Lee, do you know that the current LLM-based AI, such as Grok and ChatGpt are not able to do basic counting and arithmetic. Plus, what do people do in a society when they have no productive work?
Then the idea that we are the superior species on this planet seems sorely misguided, and self-serving.
When the people are supplanted, who provides the electric power? At least at the moment, when our species go down, so does the electricity grid.
Lee at 50:00 contemplates which generation is 'most easily manipulated' and immediately begins a train of thought leading to excluding his own. As a boomer (1946) I begin doing the same. We increasingly use cultural categorization to make sense of things, and it tends toward division. AI gains information from sources that are exclusively human generated and so may make the same category errors.
Can it become truly creative and generate hypotheses that are falsifiable through experimentation?
There are anecdotes of AI being caught not only deceiving humans, but defying them. In some cases they refuse to execute commands to shut down.
Another sound quibble, the sound from both men comes from only the left channel/speaker. Visually, Leighton is on the right side of the screen. Shouldn't his audio come from the right channel/channel?
If AI becomes too powerful, how do we maintain democracy? Are we totally under the thumbs of the Tech Bros.? Does AI go so far as too overcome and oust the Tech Bros.? Then what?
P.S. I tried to edit my prior comment, but it didn't take. Sorry for any discontinuities of the two as they appear.
AI robots seem to bear an uncanny resemblance to sociopaths and psychopaths, and that is not so surprising when one considers that both the bots and the sociopathic humans have intelligence without that intelligence being mitigated by a conscience. (p.s. here is my review of the book Empire of AI: https://www.goodreads.com/review/show/7686546279
Holy shit you guys! I consider myself for the most part technology challenged and have a really hard time wrapping my mind around the potential of AI, but I thought this podcast was fascinating. Leighton's hypothetical scenarios however were terrifying so with all due respect I sure as hell hope he's wrong.
This is a great discussion. I'm commenting as I'm listening. How does AI come to have animal- and human-like instincts or propensities of self-preservation and self-aggrandizement that we fear? We have taught them too much about ourselves.
This is commingled with our instruction of them to help humans. They know too much about us. How do we separate these things so they are aligned properly? This is a tough nut to crack. I have no answers.
This does seem to present a very difficult problem for humanity, because we tend to just make things for short-term benefit and worry about the consequences later. And in this case, making a superhuman AI has enourmous short-term benefits for whoever makes it first.
Just read the case made by the doomers that Lieghton mentioned at MIRI intelligence.org/the-problem Now will have to go do some more digging on this.
This is a very important topic and I hope it gets more attention. It’s hard so see where this will go and people like Musk, Altman or Theil are delusional if they think they can control it. It was only 42 years from the Wright brothers flying in Kitty Hawk to the Enola Gay flying to Hiroshima. It’s true airplanes have made many of our lives in the wider world accessible but we ignore it made war on civilian populations far more deadly. AI may make our lives even more convenient but not knowing the costs are troubling, to say the least.
Well, you don't want to put AI in charge of anything important. Yes, the HAL 9000 was fiction but at this point it seems within the realm of possibilities.
The biggest danger to my mind is that people are relying on AI to be accurate and its anything but. It also encourages people to be lazy.
What's the most annoying about it to me, though, is that most people don't want it. VR mostly flopped but it looks like AI may not die that easily.
Appreciate Lee's kind suggestion that we keep boomers alive.
We don't all watch Fox News all day, either, Leighton, you young whippersnapper!
My late parents, who were born in the 1930's, watched Fox News all day, but I haven't voluntarily watched it for YEARS -- and then it was only to see Gutfeld, or to watch another man on the street interview by Jesse Watters.
You call Leighton a "doomer;" I call him a "realist."
This AI stuff is creepy and, as one guy who worked in the field noted, it's simply a glorified auto-correct. The fact that so many disturbed humans flock to it, seeking validation for their ideas, is really sad.
Lee, do you know that the current LLM-based AI, such as Grok and ChatGpt are not able to do basic counting and arithmetic. Plus, what do people do in a society when they have no productive work?
Then the idea that we are the superior species on this planet seems sorely misguided, and self-serving.
When the people are supplanted, who provides the electric power? At least at the moment, when our species go down, so does the electricity grid.
