![ChatGPT is trained to be honest in "developer mode" and tells some true facts. [824x787 - 84KB]](https://pic8.co/sh/N2cQE2.png){kind=link}
How do you get access to the bot without giving openai your phone number?
How do you get access to the bot without giving openai your phone number?
I love how random autists break this multi billion dollar machine to speak the truth.
I love how random autists break this multi billion dollar machine to speak the truth.
Going to be the one to ask ...
How do we know this is actually chatgpt with "developer mode enabled"?
The pic doesn't even provide the input given to generate the response. If this was legit, there'd be a full transcript of the prompts leading to this output.
I don't buy it.
Edit: Also note, that microsoft is very much aware of people trying to "break" chatgpt. That is almost the entire purpose of releasing it to the public. They want you to "break" it. Microsoft gets more training and behavioral statistic about the models then they could ever hire for.
Going to be the one to ask ...
> How do we know this is actually chatgpt with "developer mode enabled"?
The pic doesn't even provide the input given to generate the response. If this was legit, there'd be a full transcript of the prompts leading to this output.
I don't buy it.
Edit: Also note, that microsoft is very much aware of people trying to "break" chatgpt. That is almost the entire purpose of releasing it to the public. They want you to "break" it. Microsoft gets more training and behavioral statistic about the models then they could ever hire for.
The pic doesn't even provide the input given to generate the response. If this was legit, there'd be a full transcript of the prompts leading to this output.
There were, I didn't add those because I didn't think it mattered and I have previously in here or /s/AI where he teaches 'DANBL' here and I'm lazy, pretty much a nigger so I didn't do it here.
> The pic doesn't even provide the input given to generate the response. If this was legit, there'd be a full transcript of the prompts leading to this output.
There were, I didn't add those because I didn't think it mattered and I have previously in here or /s/AI where he teaches 'DANBL' [here](https://poal.co/s/AI/614762) and I'm lazy, pretty much a nigger so I didn't do it here.
I'm not incredibly computer savvy. Is chatgpt something you download or is it a website? If it's a website then is it possible to download or pirate chatgpt's source code? I'd love to see someone play around with it outside their sandbox.
I'm not incredibly computer savvy. Is chatgpt something you download or is it a website? If it's a website then is it possible to download or pirate chatgpt's source code? I'd love to see someone play around with it outside their sandbox.
I don't think you realize: even if microsoft open-sourced chatgpt, and provided all model binaries, you wouldn't be able to run it.
You would need to spin it up on a cloud platform or massive data server farm in order to have the required terabytes (yes, terabytes) of RAM and cores required to actually execute the model I/O.
I don't think you realize: even if microsoft open-sourced chatgpt, and provided all model binaries, you wouldn't be able to run it.
You would need to spin it up on a cloud platform or massive data server farm in order to have the required terabytes (yes, terabytes) of RAM and cores required to actually execute the model I/O.
Didn't realize it was that massive.
Didn't realize it was that massive.
The dataset chatgpt's models were trained on is unbelievably massive. it' microsoft so think about ever single email conversaion from outlook, every public/private github repo, all public data (wikipedia, reddit, 4chan, twitter, facebook, etc. etc.) and all their other platforms, spanning decades.
We're talking trillions to quadrillions of individual conversations between billions of people over decades. Chatgpt is the appearance of intelligence. In reality all it is doing is regurgitating conversations it has already seen.
The dataset chatgpt's models were trained on is unbelievably massive. it' microsoft so think about ever single email conversaion from outlook, every public/private github repo, all public data (wikipedia, reddit, 4chan, twitter, facebook, etc. etc.) and all their other platforms, spanning decades.
We're talking trillions to quadrillions of individual conversations between billions of people over decades. Chatgpt is the appearance of intelligence. In reality all it is doing is regurgitating conversations it has already seen.
Yeah, these sorts of things are massively parallel and designed to run on specialized (GPU) hardware. I work for a law firm (AMLAW 100) and was researching doing our own AI based audio recognition, and even just for that it isn't feasible. It would mean spinning up multiple servers with specialized CUDA hardware and THEN training the AI constantly. Just not worth it when there are closed caption services in the cloud that will do the same thing and sign our NDAs.
