Artificial Intelligence Post 33

I’ve said that many areas are in the process of incorporating AI, and it is just a matter of time until they come to fruition.  Here is a company that seems to have a jump on their competition, and here is an article relating to what they are doing: http://ophirgottlieb.tumblr.com/post/145065914534/the-secret-is-out-this-company-is-quietly.

Here are some excerpts from the article that I thought noteworthy:

“Machine learning is the process of turning data into a computer program’s decisions…Once the basic algorithm is written, the system learns on its own. The market’s top technology companies are collecting peta-bytes of customer data, satellite images, web site content, weather sensor data, and much more. What’s left to turn all this information into cars that drive themselves, drones that carry out autonomous missions, online shopping platforms that know what you’ll buy before you do and the next step in human evolution that is virtual reality? Massive, massive, processing power. Enter NVIDIA (NASDAQ:NVDA).”

“With Nvidia, we are looking at machine learning, artificial super intelligence, big data, gaming, self-driving cars, virtual reality and a lot more.  Nvidia…literally invented the graphics processing unit (GPU) which now serves as the visual cortex of modern computers and artificial intelligence.”

“NVIDIA has its hands in the most innovative segments of technology… Deep learning is the secret weapon for self-driving car algorithms. NVIDIA already calls Tesla, Aston Martin, Rolls Royce and Audi as customers.”  The article also mentions other areas with great growth potential for NVIDIA.

Note that I don’t own this stock (but I do have on order a Tesla Model 3 which will likely use Nvidia’s technology.)  I only mention this company as an example of how AI is beginning to be integrated into our everyday lives.

There has been some discussion on how AI may never be able to truly “think” any better than a human.  That may be.  But there seems to be little doubt that given enough devices (sonar; radar; GPS; visual cameras ahead, side, and behind; etc.) and enough computer power and comparable software, that self-driving will eventually be far superior to humans.  Note that I did not say perfect.  Humans are quite poor at driving, hence the over 5 million car crashes in the US per year and 32,000 deaths.  All AI has to do is be substantially better than that.

One of the readers asked me to look at a conversation with Gary Marcus on edge.com. The conversation was so broad that it is hard to comment.  But here are a couple of observations.  Gary kept comparing AI to humans as if humans are all that good at either decision making or thinking.  We are not all that good at either!  And Gary kept inferring that the brain is so superior that we may never match it.  Well, we may not have to.  Box jellyfish have no traditional brains, but they have 24 eyes, 4 of which are quite sophisticated.  They are able to do what they have to do in their environment quite well, without a traditional brain.  So it will be with AI.  Gary noted how auto-drive is only sufficient in very controlled conditions (i.e. highways with well-marked lanes.)  They don’t do well in snow, fog, etc.  Well, nor do we!  But the algorithms being used in auto-drive are self-learning, and they may very well learn that their current sensors along with identifying the location of bridge abutments, light poles, trees, and other cars, and by using advanced GPS, that they can find where they are on the road far more effectively than a human driver in snow, fog, rain, etc.

As for very advanced thinking, we aren’t very good at this unless we have experimental data. After all these years, we still don’t understand the brain. And scientists trying to understand the Big Bang have had to add in an initial expansion that exceeds the speed of light and to incorporate their recent “findings” that the universe is expanding at an ever increasing rate.  They have included a fudge factor called black energy.  They have no idea what this is, and when Einstein did something similar he later admitted this was the dumbest thing he had ever done!

Despite all that, AI is progressing quite nicely.

As promised regards the stock market, here is an update on Elliot Wave’s prediction that the S&P 500 will drop 41.6% this year from 1880 to 1100.  I said that didn’t believe it!  The current S&P is 2102, up 11.8% since EW’s prediction early this year.

 

Advertisements

21 Responses to “Artificial Intelligence Post 33”

  1. Robert Kaufman Says:

    Here is a link to info on a new chip that might have value for AI and you might find interesting.

    https://www.knupath.com/

  2. wbrussee Says:

    Robert,

    The article gave little detail, so I cannot compare it to other chips claiming to do similar things. But there is no doubt that the growth of ever more powerful chips continues.

    From what I read there seems to be little question among computer experts that AI is coming. The biggest questions seem to be whether man will be able to control this technology and to what degree man will be made subservient.

