You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!




Results 1 to 6 of 6

Thread: Einstein on Learning, Semis and Explosive Compounding Progress

  1. #1
    Expert
    Join Date
    Apr 2013
    Location
    East SF Bay Area
    Posts
    1,567
    Thumbs Up
    Received: 453
    Given: 429

    Einstein on Learning, Semis and Explosive Compounding Progress

    Einstein's very famous quote on "Never memorize something you can look up" needs to be revisited in this age of almost limitless electronic memory as it relates to work, education, training and intelligent use of time, our most valuable asset. Our ability to look things up is no longer so simple with programs that not only help us look things up, but apply them. We need an entirely new structure where we abandon traditional learning and instead are taught how to use intelligent programs to develop even more intelligent programs. The old way of teaching and learning except the very basics is becoming more obsolete every day. Modern educational programs like Trailhead and others can take us far further and faster than traditional education/training. The ability to use intelligent platforms to create new ones will compound on itself and leave traditional organizations using traditional methods farther and farther behind as they can't compound their skill sets at even near the rates of an intelligent platform that now only automates the process, but further compounds knowledge through real time collaboration where the program itself gets more intelligent every time it's used. The future belongs to those organizations and people that use AI/ML platforms to leverage their work, just like you use a lever to lift a weight. The tradition bound forms of education and training are soon going be buried and become more obsolete at an ever increasing rate as intelligent platforms in every field using ever more sophisticated cloud based processing and memory that advances the platform every single time it is used. There is no need to waste our time or processing ability duplicating ours or others efforts when with an intelligent cloud platform with ever increasing processing and memory power has the ability to advance literally with every single action. We are already on our way in this direction and anyone that gets in the way will face extinction or irrelevance at an ever increasing rate in the "Great Acceleration". Using our time, the greatest single resource in our lives and the universe wisely and efficiently should be our ultimate goal and Albert Einstein made full use of this with what he had at the time. We now have the resources to literally have a compounding explosion of knowledge like a nuclear weapon has a compounding of energy. It is time to use our resources with all the tools the semi world has put at our disposal. We can no longer afford to waste time and resources on obsolete models for special interests bound by the past.

    Examples of new knowledge handling

    Is The Enterprise Knowledge Graph Finally Going To Make All Data Usable?

    a-j-hanson@att.net

    1 Not allowed!
    Last edited by Arthur Hanson; 09-19-2018 at 04:51 AM.
     

  2. #2
    Top Influencer
    Join Date
    Jul 2014
    Posts
    120
    Thumbs Up
    Received: 61
    Given: 28
    I still think you need to master enough knowledge to orient yourself, otherwise the things you look up are simply factoids, unanchored, conveying little understanding. Now, I am for sure appreciative of the net. It has helped me look up and absorb things as diverse as the cell wall genomics of tuberculosis and the best way to build non-linear solvers for industrial sized problems. And many other things.

    But it is hard to imagine how I could have been effective at that without the solid base of a thorough UK education including several years of hands on lab work, a broad layer of theory, and a lot of practice working through actual problems to get used to marshaling it all.

    Einstein had an excellent education, he was a bright student if a rebellious and libidinous one. His work was founded on solid skills. The things he looked up were finishing touches, not basics. Although sometimes he needed deep collaboration like learning the tensor math needed for general relativity where he leaned heavily on his more mathematically adept friends to tutor him. Perhaps that counts as look it up too. Which is fine - just don't expect all the lookups to be superficial. Sometimes he took years.

    0 Not allowed!
     

  3. #3
    Top Influencer
    Join Date
    Jul 2014
    Posts
    120
    Thumbs Up
    Received: 61
    Given: 28
    And as for the question, is all data going to become useful?

    Roughly when all questions are the right question.

