spudboy100 via Everything List <[email protected]> wrote:
*> My Trumpian amygdala has flat out taken my neocortex hostage. As we say > in Trumpville, Golly! While you are undertaking that arduous task, you > never got a chance to address this impact question I submitted at the > bottom of one of my rants. * That's probably because you throw so many garbage links into almost every post you write I've gotten in the habit of ignoring them > This is an article, and I'd like you opinion of the likely impact of GPT-4 on > research and development? I will re-supply the link. > > https://towardsdatascience.com/gpt-4-will-have-100-trillion-parameters-500x-the-size-of-gpt-3-582b98d82253 This link was uncharacteristically interesting and reinforces my opinion that human level AI can't be very far away, and once that level has been achieved the attainment of superhuman AI can only take a few days more, perhaps only a few hours. I just don't see how we can be very far away from finding the seed algorithm that would allow computers to learn how to learn anything, because we already have an upper limit on how big that algorithm must be, and it's not very big. In the entire human genome there are only 3 billion base pairs. There are 4 bases so each base can represent 2 bits, there are 8 bits per byte so that comes out to just 750 meg, and that's enough assembly instructions to make not just a brain and all its wiring but an entire human baby. So the instructions MUST contain wiring instructions such as "wire a neuron up this way and then repeat that procedure exactly the same way 917 billion times". And there is a huge amount of redundancy in the human genome, so if you used a file compression program like ZIP on that 750 meg you could easily put the entire thing on a CD, not a DVD not a Blu ray just a old fashioned steam powered vanilla CD, and you'd still have room for a few dozen lady Gaga songs. And the thing I'm talking about, the seed learning algorithm for intelligence, must be vastly smaller than that, and that's the thing that let Einstein go from knowing precisely nothing in 1879 to becoming the first person to know General Relativity in 1915. That's why I told you that whatever you do for a living it's only a matter of time, and not much time, before a machine can do your job better than you can. I always knew something like this would happen but I didn't think it would happen in my lifetime so I saw no need for an immediate change in my strict libertarian beliefs, but things have been developing much faster than I expected, and when new information becomes available an intelligent man must change his opinion. I'm still a libertarian with regard to social issues, but on economic matters not so much. It's just not rational to expect that the gargantuan and accelerating gap between the rich and the poor can continue indefinitely without causing a complete collapse of civilization. People just are not going to quietly starve to death, they are going to make a fuss. And that's also why I think all the worry Trump zombies have about Mexicans taking their jobs is ridiculous when the real problem is elsewhere. John K Clark See what's on my new list at Extropolis <https://groups.google.com/g/extropolis> bo -- You received this message because you are subscribed to the Google Groups "Everything List" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/everything-list/CAJPayv162dYfzeHxe0UUL1Wg%3DtWb68sNNGGnZYbu0caRZ54Amw%40mail.gmail.com.
