On Mon, May 10, 2021, 4:16 PM Mike Archbold <[email protected]> wrote:

> I can't speak for Colin but I do know that he isn't implementing
> algorithms....
>

Exactly. He is proposing an "Xchip" that reproduces the electrical noise
produced by real neurons. What he isn't proposing is any sort of
experiment, or any chip design, or any rational argument why this noise is
important other than that the last 70 years of trying to solve AI have
failed. He conveniently ignores all the progress we have made in AI with
neural networks that model the spiking rate as a continuous signal
representing a clamped, weighted sum of inputs and that learn by adjusting
anything that reduces the output error. It's like he is trying to
understand social networks by studying the noise from the CPU circuit board.

When Colin can answer my and WOM's questions I will take him seriously. But
I don't expect that to ever happen.

------------------------------------------
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T7c7052974ce450f1-M17b4824941e2aba759a3fe92
Delivery options: https://agi.topicbox.com/groups/agi/subscription

Reply via email to