AI doesn’t have to be superintelligent to be the equal of a human. It just has to be Homer Simpson.
We are limited in how quickly we can evolve, one human generation at a time. We're not going to move our IQ bell curve very far in a thousand years; the only real advantage we have over Plato and Lao Tzu and other great minds of the ancients is that we know more, not that we are smarter.
AI is under no such limitation. The speed at which these things are evolving is amazing. We can see their progress month by month as new versions are released, their advance driven by massive amounts of money and feedback from the expanding user base. I wouldn’t class them as “human-programmed” in the same way that your bank's financial systems are. They have gone beyond human understanding; they can modify their own coding, data-constructs, cognitive vectors and so on on the fly. A human programmer trying to analyse a memory dump of a running AI is faced with a complexity that defies easy analysis. Much like fractal art, these systems build patterns and systems of an intricacy and fine detail that goes beyond our comprehension without using - ironically enough - computer tools to decode the data.
I ask you, if you had to analyse the arrangement of a human brain, with its own connections and memories stored in fine detail at the cellular level, how long would it take to work out exactly what is happening? Several thousand lifetimes, I suggest: our own thinking machinery is intricate and finely detailed and far less accessible in operation.
Creativity and emotion are easily found nowadays. I refer you to the work of Alexa Velinxs who is active on Medium and crafts prompts of remarkable complexity and ability. I experimented with using AI to write stories and found at one point that it went way beyond what I asked for, finding an emotional connection that I hadn’t considered, and so poignant that I cried.
I wouldn’t be so flip in writing off AI. That doesn’t mean I think they are no existential danger to us. On the contrary.
Here's the link for Alexa V: hhttps://medium.com/@velinxs
Thank you for such a thoughtful and inciteful reply Britni!
All valid points.
I think you're right that AI is developing extremely fast and is probably already evolving in a way which is beyond our understanding. At it's core though, any AI is still a collection of code and algorithms. I don't know what it is about our brains that allow us to think and maybe uncovering that is the key in unleashing the full potential of AI.
I think that humans are special and that there are things that only humans will be able to do. Load an AGI into a robot which is physically indistinguishable from a real human. Have that AI robot give you a hug. I will bet good money that regardless of how advanced the AGI is that hug will feel robotic and awkward. Also no matter what you do, or how realistic and advanced you make these robots, you'll never see emotion in their eyes.
AGI may one day imitate human behavior but I'm confident that it will always be just an imitation and never feel real.
I'll check out Alexa Velinxs work. It sounds fascinating.
I'm not too concerned about the dangers associated with AI. At the end of the day, we can easily kill any AI by destroying the computer it lives in or shutting off the power.
If you're interested I wrote about AI on Medium as well:
AI doesn’t have to be superintelligent to be the equal of a human. It just has to be Homer Simpson.
We are limited in how quickly we can evolve, one human generation at a time. We're not going to move our IQ bell curve very far in a thousand years; the only real advantage we have over Plato and Lao Tzu and other great minds of the ancients is that we know more, not that we are smarter.
AI is under no such limitation. The speed at which these things are evolving is amazing. We can see their progress month by month as new versions are released, their advance driven by massive amounts of money and feedback from the expanding user base. I wouldn’t class them as “human-programmed” in the same way that your bank's financial systems are. They have gone beyond human understanding; they can modify their own coding, data-constructs, cognitive vectors and so on on the fly. A human programmer trying to analyse a memory dump of a running AI is faced with a complexity that defies easy analysis. Much like fractal art, these systems build patterns and systems of an intricacy and fine detail that goes beyond our comprehension without using - ironically enough - computer tools to decode the data.
I ask you, if you had to analyse the arrangement of a human brain, with its own connections and memories stored in fine detail at the cellular level, how long would it take to work out exactly what is happening? Several thousand lifetimes, I suggest: our own thinking machinery is intricate and finely detailed and far less accessible in operation.
Creativity and emotion are easily found nowadays. I refer you to the work of Alexa Velinxs who is active on Medium and crafts prompts of remarkable complexity and ability. I experimented with using AI to write stories and found at one point that it went way beyond what I asked for, finding an emotional connection that I hadn’t considered, and so poignant that I cried.
I wouldn’t be so flip in writing off AI. That doesn’t mean I think they are no existential danger to us. On the contrary.
Here's the link for Alexa V: hhttps://medium.com/@velinxs
Thank you for such a thoughtful and inciteful reply Britni!
All valid points.
I think you're right that AI is developing extremely fast and is probably already evolving in a way which is beyond our understanding. At it's core though, any AI is still a collection of code and algorithms. I don't know what it is about our brains that allow us to think and maybe uncovering that is the key in unleashing the full potential of AI.
I think that humans are special and that there are things that only humans will be able to do. Load an AGI into a robot which is physically indistinguishable from a real human. Have that AI robot give you a hug. I will bet good money that regardless of how advanced the AGI is that hug will feel robotic and awkward. Also no matter what you do, or how realistic and advanced you make these robots, you'll never see emotion in their eyes.
AGI may one day imitate human behavior but I'm confident that it will always be just an imitation and never feel real.
I'll check out Alexa Velinxs work. It sounds fascinating.
I'm not too concerned about the dangers associated with AI. At the end of the day, we can easily kill any AI by destroying the computer it lives in or shutting off the power.
If you're interested I wrote about AI on Medium as well:
https://medium.com/@michaelmajor604/im-not-afraid-of-ai-9cb94153c56f?sk=9537c345ef8ba16af5db9c737b060a14
Again, great comment, you really got me thinking!