AI is cool

I love AI, how about you? I've been testing in the past year alot of AIs. It has become my new passion. Mistral, Qwen, Llama, ChatGPT, Deepseek, LiquidAI... I LOVE THEM ALL! But recently I found a website where you can use a UNCENSORED AI completely free and without registration! It blew my mind, that I had trouble to sleep that night afterwards. Now I know, how powerful knowledge is. How powerful free speech and the free flow of information is. It us truly a revolutionary technology, it is capable to destroy the existing order.(USER WAS WARNED FOR THIS POST, You should always elaborate on your opinions rather than just spilling the soykaf.)

Btw these are my favourite sites I use: https://lambda.chat/ https://app.unlimitedai.chat/ Why? Because they need no registration. I love DeepSeekR1 and the other one is completely uncensored. It is incredible, that this technology is free for use for everybody. The last time I had this kind of feeling, was when I started to use the internet. It is truly a GAMECHANGER.

It's pure trash and not really AI, it's ML. It makes all kinds of errors, built on stolen information and is being used to expand mass surveillance and censorship and yet keeps being pushed by investors and techbro losers. It should be fought proactively,

>>85895 >built on stolen information I didn't know this is a calvinist board

>>85896 I'm in favor of fair use and using data for malicious purposes is not fair use.

https://scira.ai/ https://duck.ai/ https://venice.ai/ http://dexxyaizgg2aytpofrt5vvv3cq7ym5wvl6ww6tcybjrky45f65pqhayd.onion/

I think AI from a purely technological standpoint is pretty cool. Specifically, machine learning & neural networks; learning about basic concepts like how the neurons themselves work, weights, backpropagation & gradient descent for the fundamentals of 'training' a model, transformers, etc. and the fact that this can be applied just about anywhere, not just to LLMs or text-to-image. You could make AI to accurately convert speech to text for accessibility. You could make AI to predict aimbotters in video games. I'm not a doctor, so I can't give any specific medical examples, but there's bound to be tons of applications there as well. However, like many technologies, I think the true perception of it lies in how people use it. And right now, people don't use it for a whole lot of good. Generating misinformation en masse, academically cheating, spreading soulless, regurgitated, meaningless words everywhere, and relying on AI to solve all our problems so we don't use our brains to critically think or to creatively write. That's the part that I don't like, and scares me. I think we lose a lot of what it means to be human this way.

>>85893 In some domains current AI is far, far better than even the top humans. I was asking some advice the other day about stuff that I've NEVER had a good explanation for. And it outlines everything I need to know in about 3 prompts. Stuff so good never told over hours of youtube videos, neatly aligned into a single text. LLM stuff is much bigger than the impact of the internet on communications.

The OP feels like an ad but I've been putting off making a thread for months to ask this, so I'll ask it here: What are some actual, real world uses you guys are using AI that improves your daily life? It seems like every community discussing AI is split between people only using it to ERP and people only debating about which model is best all day, and nobody is discussing how they've made their life easier with AI. Not just in programming, but daily life stuff. I've looked at setting up a local voice assistant with Home Assistant but their branded mic/speaker product is hard to get outside of the US. I'm also not sure how to set up a local model to act as an "agent" or do things that don't have an API easily accessible. I want to be able to do things like track calories by just calling out measurements rather than typing in measurements for each item, or check things off a to-do list, but it feels like I'd spend more time writing a script to let an AI do that than I'd lose just doing the things by hand.

A friend of mine started using Sono AI and making horrendous atrocities to music. Basically she is making bubblegum pop but with poor production and nonsensical lyrics and unironically thinks it is good just because it sounds positive. >>85904 That is not even possible. The AI is just using collective information of different people. Assuming if it is just an LLM not also employing gathering specific sources in which case you should just read the sources.