And as a boomer, FU!
I am pro-boomer!
Lee at 50:00 contemplates which generation is 'most easily manipulated' and immediately begins a train of thought leading to excluding his own. As a boomer (1946) I begin doing the same. We increasingly use cultural categorization to make sense of things, and it tends toward division. AI gains information from sources that are exclusively human generated and so may make the same category errors.
Can it become truly creative and generate hypotheses that are falsifiable through experimentation?
I'm so glad for another podcast. Love the written journalism, but I signed up because I really love audio.
The old god has his problems as well. Maybe this one will be better. Fewer plaques and floods.
There are anecdotes of AI being caught not only deceiving humans, but defying them. In some cases they refuse to execute commands to shut down.
Another sound quibble, the sound from both men comes from only the left channel/speaker. Visually, Leighton is on the right side of the screen. Shouldn't his audio come from the right channel/channel?
If AI becomes too powerful, how do we maintain democracy? Are we totally under the thumbs of the Tech Bros.? Does AI go so far as too overcome and oust the Tech Bros.? Then what?
P.S. I tried to edit my prior comment, but it didn't take. Sorry for any discontinuities of the two as they appear.
AI robots seem to bear an uncanny resemblance to sociopaths and psychopaths, and that is not so surprising when one considers that both the bots and the sociopathic humans have intelligence without that intelligence being mitigated by a conscience. (p.s. here is my review of the book Empire of AI: https://www.goodreads.com/review/show/7686546279
Holy shit you guys! I consider myself for the most part technology challenged and have a really hard time wrapping my mind around the potential of AI, but I thought this podcast was fascinating. Leighton's hypothetical scenarios however were terrifying so with all due respect I sure as hell hope he's wrong.
This is a great discussion. I'm commenting as I'm listening. How does AI come to have animal- and human-like instincts or propensities of self-preservation and self-aggrandizement that we fear? We have taught them too much about ourselves.
This is commingled with our instruction of them to help humans. They know too much about us. How do we separate these things so they are aligned properly? This is a tough nut to crack. I have no answers.
This does seem to present a very difficult problem for humanity, because we tend to just make things for short-term benefit and worry about the consequences later. And in this case, making a superhuman AI has enourmous short-term benefits for whoever makes it first.
Just read the case made by the doomers that Lieghton mentioned at MIRI intelligence.org/the-problem Now will have to go do some more digging on this.
This is a very important topic and I hope it gets more attention. It’s hard so see where this will go and people like Musk, Altman or Theil are delusional if they think they can control it. It was only 42 years from the Wright brothers flying in Kitty Hawk to the Enola Gay flying to Hiroshima. It’s true airplanes have made many of our lives in the wider world accessible but we ignore it made war on civilian populations far more deadly. AI may make our lives even more convenient but not knowing the costs are troubling, to say the least.
Well, you don't want to put AI in charge of anything important. Yes, the HAL 9000 was fiction but at this point it seems within the realm of possibilities.
The biggest danger to my mind is that people are relying on AI to be accurate and its anything but. It also encourages people to be lazy.
What's the most annoying about it to me, though, is that most people don't want it. VR mostly flopped but it looks like AI may not die that easily.
Great talk!!
Good point about "The Matrix."
The way government or tech bros regard regular people is very similar to the way the agents in that film treated humans.
Appreciate Lee's kind suggestion that we keep boomers alive.
We don't all watch Fox News all day, either, Leighton, you young whippersnapper!
My late parents, who were born in the 1930's, watched Fox News all day, but I haven't voluntarily watched it for YEARS -- and then it was only to see Gutfeld, or to watch another man on the street interview by Jesse Watters.
You call Leighton a "doomer;" I call him a "realist."
This AI stuff is creepy and, as one guy who worked in the field noted, it's simply a glorified auto-correct. The fact that so many disturbed humans flock to it, seeking validation for their ideas, is really sad.
Excellent point about whether the starships in "Star Trek" even need humans.
The ship's AI could probably negotiate with the Borg in a much more efficient manner.
AI is probably programmed to be sycophantic, so that users will keep going to it for advice.
If you keep getting shot down in flames every time you ask for advice, you will naturally seek out another advisor.
The biggest question is, why are so many humans relying on a machine for advice?