RE: Not a good time for humanity, the ai singularity is here
May 23, 2017 at 1:11 am
(This post was last modified: May 23, 2017 at 1:21 am by MellisaClarke.)
(May 22, 2017 at 5:54 am)Aoi Magi Wrote: Just wanted to share this
Yes, I am subscribed to ColdFusion, and a few others, like 2 minute papers.
(May 22, 2017 at 1:55 pm)Tazzycorn Wrote:(May 21, 2017 at 8:15 pm)MellisaClarke Wrote: They are about to replace 75 - 80% of IT jobs:
https://futurism.com/80-of-it-jobs-can-b...-exciting/
It looks like you're correct though, in terms of mapping brains, we haven't gotten to chimp level yet. (Gotten as far as rats)
Why do you continue to post links to a website that has nothing but bullshit articles designed solely for clickbait?
Sorry, was trying to provide relevant information.
I think articles like those are preparing the public for the rapid loss of jobs due to automation.
As an econ person myself, that was not clickbait for me.
(May 22, 2017 at 1:26 pm)Mathilda Wrote: I presented a paper last month at a generic AI conference. I wrote the code and paper in my spare time over the last 7 years so I could do something really new. Last time I went to that conference was 2005, 12 years ago. It was interesting because there were far fewer people this time, and that even got mentioned at the AGM. General opinion was that most people were going to machine learning conferences. Machine learning is a limited form of AI, in the same way natural language processing is. Other reasons for low attendence is the lack of funding in academia. This was a conference in the UK.
Second interesting thing that I noticed was the complete lack of progress. They were talking about exactly the same things as 12 years ago. As far as I'm concerned we haven't had much progress since the mid 90's if you exclude advances in processing power. Intel reckons going below 5nm does not make economic sense. We're currently at 14nm so forget about Moore's law continuing.
During the panel discussions about the dangers of AI, they made a lot of good points. In a hall of about 300 people, absolutely no one mentioned the hyped up ideas about singularity from media whores such as Ray Kurzweil, Sam Harris, Elon Musk or Stephen Hawking.
The singularity is not here.
I saw a talk by Ray Kurzweil recently.
I thought Moore's law as also ending, but he described that Moore's law is not only about hardware changes, but also exponential software returns.
I am betting on Kurzweil's predictions for now.
With things like NVIDIA tpu systems, we are actually going a bit faster than Moore's law.
Kurzweil also says that we'll go into other types of computing, to maintain speed.
