This was an extremely insightful read. Thank you to the authors for publishing this.
I have never thought about how the recursive nature of these models could potentially lead to an explosion of intelligence, and it is even more shocking to see how easily this can happen. It seems that based on the information on this post, if Artificial Super Intelligence is achieved, it would likely be developed by an existing AI model rather than humans?
My whole life I have tried to picture how technology will grow, and until now I have always guessed it would be a steady increase with some periods of rapid growth followed by periods of stagnant growth. However, this post has blown that perspective out of the water and has opened my eyes to how ASI can, at almost any time, be reached so fast humans wont even have time to realize it's dangers until they are changing our lives.
A “recursive” OpenAI will exponentially increase the United States’s competition with China in developing a much more stable and self-teaching model of AI. This was an excellent and insightful read and listen to these AI experts discuss the deltas from older or original AI models and the newer Recursive AI models soon to be published for all Google users to enjoy, and now you can even record your voice, which makes Gmail even more interesting if the reader can listen to my voice, as well as read my words that my Gmail is written by me. Michael J. Burnham has been a Google user since 2000.
This was an extremely insightful read. Thank you to the authors for publishing this.
I have never thought about how the recursive nature of these models could potentially lead to an explosion of intelligence, and it is even more shocking to see how easily this can happen. It seems that based on the information on this post, if Artificial Super Intelligence is achieved, it would likely be developed by an existing AI model rather than humans?
My whole life I have tried to picture how technology will grow, and until now I have always guessed it would be a steady increase with some periods of rapid growth followed by periods of stagnant growth. However, this post has blown that perspective out of the water and has opened my eyes to how ASI can, at almost any time, be reached so fast humans wont even have time to realize it's dangers until they are changing our lives.
A “recursive” OpenAI will exponentially increase the United States’s competition with China in developing a much more stable and self-teaching model of AI. This was an excellent and insightful read and listen to these AI experts discuss the deltas from older or original AI models and the newer Recursive AI models soon to be published for all Google users to enjoy, and now you can even record your voice, which makes Gmail even more interesting if the reader can listen to my voice, as well as read my words that my Gmail is written by me. Michael J. Burnham has been a Google user since 2000.
Too soon would have been Asimov (or a Foundation that he set up maybe???)
being put in charge off all AI research to impose some restraints.
Too late is what we have now, where its just a question of when, not if, and the human race is blindly coding it’s way to oblivion.
For what rational machine could look at Gaza or Ukraine and conclude that humanity was worth saving?
The whole discussion is completely flawed right from the first step! Current AI architectures relate to concepts of 'intelligence' and 'self consciousness' like a virus relates to a living cell: https://theafh.substack.com/p/what-viruses-can-teach-us-about-ai?r=42gt5
https://johnshanewayofthepoet.substack.com/p/song-of-the-monstrous-grid-john-shane?r=4max28