    Elon Musk in a recent interview said that the most benign outcome is us being like household cats. I don’t believe it because I think that computers for the foreseeable future will need us to make spare parts and fill the function of very maneuverable robots. Mechanical device automation is far behind software/computer advances. It is sort of funny that our minds, which we think are so advanced, may be less important to our future survival than our bodies.

    • Bob Says:

      Thank goodness I get a kick out of being with my wife and child because between the threat of resource depletion, destructive climate change, terrorists using nuclear weapons, anti-biotic resistant infections, global economic collapse and the on-rush of artificial intelligence – I am finding hard to find my happy place.

      Does anyone know how this can work out without mass suffering?

  3. Bill Says:

    http://gulfcoastcommentary.blogspot.com/2016/06/the-student-loan-racket-in-action.html

  4. Albert Says:

    Warren,

    What’s your opinion regarding how the Brexit will effect the U.S. and world economy? Do you think this will lead to additional European countries exiting the Euro? Do you think the Brexit could cause a U.S. stock market crash and bear market?

  5. Bob Kaufman Says:

    Look at the new releases for a stock called Biotime [BTX] and you will see a interesting use of AI as it relates to regenerative medicine. The release is titled

    Biotime Presents Online Resource for Applying Artificial Intelligence to Stem Cell Biology at Mensa 2016 Annual Gathering.

  6. wbrussee Says:

    Bob Kaufman says “Look at the new releases for a stock called Biotime [BTX] and you will see an interesting use of AI as it relates to regenerative medicine.”

    The march towards using AI in many areas is well underway. However, there are going to be a lot of stumbles along the way. I don’t know if you have been following the articles related to the first death in a Tesla car while their Autopilot Driver Assist was being used. A trailer truck crossed in front of a Tesla and neither the driver nor the Autopilot braked the car. The car went under the trailer, cutting off the car’s roof. Apparently the driver was not paying attention (despite specific warnings from Tesla on how the system is only a driver assist) and the Autopilot did not see the truck due to a white trailer in front of a white sky. Even though the preliminary death rate per mile in a Tesla using the Autopilot is lower than the death rate per mile for US cars on average, many people are accusing Tesla of putting an unproven and not perfected system in the customers’ hands, and some people are demanding that Tesla disarm the systems in all their cars.

    Any use of AI by Biotime in regenerative medicine will have the same risks, no matter how good it is. Someone, somewhere, will feel that they were harmed by the technology, and they will sue.

    Since data from all Tesla cars is gathered and then incorporated into improved versions of Autopilot, the system will constantly improve (versus human drivers). It is therefore imperative that the Tesla system be allowed to continue in use if we are to make rapid headway in driving down the 33,000 highway deaths the US experiences every year. I presume that Biotime and others that incorporate AI systems are going to have similar incorporation needs to learn and improve.

    Incidentally, I am not sure that I want to live in the resultant world if Biotime can actually reverse the aging process. Can you just imagine the scenario that the first people getting this treatment will be the very wealthy, with all the conflict that will start? And then, if it becomes universally available, what will happen to our world population, which is already out of control? Maybe this is the way humanity will finally kill itself!

    • Bob Kaufman Says:

      Extending life well beyond 120 years and vitality well beyond 90 years appears to be a technical problem. I assume we will eventually gain the knowledge required and perfect the technology to make much longer, much more vital lives a reality. I do not think we have to worry about you or I living in such a world. We will most likely be toast well before we get a handle on how to do this. However, I think it is highly probable that we will gain the ability and use it at first for the rich and eventually, possibly, in a more widespread manner.

      If we keep reproducing at present rates and increasing our use of natural resources per capita as we have been, then the resultant suffering will be massive and the extinction of the race is possible. You have heard me say I am finding it hard to find my happy place when I think about the probabilities here. That’s why!!

      We must find a way to change human nature and the cultural norms in which humans exist. We must reduce our population by billions, convert from fossil fuel to clean energy and find a way to be happy with sufficient but much less abundant food, water, and shelter.

      How do we create an economic system and set of cultural norms that works for a world running out of resources that is already way over populated when we learn how to extend our lives to 200, 300, 400 and beyond? How do we change a set of self-selfish, short sighted human behaviors that are the result of 7000 centuries and 3500 generations of humanity? I can’t figure a way out of the path we seem to be on. Perhaps some self-programming, super-intelligent network of computers can pull a rabbit out of the hat and save our “arses”.