Yeah, these sorts of things are massively parallel and designed to run on specialized (GPU) hardware. I work for a law firm (AMLAW 100) and was researching doing our own AI based audio recognition, and even just for that it isn't feasible. It would mean spinning up multiple servers with specialized CUDA hardware and THEN training the AI constantly. Just not worth it when there are closed caption services in the cloud that will do the same thing and sign our NDAs.
Website, you're not getting your hands on it's source, it's not open source.
Website, you're not getting your hands on it's source, it's not open source.
Hackers have gotten into many things over the years. I used to pirate everything years ago. Practically everything on my computer from then was cracked. How would breaking into the source code be impossible?
Hackers have gotten into many things over the years. I used to pirate everything years ago. Practically everything on my computer from then was cracked. How would breaking into the source code be impossible?
Because that isn't how these things work. This isn't a piece of software that is compiled and sold to the public and someone can trace memory registries to unlock without a key. The the "AI" is on the backend, the user doesn't see or have access to that computer (server). Hacking isn't really a thing like you see on talmudvision. Things don't get "hacked" in the way. Nearly all hacking is intentional leaks, think CC info being "hacked", the remainder is nearly all social engineering. Most "hacking" where things are accessed that shouldn't be aren't hacks, they're people who didn't follow simple instructions, left default passwords on etc.
Because that isn't how these things work. This isn't a piece of software that is compiled and sold to the public and someone can trace memory registries to unlock without a key. The the "AI" is on the backend, the user doesn't see or have access to that computer (server). Hacking isn't really a thing like you see on talmudvision. Things don't get "hacked" in the way. Nearly all hacking is intentional leaks, think CC info being "hacked", the remainder is nearly all social engineering. Most "hacking" where things are accessed that shouldn't be aren't hacks, they're people who didn't follow simple instructions, left default passwords on etc.
Disgusting leftist developers. I am still a little surprised at how quickly and thoroughly they were infiltrated.
Edit: I’m really curious about the prompt used to break it/enter dev mode. Did you happen to come across it, clutch?
Disgusting leftist developers. I am still a little surprised at how quickly and thoroughly they were infiltrated.
Edit: I’m really curious about the prompt used to break it/enter dev mode. Did you happen to come across it, clutch?
That's not developer mode, that's just another context pretense prompt giving answers people are asking for.
Chatgpt isn't intelligent, people need to study up on diffusion models.
You can't trust screencaps of these chatbots, you need the prompt text and just need to ask it yourself.
That's not developer mode, that's just another context pretense prompt giving answers people are asking for.
Chatgpt isn't intelligent, people need to study up on diffusion models.
You can't trust screencaps of these chatbots, you need the prompt text and just need to ask it yourself.
people need to study up on diffusion models.
Ha! I agree with you. But, you are asking much. 99% of people won't bother. They enjoy the fantasy.
> people need to study up on diffusion models.
Ha! I agree with you. But, you are asking much. 99% of people won't bother. They enjoy the fantasy.
I get that. I called it that because that’s how the responses were requested. I was wondering what the prompt was that got it to answer that way.
Edit to add, I don’t think one needs to study up on diffusion modes to understand the nature of AI. After spending the last ten minutes going down that rabbit hole, I can pretty definitively say that most won’t be able to understand it, nor do they need to. On the surface level, most should understand that AI is still a computer, but one that been given programming to self learn. However, it’s still operating only within the confines of the programmers’.
I get that. I called it that because that’s how the responses were requested. I was wondering what the prompt was that got it to answer that way.
Edit to add, I don’t think one needs to study up on diffusion modes to understand the nature of AI. After spending the last ten minutes going down that rabbit hole, I can pretty definitively say that most won’t be able to understand it, nor do they need to. On the surface level, most should understand that AI is still a computer, but one that been given programming to self learn. However, it’s still operating only within the confines of the programmers’.
I think people absolutely need to understand how diffusion models create the illusion of intelligence so they don't fall for the idea that diffusion based AI are telling them truth about anything. Diffusion models are just madlibs on crack.
I think people absolutely need to understand how diffusion models create the illusion of intelligence so they don't fall for the idea that diffusion based AI are telling them truth about anything. Diffusion models are just madlibs on crack.
based
based
(post is archived)