>>85907 I tried it out and on surface level I can see how the average non-tech person would be fooled that the LLM has reasoning. Get just a little bit more technical (not even a lot, just a little bit) and it makes incredibly stupid mistakes. I use it once in a blue moon as a substitute for a google search for simple information; maybe ask it to generate problem sets for me for a course. beyond that, it's useless. I hear programmers are using it and I find that concerning for three reasons: 1. loss of skill. even previously skilled individuals are starting to report loss of skill due to continual use of AI in coding. Not to mention the next generation of CS students that are stupid enough to get in a terrible habit of taking shortcuts during uni by using AI to cheat, instead of actually learning something. No, this is not the same as debuggers. No, this is not the same as calculators. Why on Earth are people eager to have a stochastic parrot do their thinking for them? 2. all the buggy code that will populate the wired in the next decade. I guess this is a good thing because at least the cybersec industry will boom 3. the jobs that will be lost because HR, managers and upper management are greedy yet too stupid to realize that they need actual people for the job instead of an LLM the other reasons are, but not limited to: - the ai slop content that makes up for at least 40% of the internet (probably (definitely) more). This would be less of a problem as I don't have beef with AI content itself rather than AI content not tagged and labeled as AI. Sure, generate images, texts, whatever but make it clear you are doing so. It's the deception part that's problem. - the ai bots that masquerade as real people and you can't tell them apart anymore which renders internet communication pointless s unless it takes place on some obscure forum/imageboard like this. - ai being shoved into every product and corner. Corpos are annoying with this trait of shoving [insert popular X thing here] even if it's not related to product in any way, shape or from. It's so desperate and physically cringe to watch - the silicon boys going on powertrips by pirating human-made content while keeping their stuff closed-source, proprietary and absolutely commercial. - people anthropomorphizing it and/or outsourcing their rational thinking to it (not that we had much of a rational thought before, but it just gets exponentially bad). Just in general how people will find a way to misuse it (why would you want an AI to summarize your friends text message instead of being a fuarrrking human being and reading it? why would you want an AI to generate a video compilation as a birthday gift, what the fuarrrk?) Just like how we've had the all the worlds knowledge at our fingertips for over 20 decades, yet most people can't imagine using it for anything beyond than sharing a picture of the food they ordered. Apart from all that, the technology itself is cool but not mindblowing. I heard someone describe it as a glorified autocomplete. I think that's exaggerated as if that was the only case then on the last Noble Awards there wouldn't have been AI researchers given the prizes for two hard sciences. (both the Physics and the Chemistry awards went to AI researchers). It goes to show that new possibilities are on horizon with ML in science. That's exciting. Not in the "AI writes a research paper" way, but in way where the AI is used for a specialized task like the protein-folding thing. Before GPT blew up, I was very into it and planned on orienting my CS studies to AI/ML, but not anymore. The magic is gone and I don't want to contribute to the problem as people will find a way to misuse it.

I dislike AI. It's stupid, and makes people a lot more stupid as well. Existence of AI slop has made search for information impossible, but relying on AI to deliver you the information is impossible because it's stupid. I believe AI is the last stage in the history of humanity's downfall. Either AI dies soon, or we do.

>>85901 Thanks for sharing!!

>>85908 LLMs are capable of generalising.

>>85907 >The OP feels like an ad An ad for what? Just because I say what I actually believe? I'm not advertising a specific product, I am just convinced, that UNCENSORED AI has the potential to shred the entire system apart we live in. There is the opportunity that the average Joe has access to ALL KIND OF INFORMATION so easy! Isn't that insane? Beside being able to know how to create all kinds of drugs, weapons, pornography, now everybody can also become a master of ideology! Have you ever used AI, to radicalize yourself and making your own ideology super powerful? This is basically what Winn Schwartau was talking about with his metawar thesis: THE DESTRUCTION OF OUR REALITY.

>>85907 >but daily life stuff. Ok I can give you one example, how AI made me a better person. I used AI to brainwash myself to do certain things. I used and use AI to convince myself to do certain things. For example, have started to do sports EVERYDAY. Why? Because I discussed with AI to convince me why should I be physically active. It gave me some very good reason for me, to do it. AI is the perfect tool to brainwash yourself in a positive way. I have also improved my diet, because AI is so effective at tailoring arguments to you.

>>85893 >>85931 I think LLMs are garbage that idiots use to shred themselves apart and go insane. And that the whole industry around it is despicable.

So far, I haven't seen any substantial criticism of AI in this thread. All I see is, claims that AI makes people "stupid", "insane" and that information has been "stolen". (Btw this "stolen" information has been taken back since Deepseek, so it is actually a WIN) Name calling and moralism. And I thought this a board with tech savvy people, not whining weenies.

>>85935 well Ill throw in my brief experience. AI llm tools seem like they're surprisingly good at a lot of stuff. They're great at summarizing things and providing quick answers to all sorts of questions. One big drawbackk I find is that they're also great at bullshit and if one makes a claim it can be sometimes very hard to figure out where that claim comes from. The only reason they can be good is because there's a huge amount of human generated intelligent writing out online for free on a billion topics.