    0 Not allowed!
     

  4. #4
    Expert
    Join Date
    Apr 2013
    Location
    East SF Bay Area
    Posts
    1,567
    Thumbs Up
    Received: 453
    Given: 429
    Tanj, I'm not referring to just facts, but more about guiding yourself through an intelligent platform, even guiding it through virtual experiments. This is about using the very most intelligent people to build platforms to leverage their talent. As a project head on more than a few projects I was being buried alive in questions and inquires until I told everyone if you before you come to me come to me with a solution even if you didn't know if it would work. It cut down what was an excess work load almost immediately. Building a platform is far better than giving orders or instruction. Leverage, compounding and memory are all key. Building a platform for solutions is the best way to leverage talent and building and guiding platforms are going to be the skill sets of the future. This is a short, truncated explanation, but it works. Building a platform as you create and go into entirely new areas is going to be important or more important than discovery. Making an intuitive platform is going to be the most important skill set for the future. Any comments or thoughts would be appreciated. I have always been considered "Different" by teachers, professors and coworkers.

    0 Not allowed!
     

  5. #5
    Influencer
    Join Date
    Sep 2011
    Posts
    95
    Thumbs Up
    Received: 30
    Given: 2
    As you know, I think AI is total hype. Using programs to
    generate training examples does not work. I am working on a
    paper criticizing AI so I have some examples.

    First, there is an interesting recent purchase by the
    robotic surgery company Transenterix that makes a machine
    called Senhance that allows laparoscopic surgeons
    to operate with machine assistance (web site
    Senhance Surgical Robotic System by TransEnterix). It purchased Israeli
    company MST Medical Surgery Technologies. MST makes an add
    on device (sort of like Specman-e add on to Verilog) to improve
    laparoscope camera focus and aiming. Here is the PR:

    TransEnterix Acquires Assets, Intellectual Property and Retains R&D Team from MST Medical Surgery Technologies | Business Wire

    Transenterix is not going to market the add on device but
    instead it will add the software to their robotic Senhance
    surgical system. In my view this is an example of the best way
    to add computer processing to machines. MST applied the same
    improvements in projective geometry algorithms used in automobile
    collision avoidance systems to surgery. There is no AI in this
    at all.

    Another example is that AI is believed because papers that
    criticize it are not getting accepted by academic journals.
    One example is the paper "Distilling free-form natural laws
    from experimental data" by Schimdt M. and Lipson H. (2009)
    that was published in Science (324:81-85). A reply refuting
    the paper "Comment on the Article "Distilling ..." by
    UC Berkeley scientists Hillar, C and Sommer, F. was
    "eventually" rejected by Science. Here is
    the arXiv reference for the refuting paper: arXiv:1210.7273

    0 Not allowed!
     

  6. #6
    Expert
    Join Date
    Apr 2013
    Location
    East SF Bay Area
    Posts
    1,567
    Thumbs Up
    Received: 453
    Given: 429
    smeyer, AI/ML are both still in their infancy, like a Model T stage for comparison. With 300,000 AI researchers with massive computer and collaboration(which I find is key to speeding any advance) and more of all being added, we will see progress at a rate mankind has never seen in any other endeavor. I believe that within five years, we will start seeing changes in literally everything we come in contact with at a rate not even imagined. IOT is increasing dramatically the amount of data we have are increases in memory and computational performance of all our devices. The world and society are going to be changed in ways we haven't even imagined. I hope that those developing and implementing the technology have the wisdom to use it to benefit all for this will be the greatest power mankind has ever had due to its ability to increase leverage in just about everything. Any comments or additions you have or even counter arguments are solicited and welcome, operating alone is not the best way to advance and I truly appreciate outside input.

    I have been called out when I have proposed solutions to others, including teams of engineers once in a while and solved problems they haven't had solutions to. I'm never afraid to do extensive research and have been reading at least a thousand pages a week or about ten years and up to three thousand in spurts. I basically read for at living and put my own money on it. I have not worked a job or contract for ten years.

    0 Not allowed!
    Last edited by Arthur Hanson; 09-26-2018 at 12:23 PM.
     

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •