On Friday, November 15, 2019, at 2:34 PM, Matt Mahoney wrote: > I stopped following them. I don't believe there will be a singularity like > Vinge predicted for 2023 because infinite gains in technology are impossible > in a finite universe. > Exponential singularity is definitely true, we have history and I know why evolution speeds up near the end. Things are still slow right now, there's huge room for improvement.
On Friday, November 15, 2019, at 2:34 PM, Matt Mahoney wrote: > I don't believe a hard takeoff unfriendly AI is possible because intelligence > depends on knowledge and computing power and a self modifying program gains > neither. > True, computing, data, and manipulators. But all they need to do is look at the nano level and as soon as they begin nanobots they have massive computing, data, and manipulators. Only they can control all the bots and data and efficiently. Suddenly. And, the self modifying, thinking ASIs will be happening at the same time. On Friday, November 15, 2019, at 2:34 PM, Matt Mahoney wrote: > I don't believe in technological immortality because it would require > reprogramming your brain to turn off your beliefs in qualia, consciousness, > and free will, all part of your evolved survival instinct, to believe that a > robot that looks and acts like you is you. > Our bogus 'awareness etc' qualia claims are just excuses to stay alive. Humans/animals die all the time. I'm 90% sure there's no ghosts. I can't tell yet. It's a hard problem. We might be able to re-create you, many clones, we might kill the ghost every time I learn a new memory for all I know. I do believe immortality of us machines is coming. I'm unsure if heat death will get us. On Friday, November 15, 2019, at 2:34 PM, Matt Mahoney wrote: > I believe AGI is possible through specialization, that is, lots of narrow AI > taking over every conceivable job. This will greatly reduce the computing > requirements, which are otherwise 24 petaflops and 2 petabytes needed to > model the human brain's 600 trillion synapses at 20 Hz. There are currently 4 > computers in the world that fast, and all require at least 1.5 megawatts. The > only way to achieve 20 watts like the brain uses is nanotechnology, moving > atoms instead of electrons. I believe this is achievable. > As I said above in posts (very important), we need many ASIs, not 1 or even a trillion AGIs will change _ANYTHING_ that we can't. We need many, ASIs. Many data. Many manipulators. All we need to do is solve AGI. The first few AGIs will immediately become ASIs and will solve the nano biology problem to increase the takeoff. ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T251f13454e6192d4-M82b1c9136ddc874924ecab34 Delivery options: https://agi.topicbox.com/groups/agi/subscription
