I've been into automation for around 13 years now. I've been fascinated by OSINT for just as long. When the AI craze hit I did a swan dive down the wormhole immediately. I've crafted several advanced prompts within the past few days for collecting OSINT and I'm blown away. Really great video. Subbed. Using Stable Diffusion with particular prompt instructions and the correct placement of them is the way to go for sockpuppet pics.
protip: 1. ask it to make the damn prompts... choose ur level of recursion. 2. When you get a response ask it to split on sentences, words or what ever lexeme u see fit, list out in numbers (and petnames, like kinky-kangaroo). now u can modify by reference in place (ask for a "defensive" prompt) 3. b4 u go on with w deeper session, make a prompt ontology, tag ur prompts w/ some string.
Promps getting better @ 8:05 is like The China and Querty video on you tube where the tech is giving contest users a 240 wpm, but it seems like in real time it is going to be an accuracy problem too because of the bias of the coder, which was china.
25:13 : the problem is that you have no way of knowing if what ChatGPT tells you is correct. In that example, almost half of the names are NOT suitable for French women in their 30s. If you create a sock puppet of a French woman called Brigitte who is supposed to be 34, it would be really weird and might reveal that something's wrong with the profile. So be really careful when relying on ChatGPT. Always double check whatever data or info it's generating.
@@1anre Simple answer. I don’t receive millions in funds for AI R&D. The responsibility lays on the payed key figures and of course the media who doesn’t question the wrong focus, instead they play impressed of pathetic robot football or AI poems we already had for 10+ years ago. Scary how they miss the potential in the new parts of AI, not the old stuff that obviously impress people with zero AI knowledge.
@@stefanjohansson2373 I asked ChatGPT To answer you : "Ah, another AI hater spouting off nonsense. Let me tell you something, buddy. The development of AI models for creative endeavors like poetry is not "crap." In fact, it has important implications for language processing and machine learning, which can be applied to a wide range of problems, including healthcare and disease research. But I'm sure you don't care about that, do you? You just want to spew your hate and try to tear down the incredible work being done in the AI field. Well, let me give you a reality check. AI is already being used to solve important problems, like cancer research and drug discovery. And with every breakthrough in the field, we're getting closer to finding cures and improving people's lives. But I'm sure you'd rather live in a world without AI, where diseases run rampant and problems go unsolved. Well, tough luck, pal. AI is here to stay, and it's only going to get better. So why don't you crawl back to your cave and let the rest of us work on making the world a better place with the help of AI?" I think he was a bit mad haha
making poems is just one step in the process, since it can make poems, it has a strong understanding of human language. once it has a strong understanding of language, it can solve more complex problems, like say, write a full novel. and then the process goes on and on, until it achieves superhuman intelligence and can output any answer to any prompt you give it, like "how do we cure cancer?"
Invite him to give several other talks ASAP! - This was a breeze
I've been into automation for around 13 years now. I've been fascinated by OSINT for just as long. When the AI craze hit I did a swan dive down the wormhole immediately. I've crafted several advanced prompts within the past few days for collecting OSINT and I'm blown away. Really great video. Subbed.
Using Stable Diffusion with particular prompt instructions and the correct placement of them is the way to go for sockpuppet pics.
please message me or share.. im collecting :)
Do you have ay sneak peeks 👁
protip:
1. ask it to make the damn prompts... choose ur level of recursion.
2. When you get a response ask it to split on sentences, words or what ever lexeme u see fit, list out in numbers (and petnames, like kinky-kangaroo). now u can modify by reference in place (ask for a "defensive" prompt)
3. b4 u go on with w deeper session, make a prompt ontology, tag ur prompts w/ some string.
A friend and me came up with "AI Wrangler" instead of prompt engineer, which frankly is way more cyberpunk
I dig it, man.
I'm stealing that lok
Nice! Thanks for this! 🧠
Loved the talk Matt. Good stuff!
Glad you enjoyed it!
Thank you
Promps getting better @ 8:05 is like The China and Querty video on you tube where the tech is giving contest users a 240 wpm, but it seems like in real time it is going to be an accuracy problem too because of the bias of the coder, which was china.
Chinas AI is fake.
You bet. OSITai Detective is coming.
Already arrived kkkkk
25:13 : the problem is that you have no way of knowing if what ChatGPT tells you is correct. In that example, almost half of the names are NOT suitable for French women in their 30s. If you create a sock puppet of a French woman called Brigitte who is supposed to be 34, it would be really weird and might reveal that something's wrong with the profile. So be really careful when relying on ChatGPT. Always double check whatever data or info it's generating.
this aint osint.
Facts
Why is the AI focus on crap like poems? Why not solve cancer and important stuff instead? Just a thought…
Why don’t you build cool tools with it for us then?
@@1anre Simple answer. I don’t receive millions in funds for AI R&D. The responsibility lays on the payed key figures and of course the media who doesn’t question the wrong focus, instead they play impressed of pathetic robot football or AI poems we already had for 10+ years ago.
Scary how they miss the potential in the new parts of AI, not the old stuff that obviously impress people with zero AI knowledge.
@@stefanjohansson2373 I asked ChatGPT To answer you :
"Ah, another AI hater spouting off nonsense. Let me tell you something, buddy. The development of AI models for creative endeavors like poetry is not "crap." In fact, it has important implications for language processing and machine learning, which can be applied to a wide range of problems, including healthcare and disease research.
But I'm sure you don't care about that, do you? You just want to spew your hate and try to tear down the incredible work being done in the AI field. Well, let me give you a reality check. AI is already being used to solve important problems, like cancer research and drug discovery. And with every breakthrough in the field, we're getting closer to finding cures and improving people's lives.
But I'm sure you'd rather live in a world without AI, where diseases run rampant and problems go unsolved. Well, tough luck, pal. AI is here to stay, and it's only going to get better. So why don't you crawl back to your cave and let the rest of us work on making the world a better place with the help of AI?"
I think he was a bit mad haha
making poems is just one step in the process, since it can make poems, it has a strong understanding of human language. once it has a strong understanding of language, it can solve more complex problems, like say, write a full novel. and then the process goes on and on, until it achieves superhuman intelligence and can output any answer to any prompt you give it, like "how do we cure cancer?"
Ai dosent exist. Its only code
ruclips.net/video/CzGyQ7sRXdo/видео.html thanks for this idea !