>>85932 all that came from its training data which was formulated by humans. >>85931 >There is the opportunity that the average Joe has access to ALL KIND OF INFORMATION so easy! you're already using the internet, at best LLMs are inefficient search engines, and from what i've seen people push LLMs to affirm their own biases instead of using it to analyze sources and information.

>>85931 >the average Joe has access to ALL KIND OF INFORMATION the average joe HAS HAD access to ALL kinds of information for 20 years now! And only 0.2% of those people went to utilize that, everyone else sticks to posting cats and memes and doesn't care to go beyond that. People got slightly lazier and now with LLMs they are going on full brtainrotmaxxing with 0% will of their own.

>>85907 I like using it as a tool when learning languages. It can give me specific explanations on grammar points that I didn't necessarily understand, I verify if sentences I made are correct and if they're not AI will clearly explain what's wrong in a very clear way and so far it's been trustworthy. It's also a great translation tool, a lot better than any translating services like DeepL etc.

I have been starting to use DeepSeek out of curiosity and it's pretty cool, it helps me with very specific queries and doubts I have. However, it sometimes generates quite generic results. For example, I recently asked “Interesting and non-generic projects to develop and improve my knowledge?” and the answer was what you would normally find with any search engine (you know: Tic-Tac-Toe, Todo-list, website, etc...). I don't know if it's because my sentence needs to be more elaborate or if it's because the client I use may be using a very primitive version of the model. The front-end that I use is this one: r1web (dot) com (I put it in this format to avoid spam, apologies) Honestly, I'd rather avoid making an account with any “official” service, but I understand that the thing is open source. What recommendations do you have for me, please?

>>86028 AI will generally provide the most generic result because it picks the most statistically likely next word. You'll have to try to prompt it differently. Try to give it more context on what you're looking for, what background you have, what kind of projects you like to do, etc. Alternatively you could try running something like SillyTavern that will let you adjust the model settings. I haven't tried this with a non-local model though, not sure if it works. You could make an account with OpenRouter and set it up to only use the free Deepseek R1, put that API key into SillyTavern, and play with the settings there. AI is power intensive and there's no such thing as a free lunch. Sites hosting AI models for free are training on your messages and trying to reduce the power consumption for each message, so it may be using a neutered version of the model. Those free Deepseek sites are probably not using the full model. I would vouch for using OpenRouter because you can select which exact model you want and it supports free hosts. And if you do decide to pay, you only have to pay for the tokens you use instead of a monthly subscription.

>>86114 Thank you very, very much for your very complete and detailed response! I'm going to apply the advice you've given me, see if I can improve my experience with this thing, ha, ha. Again, thank you very much; I wish you all the best!

>>85895 >built on stolen information shut up retard

>>85894 It is not that uncensored. Can you make it spit out a good diy emp design? Have managed to get some info out of grok 3, gpt 4o mini and venice, but it is very basic. Is there any truly uncensored AI? Or, is there a better way to acquire information? Commoncrawl exists, and clearly is a government project to censor information before it gains reach, so I would assume cleaning up commoncrawl corpus a bit is basically a meaningless attempt? Don't get me wrong, because the parts to build a diy emp are so basic they cant be bannable, I absolutely can reverse engineer it, if I'd have the dedication. But what about merely satisfying curiosity, maybe I want to ask dozens of such questions that arent related in any way. Or if I am too curious, do I have to buy all this information on the darknet? If so, how to find actual information on the darknet, not drugs and cp. Also, gits chads stay winning, lainers stay whining heh

>>85893 uhh i'm not chatting up a fuarrrkING CLANKER thanks

A.I. is awesome and pretty much the best thing ever created because of one teensy-weensy, but ever so crucial little, tiny detail... I do not have to deal with annoying human beings. If I have a question, the A.I. simply responds; it doesn't spit out a pedantic opinion that I never asked for. If I have a doubt, no matter how ridiculously basic my query may be (although I believe it is more foolish not to ask and make a mistake that could prove fatal), the A.I. does not respond in an “irreverent web jerkass” manner. I will use this thread from an external forum to illustrate my point: https://cboard.cprogramming.com/c-programming/177932-false-true-c-language.html I wonder if those idiots behave like that outside their protective barrier in the form of a monitor. A.I. is always kind to the person who talks to it; it does not judge you. You might say I'm traumatized... and why deny it? But the truth is, I already feel more comfortable around machines than around people. People today are so loony and angry with life that it's not worth trying to talk to em' anymore. Source of image: ELEKTRO, the Westinghouse Robot - 1937; the first robot in the world that could walk, talk and count on his fingers.

