the people funding these corps before they end up somehow turning profitable, because the minute one of them actually starts turning a profit, they will start to monopolize and do everything they can to make everyone dependent on them though a subscription service.

🧵3/3

Question for Anti-AI folks: Would you still be against usage of LLMs if the following conditions had been met:

• Anything made by them is automatically considered public domain and a derivative work.
• Models cited all training data they used, including copyrighted and otherwise unauthorized data.
• The Models could be entirely self hosted, running on a home computer (no data center or $2000 GPU needed)
• Corporations did not own or control any of this, it was just discovered by some random college student or something and released publicly to everyone for free.

It seems to me like the actual problems people have with “AI” are really just problems of venture capitalists having almost complete dominance over the field.

Now I would Imagine in a parallel timeline where such a technology was created and released by something like a random college student or some underground group, it would get the same kind of treatment Napster did, and lets never for get that Aaron Swartz was driven to suicide by the legal system for doing a much much smaller version of the exact some thing these AI Corporations have been doing.

The problem isn’t the technology, its not the tools being made, its the people who are claiming ownership of those tools when they had no right to them in the first place. Either LLMs belong to all of us as public domain, copyright laws need a complete re-thinking, OR none of this is legal and the rules just don’t apply to corporations.

🧵2/3

I consider myself “Anti-AI” but “Pro-LLM”, I’m going to try to explain a bit on what the difference is why i have this position, and you can all feel free to argue or hate on me in the replies:

🧵1/3

Before LLMs i did little more then small simple bash and python scripting and the most basic of HTML or anything else. My “skills” such as they are mainly focus on higher level system and network administration tasks. I don’t consider myself a developer or a coder, but i know enough to make changes under the hood, but not so much that I’d be creating anything of much use outside of my own very specific environment. My usecase for coding in general is basically just automation and/or filtering. I’m either trying to make something i do regularly easier to do in less steps, or I’m trying to parse a large amount of data (ie logs) to find the specific info i need.

This being the main way i use any sort of “code” LLMs end up being a great deal useful for me becasue to start with, you don’t even need very large ones for them to be useful. Instead of having to type out some long complicated command with specific syntax and ordering, I can just say “deploy a copy of project_a to server_b” and it writes the script for me, which i then look over, see it does pretty much exactly what i want and is most likely way better then whatever crude thing i’d come up with. Sometimes I even end up learning a bit because i see how the LLM does things, which it “chooses” because those things are the most likely things to work according to its training.

That all being said, I find LLMs useful and i see no reason not to use them any more then i see a reason to not use swipe typing or auto-complete or spelling/grammar checking. But I don’t consider these things “AI”. What I consider to be AI is the corporate run slop machine that stole all of humanity’s creative works and want to sell it all back to us by the token… I hope those companies fail spectacularly and whatever useful research they did becomes public domain which is exactly what it should have all been in the first place.

It’s probably just my bubble but it really seems to me like the majority of people who are anti-llms for code don’t actually code, while those who do code see it for what it is, (spicy auto-complete).

And stay tuned next for the next hard hitting article “Now that deaths from lung cancer are way down from the 1970s, is it time to rethink whether smoking is really that bad for you?”

We’re so fucked as species ya’ll. These people don’t just exist, they are paid to write this kind of garbage for the NEW YORK TIMES. Like seriously I would not be surprised if in 5-10 years we start seeing a new trend of “organic vaping” or some other stupid bullshit way of trying to bring back smoking.

@MxRemy I wouldnt count KeePassXC out just for using co-pilot to do code reviews, but personally I’m thinking of setting my my own vault-warden instance at this point.

This is still very much a work in progress but i think its coming along nicely. Eventually I want to make an actual flash game where you just shoot billionaires, but that and the “prediction” market are on hold for now while i focus on actual infrastructure things. Meanwhile though some of the links will work and go to things like jellyfin and peertube.

I know im making this way more complicated then it needs to be, but i think once everything is sorted things will be a lot more modular and easy to manage.

Rahul Sidhu’s is the CSO of #Flock “safety” lives in an 11 million dollar home located at: 25.80922531171, -80.1343327890

He think’s he has a right to invade everyone’s privacy and make a profit while doing it. Lets invade his privacy shall we?

It’s still going to be a minute before most stuff is back up with consistency though, the main jellyfin server has an issue i can’t seem to resolve that causes it to freeze once a day or so and requires reformatting.

I recently registered the domain “Throbbing.click” and will be using it as something of an experimental alternative for services and things.

I am so terrible about posting updates on gravitywell and related things. Nothing really feels right for it, I keep thinking i want to try this or that federated thing but Im starting to realize what i really need is for the homepage to have a news widget of some kind.

I’ve been wasting $1000’s of AI corps money. Why should they be the only ones who get to take and give nothing back?

I have over 200 accounts across 20+ providers and I don’t pay for a single one. Sometimes they get rate limited but never enough to completely stop me.

Guys who work in office jobs from home: i need like 10000g of protein or i will literally die

Guy who actually does manual labor: did you know seven elev has a deal where two taquitos free if you get monster energy drink

“People who lose weight on Ozempic are viewed worse than people who don’t lose weight at all”

A lot of the fear-driven coverage around GLP1 (ozempic) meds makes more sense when you look at the economics of the weight-loss industry. They spent years cashing in on useless pills, gimmick diets, and endless programs that failed over and over, while customers kept blaming themselves and coming back for more. Now we have A treatment that actually delivers results and this is a nightmare for the health and fitness grifters.

NY Post Link - - - Archived link

Finally got around to trying out data only mobile service from smspool.net, wish i’d learned about this sooner but i guess the name threw me off and i had not realized they also offered data service in addition to sms verification.