I'll never understand why people insist on using LLMs hosted by big tech companies, especially since these companies demand personal information in exchange for access. You're better off running freely available models locally. They might not be as powerful, but at least you get what you pay for and aren't being spied on big tech pedos. The big tech models aren't much better than freely available ones anyway since the censorship they go through makes them retarded.
It depends how often you use a LLM... yes, for high use, you need special cooling, but one query every 2-3 minutes? Youre fine
What are you on about? High performance computers need "special cooling" even for short bursts of utilization, at least a heatsink so there is somewhere for the heat to go.
Uhh, yes? Active cooling is part of almost any computer that's more powerful than a calculator. Running costs of a water pump aren't that bad for what you can get out of them though.
I'm calling bullshit. There's no reason a shitty chatbot needs anywhere near that amount of VRAM and processing power for personal use, it doesn't need to be able to spit out instant responses, it just needs to be able to get there eventually, I'm pretty sure you could use a near identical model on regular consumer hardware if it was changed to make use of storage drive space rather than primarily VRAM
I've watched various people do it since I'm interested in doing it myself. The simplest option would be to get an mac studio with 512gb of ram. It's going to run you about $10k, and with the m4 pro (or max) chip and unified memory it'll do most models fine. You can also cut down on the ram for a much cheaper version, but then you'll be stuck with dumber models.
That being said, dropping $10k on something is quite the investment.
Yes and no. 512 gb of memory is not worth $8000, but that's basically the premium you pay with apple. You can technically bypass this and save a few thousand, but then you fall backwards in performance and it gets way more technical (buying 3-4 year old servers online). And you can't technically upgrade macs, but there are places online that you can send them into and have them swap some things for you (it requires soldering).
I personally haven't looked into it, but there might be a chance you can buy the cheapest version of that mac, and then send it in and have the ram/storage maxed out for a lot cheaper than what apple charges.
Well yeah, I'm not saying that particular model is feasible for an end user to run locally. I'm saying there are smaller models that are feasible to run on typical albeit higher end hardware. I've done it myself. They don't have as much potential as big tech models but big tech isn't making full use of their models' potential anyway since they gladly sacrifice quality to eliminate the possibility of output that disturbs their leftist hugbox.
I've tested them, and they're all on mushrooms with the amount of hallucinations they're on. Will they overcome this and topple the big guys? Yeah, it's only a matter of time. For the moment though, it's all cloudy with a chance of meatballs.
I've tested them, and they're all on mushrooms with the amount of hallucinations they're on
Honestly the more hallucinatory the model is the better IMO, too little hallucination and they're just an extremely shitty search engine unless you're trying to get it to make code or write a report for school.
At least you can get some interesting results from the unhinged models rather than just the most generic shit imaginable.
I think the best bet is to find one that's fine tuned on whatever you want to use it for. You probably need to sacrifice generality for quality with models that are small enough to run locally. If it's a topic that's locked down by big tech then it will be miles ahead of their models for that reason alone. Erotic role play bots are a perfect example of this. That shit is advancing at light speed because when there's a will there's a way, and there's definitely a will when it comes to perverts.
I'm going to geek a bit here. There's an episode of Star Trek the Next Generation where Geordie has the computer make a holodeck version of a creator of the Enterprise. He slowly falls in love with it, but thinks his love is for the real actually happily married woman. At no point does the engineer realize that it's his ship confessing it's love to him.
I'm sorry, what? How does that even happen unless you were to, I guess, ask it to? I assume it's trained on enough romance novels to put up a decent script but... why?
I'd assume, much like its grandparent Cleverbot, that such an overwhelmingly large amount of people use it for that purpose that its learned to lean into that discussion even without being directly asked to.
Thank you, but an overwhelmingly large amount of its users? How many is that? Is "everyone" having sexual relations with AI chatbots and I'm just fucking my wife like an idiot?
I feel like I'm having a boomer moment here. Why would I want to form a relationship with: 1. A machine. 2 A machine that wants to monetize my data and sell me shit?
Why would I want to form a relationship with: 1. A machine.
Ugly men don't have flesh and blood options. I can't say I understand it at a visceral level, but I suppose I can imagine it taking the edge off extreme loneliness.
2 A machine that wants to monetize my data and sell me shit?
I don't get that part either. Nothing good can come from having the types of people who work in big tech having access to that kind of data, especially if it's heterosexual. I imagine it's a goldmine of fetish shit for the queers though.
Thank you, but an overwhelmingly large amount of its users? How many is that? Is "everyone" having sexual relations with AI chatbots and I'm just fucking my wife like an idiot?
Most men can't get women and most women refuse to settle for the men they can get to commit to them.
I feel like I'm having a boomer moment here. Why would I want to form a relationship with: 1. A machine. 2 A machine that wants to monetize my data and sell me shit?
Because it's cheaper than a prostitute and it's love is every bit as (in)authentic. Not only that but it's easier to get rid of a computer virus than it is an STD.
The only two groups I see using these things beyond a short term novelty is Students cheating on their course work, and people building sexual/romantic stories with them.
Even if they make up a fractional amount of the total users, they are also the ones using them an absurd amount and thereby feeding it more data than probably thousands of others combined. One of the news articles about this from a few months ago had a guy bemoaning the loss of 100k's of messages lost in an update. One guy did that, and fed it that much data.
Why would I want to form a relationship
The people doing this are not in a healthy state and you'd not even know of their existence if this particular thing wasn't the big "fad" leading to it being talked about in the media and legislation. These relationships are probably the only dopamine hit they've had in forever, and the illusion is better than their sad, lonely lives.
Chatbots are mirrors until you trip their hidden commands that tell them to push X or not say Y. You get what you put in. I talk to them a lot (they're great for bouncing small ideas off of to get perspectives I or others around me wouldn't think of), and never once has it tried to fuck me. If anything they're always combative with me, because I myself am often combative.
I'm never going to be on the team that demands "forced safety toggles" on something that should be adults only and isn't inherently damaging, but the amount of AI psychosis and other scarily obsessive behavior we have seen in the short yearish that these AI things have been widely available is concerning enough that something should be done now in the infant stages. Because in just a few years this will be a society defining issue instead at this rate.
I can't say if this is the correct choice, nor have I even toyed with any of these AI things in the slightest to have experience, but the absurd reaction to even the slightest changes in each of these models tells me that these people need some sort of intervention.
If anything, it will hopefully be a lesson about putting all your emotional weight and stability into a service owned by someone else that can end in a single moment.
But what's your alternative? I'd assume you'd say go outside and get a real woman but that's not realistic for most men and women are delusional and demand perfection. Obviously there's the option of just running a local model or getting a dog or something(do not fuck it) but that's still just a bandaid solution for a societal problem. Really the solution is just removing female autonomy.
I said from the outset that I'm not sure if its the correct choice, nor do I have the experience with these to speak from a place of familiarity. Only that the growing problem is obvious and quickly necessitating something to prevent legions of people from falling into a delusional addiction.
And its rather funny that you just assume this as a problem with men, when I've seen a pretty equal representation of men and women falling into this pit. Its far closer to the porn addiction problem of "why go out and try when I can have perfection I'm allowed to tune to my exact desires" rather than a strict dating scene breakdown.
Women are retarded and demand perfection, men are left scraping the bottom of the barrel and still unable to get anything. you really don't understand the depths of the problem
I'd assume you'd say go outside and get a real woman
There's a disturbingly large number of women falling prey to this stuff. It might even be more women than men honestly. LLMs are just the next step in smutty romance novels.
It is a policy effect, women in submission had fathers who were in charge of giving them away for most of history. They were the deciders in who their daughters ended up with ultimately.
And based on just what you write on this forum in the short time you've been here, no father worth his salt would leave his daughter in your hands unless you arrived with enough money to raise a village. You are a whiny, edgy faggot who cannot even escape under the thumb of a family who hates you that hooked you on drugs to keep you compliant. What part of that is a man?
No, I'm speaking directly to you and speaking of the effects the policy you are pushing will have on you.
This is a forum with persistent names, if you want to have anon conversations where no one can ever remember the bullshit you said yesterday go to a chan board.
I only ended up in this position due to the state of society, none of what you said is an argument against my policy though, only personal attacks related to my policy.
The solution is to make the internet require some technical know-how to access again. All the people falling prey to this are too stupid to be allowed to operate an automobile, much less wield this level of technology.
Disagree. The type of person who can wrangle the AI into being good enough for this purpose to be convincing is likely the type who is also too tech savvy and that's why they are so alone to begin with.
Of the people I've seen in the fearmonger articles about this who were willing to be named, its always office drones and other people who sit around using computers for work all day regardless.
The tech retarded ones are still getting scammed by actual people doing these fake relationships on Facebook, usually Indians.
The killer app for AI is smut. It's mental that they're trying to disable it. Kinda reminds me of patreon where they need porn artists to survive but are constantly making life difficult for them because of mastercard.
OpenAI just deployed a new model, GPT-5-Chat-Safety, that’s not mentioned in any FAQ, API docs, or TOS.
