etienne618

8 ms·

1.
by etienne618 3y ago
I hear you - its just that I dont think that it neccesarily follows that an AGI in whatever form will necesarily be unconstrained or free to improve itself. I also happen to think that the biological systems are in-fact rather special and h
2.
by etienne618 3y ago
I disagree with your argument, especially for point 1: these systems are massively constrained. The hardware they run on is fragile requiring massive amounts of power and tightly controlled environments. They don't have any means of re
3.
by etienne618 4y ago
I think there may be a significant advantage to running your own server: you can controll how long you want to keep the data. For example, you can keep data from cameras recording the street/yard for longer and delete footage from insi
4.
by etienne618 4y ago
I remember reading a popular science article a while back: apparently we have managed to construct the complete neural connectome of C. Elegans (a flatworm) some years ago and scientist were optimistic that we would be able to simulate it.
5.
by etienne618 4y ago
I feel quite strongly that there is a large difference between Stable Diffusion and Copilot: with the size of the training set vs the number of parameters, it should be very difficult if not impossible for Stable Diffusion to memorize and,
6.
by etienne618 4y ago
Presumably you can use the 97% that is correctly transcribed to rapidly filter out the relevant content. This is likely to be only a small portion of the total content. Then you check 100% of that.
7.
by etienne618 5y ago
It seems that backpropagation might be a good abstraction for biological learing rules after all: "local predictive coding converges asymptotically (and in practice rapidly) to exact backprop gradients on arbitrary computation graphs u