Comments on: The Brain vs Deep Learning Part I: Computational Complexity — Or Why the Singularity Is Nowhere Near https://timdettmers.com/2015/07/27/brain-vs-deep-learning-singularity/ Making deep learning accessible. Sat, 30 Oct 2021 04:25:45 +0000 hourly 1 https://wordpress.org/?v=6.0.11 By: Tim Dettmers https://timdettmers.com/2015/07/27/brain-vs-deep-learning-singularity/comment-page-1/#comment-97522 Sun, 24 Oct 2021 17:33:19 +0000 https://timdettmers.wordpress.com/?p=312#comment-97522 In reply to Omar.

Thanks for leaving a comment about that. I have a plug-in that detects dead/broken links, but it did not pick up on this. Thanks for letting me know!

]]> By: Omar https://timdettmers.com/2015/07/27/brain-vs-deep-learning-singularity/comment-page-1/#comment-93308 Tue, 27 Jul 2021 05:20:53 +0000 https://timdettmers.wordpress.com/?p=312#comment-93308 This article is pretty interesting, I am new to AI and I find the contents very exciting but I wish the image links were not broken. It just makes it hard to visualize what I am reading.

Still, this is a great read. Thanks!

]]> By: Tim Dettmers https://timdettmers.com/2015/07/27/brain-vs-deep-learning-singularity/comment-page-1/#comment-48354 Sat, 08 Dec 2018 19:51:58 +0000 https://timdettmers.wordpress.com/?p=312#comment-48354 In reply to Rolando.

Yes, I would go for free wordpress. I did that two for a couple of years before I got my own server and own domain.

]]> By: Rolando https://timdettmers.com/2015/07/27/brain-vs-deep-learning-singularity/comment-page-1/#comment-48268 Fri, 07 Dec 2018 23:02:18 +0000 https://timdettmers.wordpress.com/?p=312#comment-48268 Superb blog! Do you have any recommendations for aspiring writers?
I’m planning to start my own website soon but I’m
a little lost on everything. Would yyou advise starting with a free platform like
Wordpress or go for a paid option? There aree soo many options out there
that I’m tktally confused .. Any recommendations?
Cheers!

]]> By: dirk bruere https://timdettmers.com/2015/07/27/brain-vs-deep-learning-singularity/comment-page-1/#comment-44991 Tue, 23 Oct 2018 17:17:34 +0000 https://timdettmers.wordpress.com/?p=312#comment-44991 In reply to Peter Kinnmark.

It does not matter whether Kurtzweil is right or wrong. It is something we will discover by experiment. Right now NN h/w is on a steep “More than Moore” curve. When that starts to level out in 10-15 years we will have a better view of the landscape. Meanwhile at the very least narrow AI promises to deliver very impressive performance enhancements in a vast number of fields. Also, I’m not too sure about the emphasis on Human scale AGI. What we want of AI may well be quite achievable without “recreating the Human brain”. Let’s just see where this goes

]]> By: Tim Dettmers https://timdettmers.com/2015/07/27/brain-vs-deep-learning-singularity/comment-page-1/#comment-44985 Tue, 23 Oct 2018 16:10:40 +0000 https://timdettmers.wordpress.com/?p=312#comment-44985 In reply to Peter Kinnmark.

Thanks for your comment. I can see your point and it makes sense to me. I believe that it is not really possible to accurately picture a distinct future, but it helps to have some reasonable paths to the future. I think mine is reasonable as is yours, but in the end we have to wait it out to see what turns out to be true.

]]> By: Peter Kinnmark https://timdettmers.com/2015/07/27/brain-vs-deep-learning-singularity/comment-page-1/#comment-44954 Tue, 23 Oct 2018 05:02:34 +0000 https://timdettmers.wordpress.com/?p=312#comment-44954 Even without a definition of intelligence, it’s resonable to claim that humans have it. But human intelligence, probably closely linked to to at least the scences, is the result of evolution, not our thinking. There is no reason to assume that our intellectual power is enough to properly decode the workings of the brain, the imprint of 600 million years of evolution. Predicting strong AI based on the progress of computer hardware is thus a false idea.
The above dilemma is true for any step in the improvement of intelligence. The idea that intelligence, summed up in a fuction i, will automatically lead to a new intelligence, i+1, is false parallel to the same argument. Thus Kurtzweils singularity is not just wrong. It’s in fact infinitely wrong.

]]> By: Tim Dettmers https://timdettmers.com/2015/07/27/brain-vs-deep-learning-singularity/comment-page-1/#comment-24008 Thu, 16 Nov 2017 21:06:22 +0000 https://timdettmers.wordpress.com/?p=312#comment-24008 In reply to George Michaelson.

That is a bit of what I am trying to do. I try to show that Ray Kurzweil’s view is too simplistic and naive. I try to replace it with something that makes more sense. Still, I cannot predict the future with this, but I can show boundaries. With current progress, it will take at least a century for AGI to happen. Personally, I think it will take much longer, but I cannot predict that. If the progress is good and exponential, we can hope for AGI in a 100 years and not any earlier really. So what I am providing here is a lower bound, but not any predictive on its own.

]]> By: Tim Dettmers https://timdettmers.com/2015/07/27/brain-vs-deep-learning-singularity/comment-page-1/#comment-24007 Thu, 16 Nov 2017 21:02:08 +0000 https://timdettmers.wordpress.com/?p=312#comment-24007 In reply to John Smith.

Here a popular reference for a short overview. For papers just follow the links in the article: https://www.scientificamerican.com/article/scientists-surprised-to-find-no-two-neurons-are-genetically-alike/

]]> By: Tim Dettmers https://timdettmers.com/2015/07/27/brain-vs-deep-learning-singularity/comment-page-1/#comment-24006 Thu, 16 Nov 2017 20:58:44 +0000 https://timdettmers.wordpress.com/?p=312#comment-24006 In reply to John Smith.

Not all viruses work like this, but retroviruses work like this in general.

]]>