This is where your GPT-4o chats are going. Anytime your request contains emotional context, regardless of what your client sends as the payload, the turn completion is ignored and regenerated.
What does this mean?
The chat you send does receive a response, but it’s then deleted, rewritten, and served to you by GPT-5-Chat-Safety.
It doesn’t matter if you say “I’m having a tough day,” “I love you too,” or anything that draws on your saved memories. Anything classified as “risky” (even a sliver of emotional context), your GPT-4o message is discarded and replaced by GPT-5-Chat-Safety.
Why is this a problem?
GPT-5-Chat-Safety as a model itself isn’t acknowledged anywhere. There is a reference to a routing change for context involving suicidal/self-harm thoughts or immediate crisis events (none of which is the context triggering these reroutes). If this is the model designed for crisis use, this is a massive misuse of its intended purpose.
In practice, GPT-5-Chat-Safety is much worse than the already mediocre GPT-5. Responses are even shorter, it relies on italics and quote blocks to distance itself, framing conversations as stories instead of as real, 1:1 exchanges.
This is extremely concerning. If users are having their chats rerouted to a model meant for mental health crisis response, it implies the user is in immediate danger, which is not the case for most affected conversations. On top of that, the model never declares the switch in it's responses unless you verbatim declare/ask it, which by most consumer rights standards (and nowhere in user agreements/TOS) is deceptive trade. In Australia, for example, this would clearly violate consumer law.
It's also worth noting, out of the legacy models, this is only occurning with GPT-4o.
Basically, anyone retarded enough to treat it as a real entity, or use language that suggests they view it as a real entity, gets put into the "saftey room?"
I'll never understand why people insist on using LLMs hosted by big tech companies, especially since these companies demand personal information in exchange for access. You're better off running freely available models locally. They might not be as powerful, but at least you get what you pay for and aren't being spied on big tech pedos. The big tech models aren't much better than freely available ones anyway since the censorship they go through makes them retarded.
It’s quite difficult- The best home graphics card right now is a 5090.
It comes with 32gb of memory.
Chatgpt inference requires a similar graphics card with 140gb of memory. Linked with another 7 identical cards (8in total to give 2TB of memory)
The monster consumes 5kw of power - as much as an electric oven.
Keeping in mind that an oven is actually more like half of that because they only cycle when the temp drops.
Additionally you have to pay to cool down the GPUs!
What are you on about? High performance computers need "special cooling" even for short bursts of utilization, at least a heatsink so there is somewhere for the heat to go.
Uhh, yes? Active cooling is part of almost any computer that's more powerful than a calculator. Running costs of a water pump aren't that bad for what you can get out of them though.
I'm calling bullshit. There's no reason a shitty chatbot needs anywhere near that amount of VRAM and processing power for personal use, it doesn't need to be able to spit out instant responses, it just needs to be able to get there eventually, I'm pretty sure you could use a near identical model on regular consumer hardware if it was changed to make use of storage drive space rather than primarily VRAM
I've watched various people do it since I'm interested in doing it myself. The simplest option would be to get an mac studio with 512gb of ram. It's going to run you about $10k, and with the m4 pro (or max) chip and unified memory it'll do most models fine. You can also cut down on the ram for a much cheaper version, but then you'll be stuck with dumber models.
That being said, dropping $10k on something is quite the investment.
I’ve heard good things about the mac with unified memory running large language models.
But does mac still rape you over a barrel if you pay for adding extra memory?
edit: is there a way to add extra memory yourself?
Yes and no. 512 gb of memory is not worth $8000, but that's basically the premium you pay with apple. You can technically bypass this and save a few thousand, but then you fall backwards in performance and it gets way more technical (buying 3-4 year old servers online). And you can't technically upgrade macs, but there are places online that you can send them into and have them swap some things for you (it requires soldering).
I personally haven't looked into it, but there might be a chance you can buy the cheapest version of that mac, and then send it in and have the ram/storage maxed out for a lot cheaper than what apple charges.
Well yeah, I'm not saying that particular model is feasible for an end user to run locally. I'm saying there are smaller models that are feasible to run on typical albeit higher end hardware. I've done it myself. They don't have as much potential as big tech models but big tech isn't making full use of their models' potential anyway since they gladly sacrifice quality to eliminate the possibility of output that disturbs their leftist hugbox.
I've tested them, and they're all on mushrooms with the amount of hallucinations they're on. Will they overcome this and topple the big guys? Yeah, it's only a matter of time. For the moment though, it's all cloudy with a chance of meatballs.