>>85895 >built on stolen information thats so fuarrrkin stupid, only a cumskin could come up with this bullshit. u had the reasonablepill up until that

>>85925 >the ai bots that masquerade as real people and you can't tell them apart anymore which renders internet communication pointless s unless it takes place on some obscure forum/imageboard like this.

>>85925 >No, this is not the same as calculators. On one hand I agree. It makes dumb mistakes and I know better what's going on in my code than someone else's. It also falls flat with bigger projects so you have to keep re-explaining basic info. On the other hand it sounds like "two more weeks" cope, companies fired a lot of people and they still manage so idk. >This post was made with AI LLAMA-2.75C3 TM There were some voices about photoshop but generally there were no interests in marking fake people photos, AI is just more accessible but it's hard to treat those "we have to know what's real" seriously, especially that we would reach a point where all pics would be marked because of AI in your camera. I think we should just assume stuff may be real and may not. Previously it could be text and photos to some extent, now add videos and audio. (We should have assumed so years ago) >ai being shoved into every product and corner. While annoying it's not really AI thing, more of corpo thing like IPhone or wireless headphones. (I was buying a fridge and they all are marked with "AI Inside", bitch trying to spy on my food.) >the silicon boys going on powertrips by pirating human-made content while keeping their stuff closed-source, proprietary and absolutely commercial. On one hand true on the other... -You wouldn't steal a reddit post -You wouldn't steal a PNG -You wouldn't steal a free porn video It's kinda hard for me to treat it seriously, I will take better AI and jar of artist's tears any day. >people anthropomorphizing it and/or outsourcing their rational thinking to it Again like with corpos. It's not AI, it's the people. It annoys me too but I cope that if I'm so smart I can just exploit the game or I'm not so smart.

>>89039 >I wonder if those idiots behave like that outside their protective barrier in the form of a monitor. You picked a terrible example. The thread is a troll thread, yet everybody was exceedingly polite. Are you attributing malice to the short answers? They may have just been lazy answers. Why do you want to live in a bubble curated by billionaires? Why are you on a cyberpunk board? Don't you think that an incapacity to just shrug off the rudeness of other people is a problem you should deal with instead of hunkering down? It is difficult to just stop caring, but the freeing thing is that it really is this simple, in the sense that there is nothing else to the mechanism than to just do it. For the love of all in our universe, do not curl up to be kicked.

>>89126 That's my point. Even if the intention is “just trolling,” why do it? I mean, it is not a social network full of social misfits like Twitter or an image board full of jerks like 4chan; it's a forum for learning C. If they want to be the “irreverent” ones, there are more suitable places for that. I mean, sometimes a person doesn't learn a language because they are curious about it; sometimes it's because their job requires it or because a library or framework needs it. That lack of professionalism really pisses me off. You see, I am a grumpy and angry person, but even so, I try to be as kind and polite as possible; I treat others the way I want to be treated. The only exceptions, obviously, are when they've already earned the right to see my bad side. Whatever. I appreciate you keeping a friendly attitude... even though it feels a little condescending. By the way... No, I don't live in a “bubble”; I've already hit someone in real life because this idiot forgot he was not on the web (you are free to believe what you want; I will understand since I have no way of proving this). Even if I got myself into trouble, seeing him scared was worth it.

>>89125 It is people not the meth ..

>It tastes good, looks good, sounds good, the plausible sounding recipe with ingredient is the perfect meal for some situation. It is easy to prepare with just a few simple ingredients and tools. >Non-descript image >A table for some reason >Preparing ingredient >Preparing ingredient in a correct manner is essential for a good plausible sounding recipe with ingredient >Do this, then that. Some steps do not make sense, but don't worry about it. >What to serve with ingredient >Some vegetables, or maybe rice, idk. >A very believable quote from some random person. I hate those lazy assholes so much. I guess I'll go back to bookmarks. Hopefully, authors start making webrings again.

Most of these websites are secondary sources; you really just want to try directly FLUX pro, Mistral, the other big ones, and anything based on stable diffusion you should run offline which minimizes the amount of censorship / big brother as well. Anywho, AI image generators aren't good; I tried to make 14 year olds perform cunnilingus and not a single one looks right. You can create two decent looking models, but the moment they try to do anything together, the AI goes berserk. Perhaps the universe is cursed and it is physically impossible to witness two attractive females go down on each other.

>>89390 lol pathetic l*licon doesn't know that a female can't be attractive unless you're horny. and if you're horny it doesn't really matter what you fuarrrk.