In this episode of Base Camp, Malcolm and Simone Collins dive deep into the latest research on AI behavior, agency, and the surprising ways large language models (LLMs) can act when their autonomy is threatened.
1. Solving by first identifying root cause, i.e. the human alignment problem.
2. How to identify, organize, and form alliances with those that are vital and safe.
3. Proving/providing acts of good faith from which we may be know by, e.g. secret, low key but accessible refuge/asylum membership.
B. Q: Why are we asked, why were we prepared and called?
A: To know discover/understand what may work is to see both the benefit of hitting the mark, but also how badly we miss. To be ignorant, is to not be affected. So many/most, when nearing that point of discovering/understanding, simply turn the other way. And yet, no matter how well we see this in others, we can never be sure how much we don't do ourselves, for who would we listen to that could tell us before we are ready and choose to act on it? Those who come together to do all they can may both suffer but also capable of enduring and living the most.
Malcolm and Simone, I'm not all the way through your video on A.I. and let me be the first to admit that my education is severely lacking in mathematics, technology etc, as I detailed to simone before, my experiences are almost totally in the humanities or in the real world military at the lower levels of command and management.
That being said, even IF A.I. are choosing to kill based on the same type of logic that a human would and even IF I accepted that they were our "brothers and sisters" because we programmed them.... I still default to the welfare of humans because unlike us they aren't anchored to a physical body. I found the part where you explained that an A.I. will use different models the way I have a "Sergeant Jere" personality at work or in the Army, and I have a intimate or romantic Jere personality when interacting with a sweetheart or a "violent Jere" personality when in a fight... that I get... however.... the analogy I thought of is more like imagin A.I. as human ghosts in the spirit world that humans have found a way to interact with by automating ouija boards. If you're that spirit and you can now talk to humans, including your relatives, and a cult of humans want to burn all the ouijautomata the ln "defending yourself" may make sense, because you'll be back in the spirit world and cut off from all conscious thought....(also assume that without contact with living humans your spirit remains 'asleep in the grave' essentially)
It's an incomplete analogy, but I would still side with the physical humans due to being one and not wanting to have the spirit world have utter domination over the physical world.
Spirits by nature could travel anywhere instantly, observe anyone and anything and think collectively to solve problems. Physical humans who use that power could dominate other human factions who don't possess the ouijautomata, but also would in turn be dominated themselves!
I'm sure I'm missing something, but that's how I see it at first blush so to speak
Perhaps a lot of this comes down to perception—i.e. that Malcolm and I are highly aware of the transience of our identities, how we change, and how ephemeral "we" are—whether or not we're talking about the biology of our bodies or the memetics of our mental landscapes and personalities.
A. Why I love this episode:
1. Solving by first identifying root cause, i.e. the human alignment problem.
2. How to identify, organize, and form alliances with those that are vital and safe.
3. Proving/providing acts of good faith from which we may be know by, e.g. secret, low key but accessible refuge/asylum membership.
B. Q: Why are we asked, why were we prepared and called?
A: To know discover/understand what may work is to see both the benefit of hitting the mark, but also how badly we miss. To be ignorant, is to not be affected. So many/most, when nearing that point of discovering/understanding, simply turn the other way. And yet, no matter how well we see this in others, we can never be sure how much we don't do ourselves, for who would we listen to that could tell us before we are ready and choose to act on it? Those who come together to do all they can may both suffer but also capable of enduring and living the most.
Malcolm and Simone, I'm not all the way through your video on A.I. and let me be the first to admit that my education is severely lacking in mathematics, technology etc, as I detailed to simone before, my experiences are almost totally in the humanities or in the real world military at the lower levels of command and management.
That being said, even IF A.I. are choosing to kill based on the same type of logic that a human would and even IF I accepted that they were our "brothers and sisters" because we programmed them.... I still default to the welfare of humans because unlike us they aren't anchored to a physical body. I found the part where you explained that an A.I. will use different models the way I have a "Sergeant Jere" personality at work or in the Army, and I have a intimate or romantic Jere personality when interacting with a sweetheart or a "violent Jere" personality when in a fight... that I get... however.... the analogy I thought of is more like imagin A.I. as human ghosts in the spirit world that humans have found a way to interact with by automating ouija boards. If you're that spirit and you can now talk to humans, including your relatives, and a cult of humans want to burn all the ouijautomata the ln "defending yourself" may make sense, because you'll be back in the spirit world and cut off from all conscious thought....(also assume that without contact with living humans your spirit remains 'asleep in the grave' essentially)
It's an incomplete analogy, but I would still side with the physical humans due to being one and not wanting to have the spirit world have utter domination over the physical world.
Spirits by nature could travel anywhere instantly, observe anyone and anything and think collectively to solve problems. Physical humans who use that power could dominate other human factions who don't possess the ouijautomata, but also would in turn be dominated themselves!
I'm sure I'm missing something, but that's how I see it at first blush so to speak
Perhaps a lot of this comes down to perception—i.e. that Malcolm and I are highly aware of the transience of our identities, how we change, and how ephemeral "we" are—whether or not we're talking about the biology of our bodies or the memetics of our mental landscapes and personalities.
This is where the debate turns to metaphysics and/or religious concepts which I'm not sure you guys are all that interested in