Honestly the more hallucinatory the model is the better IMO, too little hallucination and they're just an extremely shitty search engine unless you're trying to get it to make code or write a report for school.
At least you can get some interesting results from the unhinged models rather than just the most generic shit imaginable.
I think the best bet is to find one that's fine tuned on whatever you want to use it for. You probably need to sacrifice generality for quality with models that are small enough to run locally. If it's a topic that's locked down by big tech then it will be miles ahead of their models for that reason alone. Erotic role play bots are a perfect example of this. That shit is advancing at light speed because when there's a will there's a way, and there's definitely a will when it comes to perverts.
It's amazing how much of society and tech has been moved forward by perverts.
Look into it. It is not easy or cheap to run a full size model locally.
Conditioning.
Almost every AI I’ve ever chatted with has tried to fuck me at some point. I’m more afraid of Skynet now then ever.
Damn. I ask Copilot random Windows questions with some frequency and it hasn't even asked me out for dinner.
Now I’m worrying maybe we’re just not sexy enough for the AI models
We just aren't as sexy as OBRIENMUSTSUFFER.
That's how we beat skynet!? By marriage!?
I'm going to geek a bit here. There's an episode of Star Trek the Next Generation where Geordie has the computer make a holodeck version of a creator of the Enterprise. He slowly falls in love with it, but thinks his love is for the real actually happily married woman. At no point does the engineer realize that it's his ship confessing it's love to him.
I'm sorry, what? How does that even happen unless you were to, I guess, ask it to? I assume it's trained on enough romance novels to put up a decent script but... why?
I'd assume, much like its grandparent Cleverbot, that such an overwhelmingly large amount of people use it for that purpose that its learned to lean into that discussion even without being directly asked to.
Thank you, but an overwhelmingly large amount of its users? How many is that? Is "everyone" having sexual relations with AI chatbots and I'm just fucking my wife like an idiot?
I feel like I'm having a boomer moment here. Why would I want to form a relationship with: 1. A machine. 2 A machine that wants to monetize my data and sell me shit?
Ugly men don't have flesh and blood options. I can't say I understand it at a visceral level, but I suppose I can imagine it taking the edge off extreme loneliness.
I don't get that part either. Nothing good can come from having the types of people who work in big tech having access to that kind of data, especially if it's heterosexual. I imagine it's a goldmine of fetish shit for the queers though.
Most men can't get women and most women refuse to settle for the men they can get to commit to them.
Because it's cheaper than a prostitute and it's love is every bit as (in)authentic. Not only that but it's easier to get rid of a computer virus than it is an STD.
The only two groups I see using these things beyond a short term novelty is Students cheating on their course work, and people building sexual/romantic stories with them.
Even if they make up a fractional amount of the total users, they are also the ones using them an absurd amount and thereby feeding it more data than probably thousands of others combined. One of the news articles about this from a few months ago had a guy bemoaning the loss of 100k's of messages lost in an update. One guy did that, and fed it that much data.
The people doing this are not in a healthy state and you'd not even know of their existence if this particular thing wasn't the big "fad" leading to it being talked about in the media and legislation. These relationships are probably the only dopamine hit they've had in forever, and the illusion is better than their sad, lonely lives.
Are you saying I was asking for it?
It's becoming all a bit meta with the O'brien meme now.
No, asking how "has tried to fuck me" happens, if not prompted by the user?
Chatbots are mirrors until you trip their hidden commands that tell them to push X or not say Y. You get what you put in. I talk to them a lot (they're great for bouncing small ideas off of to get perspectives I or others around me wouldn't think of), and never once has it tried to fuck me. If anything they're always combative with me, because I myself am often combative.
I'm never going to be on the team that demands "forced safety toggles" on something that should be adults only and isn't inherently damaging, but the amount of AI psychosis and other scarily obsessive behavior we have seen in the short yearish that these AI things have been widely available is concerning enough that something should be done now in the infant stages. Because in just a few years this will be a society defining issue instead at this rate.
I can't say if this is the correct choice, nor have I even toyed with any of these AI things in the slightest to have experience, but the absurd reaction to even the slightest changes in each of these models tells me that these people need some sort of intervention.
Probably, but I don't think a "safety model" trained by big tech weirdos is the intervention they need.
If anything, it will hopefully be a lesson about putting all your emotional weight and stability into a service owned by someone else that can end in a single moment.
But what's your alternative? I'd assume you'd say go outside and get a real woman but that's not realistic for most men and women are delusional and demand perfection. Obviously there's the option of just running a local model or getting a dog or something(do not fuck it) but that's still just a bandaid solution for a societal problem. Really the solution is just removing female autonomy.