      I just hope we don’t end up being the pets of a super race of our own creation.

  7. Robert Kaufman Says:

    That should have been 700 centuries; not 7000. Typing and proofing are not by strongest suits.

  8. Bob Kaufman Says:

    My cousin is 80. Just totaled his car although he was unhurt. My wife’s aunt was 90 and died in a car crash that was her fault. I am 67. Come on autonomous automobiles.

    • wbrussee Says:

      I am 75 and in great health. But I try to limit my distance driving just because I know that my responses are not as quick as when I was 40, especially when I get tired! I am planning to get more automated driving assists in my next car. I have an electric Tesla Model Three on order. Almost 400,000 people have put $1,000 down for one of these cars that aren’t even scheduled to start production until the middle of next year. But I am fairly early on the waiting list because I went to Charlotte (my son-in-law drove) and stood in line for over two hours to put down my deposit before even seeing the car. Tesla has had some questionable press about some incidents that have happened to their cars with or without their Auto-Pilot engaged. But even at this early stage in development, its use makes it safer than a car without this level of driver assist. Given that many people are misusing the systems by not paying attention to the road at all, it has done remarkably well. And, given that the cars give continuous feedback to Tesla as to their driving experiences, these systems are being improved continuously. By the time I get my Model Three, I expect that the systems will be quite remarkable.

  9. Robert Kaufman Says:

    Here are three recent links from a website called Futurism that relate to advances to AI. Just keeping in touch as I am still interested to developments in this space.

    http://futurism.com/computer-memory-could-speed-up-by-1000-times-thanks-to-new-computing-breakthrough/

    http://futurism.com/the-future-of-quantum-computing-is-almost-here-research-extends-qubit-life-10x/

    http://futurism.com/dawn-of-synthetic-reason-deepmind-can-learn-from-its-own-memory/

    • wbrussee Says:

      The articles you referenced on computer advances are still in the early experimental stage, so it is hard to comment on them. But the article on Google’s Deep Mind AI is interesting, in that this is a continuing series of steps that show that Deep Mind can figure things out by itself without being fed all the possible results by a human or from existing results in data. This certainly is a big step to showing that AI can actually reason!

      I have been following the work that many of these systems are doing, including Watson. Most of Watson’s work is proprietary, being done on specific problems that are not being made public. So it is hard to measure real progress.

      One of the most visible areas to measure gains is with Tesla’s progress on automatic driving. Elon Musk has just recently announced that they have changed their earlier approach of using visible images and lasers as their primary inputs, to using radar. In the past, radar images were thought to be too fuzzy to use other than as a reference. But Tesla, working with their vendor, has apparently figured out a way to squeeze more information out of these images, such that the car can now “see” in a fog, and even “see” the car two cars in front of them. So Tesla car owners will not only get the benefit of a system that doesn’t get tired or distracted like a human, but also can see things we cannot see. It plans to couple all this with maps that are continuously updated by the vehicles themselves. And the computers are already capable of handling all this data. By this time next year I hope to have my Tesla Model 3 with all these goodies and more!

      I suspect that similar advances in medicine, investing, law, and other areas will soon be shown to the general public. None of this is true AI. But, as I have said before, I believe that AI will develop incrementally, and finally be integrated into massive computer systems, perhaps similar to ones described in the other articles you referenced. Then true AI can begin to happen!

  10. Bill Says:

    We could use some AI. Apparently we aren’t smart enough to deal with our problems.

  11. Bob Says:

    Did AI influence the election?

    https://www.washingtonpost.com/politics/after-working-for-trumps-campaign-british-data-firm-eyes-new-us-government-contracts/2017/02/17/a6dee3c6-f40c-11e6-8d72-263470bf0401_story.html?utm_term=.fa4fa3a5bb36

    • wbrussee Says:

      I see no evidence of AI anywhere yet. Sure, loads of data gathering/analysis, but it all seems to be human directed.

  12. Don Jenver Says:

    What’s a bigger near term threat? AI or climate change?

  13. Don Jenver Says:

    What would be your earliest prediction for AI becoming a problem?

  14. Don Jenver Says:

    Do you think that any world leaders are aware of the issues with AI? Do you think they are preparing in any way? Are they likely to become more authoritarian? They might have to be.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: