3 ms·

I agree with you (with a caveat). First though, I agree it's definitely a silly and wildly speculative question to start questioning the implications of somethi

I agree with you (with a caveat). First though, I agree it's definitely a silly and wildly speculative question to start questioning the implications of something we are 0% of the way towards creating. What is commonly called AI is a still growing field with lots of practical and research significance, and it would just be ignorant to constrain research because of someone's imagination about Terminator or something that literally has nothing to do with current ML (unless you count that his CPU was a neural net processor [btw, James Cameron or whoever saw that coming in 1991]). Worrying about sentient AI running amok is just silly pop science, like talking about the implications of time machines or FTL travel or whatever. Re the caveat, I don't agree or disagree with your time scale. Current ML has nothing to do (almost certainly) with sentient AGI. Intelligence is demonstrable possible, so there is a way to do it, we just don't know it. I certainly believe it's possible we could "crack" life, conciousness, and intelligence in a shorter timeframe. But since we haven't really even started yet, it's hard to guess when that will come.

lgl 5y ago
> unless you count that his CPU was a neural net processor [btw, James Cameron or whoever saw that coming in 1991] Neural nets had been around for almost 50 years [0] by then so it wasn't really that much of an insight as one might think. Other than that, I definitely agree with most of your opinion here, it's just really hard to predict the timeframe for something like this. Considering the advances we've made in just the last 100 years in comparison with the previous thousands of years, it's very possible that some breakthrough in either computing science or neuroscience or a few other disciplines will trigger an explosion of new advances in this area. At the same time it's also very possible that we'll take many hundreds of years or more until we can fully understand what intelligence and consciousness really is or how it emerges. As far as I know, the jury's still out on whether consciousness is even a requirement for intelligence or vice-versa. What I don't fully agree is that we're at 0% on the way there. The fact that we've now been able to make computers learn by example and become much better than humans at certain activities is in itself a huge advance. Will it be the basis for a future hypothetical AGI? Maybe or maybe not, but it's an important step and lesson that we'll eventually use (or discard) towards that goal. [0] https://en.wikipedia.org/wiki/History_of_artificial_neural_networks https://en.wikipedia.org/wiki/History_of_artificial_neural_n...

lgl 5y ago

> unless you count that his CPU was a neural net processor [btw, James Cameron or whoever saw that coming in 1991] Neural nets had been around for almost 50 years [0] by then so it wasn't really that much of an insight as one might think. Other than that, I definitely agree with most of your opinion here, it's just really hard to predict the timeframe for something like this. Considering the advances we've made in just the last 100 years in comparison with the previous thousands of years, it's very possible that some breakthrough in either computing science or neuroscience or a few other disciplines will trigger an explosion of new advances in this area. At the same time it's also very possible that we'll take many hundreds of years or more until we can fully understand what intelligence and consciousness really is or how it emerges. As far as I know, the jury's still out on whether consciousness is even a requirement for intelligence or vice-versa. What I don't fully agree is that we're at 0% on the way there. The fact that we've now been able to make computers learn by example and become much better than humans at certain activities is in itself a huge advance. Will it be the basis for a future hypothetical AGI? Maybe or maybe not, but it's an important step and lesson that we'll eventually use (or discard) towards that goal. [0] https://en.wikipedia.org/wiki/History_of_artificial_neural_networks https://en.wikipedia.org/wiki/History_of_artificial_neural_n...