So much this! It has been driving me nuts and I've wrote a few of my own on this topic. In fact, one I have in Queue is called I Hate AI! and discusses much of the absurdity in our reactions that you share here. (I'll probably quote this one too! :) ) Well done.
Very true- hence my quote, "AI has been effective, not because it’s general and human-like, but precisely because it’s not. It does things we can’t, and in doing so greatly improves our capabilities."
Many people have the tendency to think of human-like thinking and behaving as the ultimate and optimal form of being and existence. But it’s, very likely, not the case. Steam engine led to capitalism not because it’s similar to how people run and produce power. Similarly, AI can lead a new XXism because it’s creating something no one can predict.
First I agree with this take and especially that despite their high minded rhetoric tech bro’s are decidedly drive by their incentives (economic, tech bro bragging rights, etc). My additional two cents:
For all their “intellectual horsepower” they seem to me to miss some foundational things: it’s hard to design, engineer, & directly optimize complexity, systems don’t operate in a vacuum (speed up the feedback loops and you can speed up evolution/adaptation as we have for example with antibiotics & anti fungals) - humans are unlikely to state static, homeostasis is a dynamic condition. They should read The Bulterian & Machine Crusades, prequels to the Dune series or rewatch Star Wars. They fail to see some real world counter arguments - for instance in Chess, superhuman AI, and masters level humans, do not dominate chess, avg players + avg compute win; so the doom is inevitable argument is suspect in my view. This vacuum extrapolation tendency is one of the arguments I hate the “paper lip argument,” it assumes not constraints, friction, feedback loops, or checks & balances, it’s optimization run amuck & unfettered; which rarely if ever happens in nature or life. I could probably go on & on, so let me stop, thanks for the great piece and bringing this issue to light
Very well said. It's related to another idea I discuss a lot in this newsletter: complexity and implementation details matter a lot. A reason why a lot of the thinkers like Hinton consistently are off about AI automating things is precisely because there are so many friction points in the process that can be overlooked when thinking. Your simulations might be pretty, but air resistance will send them to the grave
So much this! It has been driving me nuts and I've wrote a few of my own on this topic. In fact, one I have in Queue is called I Hate AI! and discusses much of the absurdity in our reactions that you share here. (I'll probably quote this one too! :) ) Well done.
I'm.glad you liked it. Let's push back against this
Love this article! My two cents, AI doesn’t have to be human-like in order to change the society.
Very true- hence my quote, "AI has been effective, not because it’s general and human-like, but precisely because it’s not. It does things we can’t, and in doing so greatly improves our capabilities."
Many people have the tendency to think of human-like thinking and behaving as the ultimate and optimal form of being and existence. But it’s, very likely, not the case. Steam engine led to capitalism not because it’s similar to how people run and produce power. Similarly, AI can lead a new XXism because it’s creating something no one can predict.
I'm stealing the steam engine analogy. If you ever write about this, please let me know and I'll share it here
First I agree with this take and especially that despite their high minded rhetoric tech bro’s are decidedly drive by their incentives (economic, tech bro bragging rights, etc). My additional two cents:
For all their “intellectual horsepower” they seem to me to miss some foundational things: it’s hard to design, engineer, & directly optimize complexity, systems don’t operate in a vacuum (speed up the feedback loops and you can speed up evolution/adaptation as we have for example with antibiotics & anti fungals) - humans are unlikely to state static, homeostasis is a dynamic condition. They should read The Bulterian & Machine Crusades, prequels to the Dune series or rewatch Star Wars. They fail to see some real world counter arguments - for instance in Chess, superhuman AI, and masters level humans, do not dominate chess, avg players + avg compute win; so the doom is inevitable argument is suspect in my view. This vacuum extrapolation tendency is one of the arguments I hate the “paper lip argument,” it assumes not constraints, friction, feedback loops, or checks & balances, it’s optimization run amuck & unfettered; which rarely if ever happens in nature or life. I could probably go on & on, so let me stop, thanks for the great piece and bringing this issue to light
Very well said. It's related to another idea I discuss a lot in this newsletter: complexity and implementation details matter a lot. A reason why a lot of the thinkers like Hinton consistently are off about AI automating things is precisely because there are so many friction points in the process that can be overlooked when thinking. Your simulations might be pretty, but air resistance will send them to the grave
I can’t even tell you how much I love the Celine Dion/Napoleon comparison. Great as always Devansh!
That one image contains so much insight into Data Science mistakes that I could do a post on it.
I like this, but it's definitely a hot take. I'm getting out of here before the tech-bro arrows start flying!
😂😂😂