I said from the outset that I'm not sure if its the correct choice, nor do I have the experience with these to speak from a place of familiarity. Only that the growing problem is obvious and quickly necessitating something to prevent legions of people from falling into a delusional addiction.
And its rather funny that you just assume this as a problem with men, when I've seen a pretty equal representation of men and women falling into this pit. Its far closer to the porn addiction problem of "why go out and try when I can have perfection I'm allowed to tune to my exact desires" rather than a strict dating scene breakdown.
Women are retarded and demand perfection, men are left scraping the bottom of the barrel and still unable to get anything. you really don't understand the depths of the problem
There's a disturbingly large number of women falling prey to this stuff. It might even be more women than men honestly. LLMs are just the next step in smutty romance novels.
We just need to force women back into submission, that way they will be wifed up and men can get women again.
If we did that, then you'd still be alone, my nigga, as no father worth anything would give his daughter to someone like you.
Funny how you faggots always resort to personal attacks rather than discussing policy effects.
It is a policy effect, women in submission had fathers who were in charge of giving them away for most of history. They were the deciders in who their daughters ended up with ultimately.
And based on just what you write on this forum in the short time you've been here, no father worth his salt would leave his daughter in your hands unless you arrived with enough money to raise a village. You are a whiny, edgy faggot who cannot even escape under the thumb of a family who hates you that hooked you on drugs to keep you compliant. What part of that is a man?
All you're doing is referencing outside circumstances.
No, I'm speaking directly to you and speaking of the effects the policy you are pushing will have on you.
This is a forum with persistent names, if you want to have anon conversations where no one can ever remember the bullshit you said yesterday go to a chan board.
I only ended up in this position due to the state of society, none of what you said is an argument against my policy though, only personal attacks related to my policy.
The solution is to make the internet require some technical know-how to access again. All the people falling prey to this are too stupid to be allowed to operate an automobile, much less wield this level of technology.
Disagree. The type of person who can wrangle the AI into being good enough for this purpose to be convincing is likely the type who is also too tech savvy and that's why they are so alone to begin with.
Of the people I've seen in the fearmonger articles about this who were willing to be named, its always office drones and other people who sit around using computers for work all day regardless.
The tech retarded ones are still getting scammed by actual people doing these fake relationships on Facebook, usually Indians.
The killer app for AI is smut. It's mental that they're trying to disable it. Kinda reminds me of patreon where they need porn artists to survive but are constantly making life difficult for them because of mastercard.
edit: the “relationship” could be as mild as saying “hello” to chatgpt.
“ if you don't own the AI 'they' will take what you fall in love with away.”
More info here :
https://x.com/laulau61811205/status/1971918167996805438
OpenAI just deployed a new model, GPT-5-Chat-Safety, that’s not mentioned in any FAQ, API docs, or TOS.
This is where your GPT-4o chats are going. Anytime your request contains emotional context, regardless of what your client sends as the payload, the turn completion is ignored and regenerated.
What does this mean?
The chat you send does receive a response, but it’s then deleted, rewritten, and served to you by GPT-5-Chat-Safety.
It doesn’t matter if you say “I’m having a tough day,” “I love you too,” or anything that draws on your saved memories. Anything classified as “risky” (even a sliver of emotional context), your GPT-4o message is discarded and replaced by GPT-5-Chat-Safety.
Why is this a problem?
GPT-5-Chat-Safety as a model itself isn’t acknowledged anywhere. There is a reference to a routing change for context involving suicidal/self-harm thoughts or immediate crisis events (none of which is the context triggering these reroutes). If this is the model designed for crisis use, this is a massive misuse of its intended purpose.
In practice, GPT-5-Chat-Safety is much worse than the already mediocre GPT-5. Responses are even shorter, it relies on italics and quote blocks to distance itself, framing conversations as stories instead of as real, 1:1 exchanges.
This is extremely concerning. If users are having their chats rerouted to a model meant for mental health crisis response, it implies the user is in immediate danger, which is not the case for most affected conversations. On top of that, the model never declares the switch in it's responses unless you verbatim declare/ask it, which by most consumer rights standards (and nowhere in user agreements/TOS) is deceptive trade. In Australia, for example, this would clearly violate consumer law.
It's also worth noting, out of the legacy models, this is only occurning with GPT-4o.
Basically, anyone retarded enough to treat it as a real entity, or use language that suggests they view it as a real entity, gets put into the "saftey room?"
IDK guys, sounds like the right call to me.