The LENR Revolution and Robots (Albert Kallal)

The following post has been submitted by Albert D. Kallal

The promise of LENR and the delivery of low cost energy is not only the holy grail of physics, but a story that has the potential to change the world as we know it.

Energy = high potential standard of living. The instant mankind replaced muscle power with chemical power was the instant we became freed from our limitations. The result was the rise from near slavery conditions to grow our food and provide the resources to that of a high standard of living. At least for those that have energy.

Today, even high population of counties like China are able to feed their large population. In fact we even see some of those areas now exporting rice and other food stuffs. A quick glance on the grocery shelfs and was surprised at the variety of foods coming all the way from China to be sold in my native country of Canada. All this food is the result of energy.

It is without question that adoption of energy is what feeds us, clothes us, and keeps us warm. Our use of energy is a significant part in near everything we consume.

There are three areas that I see as booming and benefiting from LENR.

First up will be the ability to grow food indoors. Even today we see an “active” market rising in indoor growing, and controlled climate production of food.

In rather hot climates, such energy can control the climate, and of course provide the needed water (from sources like the ocean). So as part of this indoor growing revolution, I include the ability of LENR to provide clean drinking water from the ocean.

And in cold climates, again LENR can provide the energy to grow food year round. In my local market, local grown cucumbers of excellent quality grown in greenhouses year round are about $1.60. This occurs without pesticides and even use of soil. (And sure, they are about 45 cents more than other products, but they are local grown, and their quality is beyond outstanding.

Each cucumber is in perfect condition – they are grown in perfect controlled conditions.

In the winter time due to increased heating costs, the same product almost doubles in price. So this area of the economy will boom with low cost energy inputs.

So with low cost energy, the ability to grow food just about ANYWHERE on the planet will the result of such an energy revolution. And such food grown this way is achieved without soil, and without pesticides and without soil stress (you really don’t need soil). Who needs to plow a field to grow such food when you have greenhouses? And automation is FAR more possible with a control environment then that of traditional farming.

The other revolution of course would be transportation. Transportation and movement of goods is mostly driven by chemical (fossil) fuels. LENR can transform this area.

And speaking of automation?

While I feel the above two areas benefit the most from LENR, the largest impact in our lives will be that of Robots.

I don’t mean the factory robots, but autonomous ones – ones that can run around and deliver things, and help handicapped people at home.

The “major” limitation for such robots is of course the power source. Unless some amazing battery technology breakthrough occurs, then such robots will require a power source.

Imagine a “swarm” of small drone travelling into a forest. With intelligent software and vision ability (like a typical smartphone), such small drones could spend their whole day and night hunting for pine beetles, and zap them with a low power laser. Thus you can eliminate pests without pesticides.

Once again, battery technology not up to the task, but a LENR power pack is.

Even better is robots to help humans.

Here is an excellent video that explains outlines where robots are going. We are VERY close to such robots resulting in a practical and significant contribution to society – but such a revolution cannot occur unless we find a new power source like LENR.

Boston Dynamics: The Coming Robot Revolution – Marc Raibert

The above is a great video – since it gives hints of how far robotics has progressed in the last few years, and what is in store for us. However without LENR, I can’t see how such machines will be powered.


Albert D. Kallal

Edmonton, Alberta Canada

  • Thomas Kaminski

    To food production, I would add clean water sources. Perhaps LENR can help with de-salinization and water distribution to those areas undergoing severe drought. Also, don’t forget basic heating and cooling environmental controls — they consume a good fraction of the energy use in advanced nations. Food, water, shelter — basic human needs all requiring energy.

    Robotic energy sources are important, but so is powering the “cloud” computing. It is estimated that powering and cooling distributed computing will be increasingly important as more AI is used to enhance the robots’ ability to understand and interact with the environment.

    • roseland67



      I have built power distribution systems for data centers all across the states, and have seen power consumption go from
      30 Watts/ft^2 to well over 200 in some of the newer ones
      Multiple utility feeds, rows of generators, banks of battery’s, thousands of tons of rooftop chillers or cooling towers, all to keep the data center running 24/7.

      The amount of power and energy required is staggering and will only get higher.

      And to think these types of facilities did not commercially exist 25-30 years ago.

      • orsobubu

        Funny to remember all that old reviews of Asimov’s sci-fi novels, always describing the steadily wrong vision of future worlds by him and other authors; in particular about computer dimensions, Asimov predicted computer sizes growing til encompassing entire planets and absorbing the energy of stars, and this contrasted with actual chip miniaturization. Now he takes a sort of revenge…

        • Albert D. Kallal

          To be fair, Popular science had us living in bubble homes from the 1950’s to now.

          I will in fact say that computers are about the ONLY thing that has come true in terms of science
          fiction. Well sans the self-aware thinking computers!

          So the small computer devices we have today combined with communication really is science fiction
          come true. I think computers are about the only technology that really delivered on science fiction promise and predictions!

          And LENR would be another science fiction dream and prediction come true!

          Albert D. Kallal
          Edmonton, Alberta Canada

      • Albert D. Kallal

        Yes, they need two things:
        1 – Power
        2 – relative availability of lots of water (and water of drinking quality).
        A slam dunk for LENR!

    • Albert D. Kallal

      I likely should have included drinking water as a big possible change by adopting LENR. that alone could be the greatest contribution from LENR.

      And if LENR does not pan out? Well, then thorium reactors will be our only real great choice.

      However, I don’t see such thorium reactors as being scaled down to small devices like LENR can be. So a new battery technology will have to come along for those robots if LENR don’t pan out.

      We can well provide our future energy needs with nuclear (thorium), but I would rather LENR pops up first and grabs our time, money and resources to provide our future energy needs this way before thorium takes off.

      If LENR don’t pop out of the bag soon, then the thorium revolution will occur and beat LENR to the punch.

      Albert D. Kallal
      Edmonton, Alberta, Canada

  • sam

    November 27, 2017 at 12:33 PM
    Dear Dr. Rossi,

    What I gather from the Nov.24 presentation is, that there is still some R&D work to be done on the Hot Cat. Can you give a time estimate, even tentative, of when the E-Cat QX is finally going to be ready for mass production?

    Kindest regards,

    Andrea Rossi
    November 27, 2017 at 3:01 PM
    I will do all my best to make it happen within 2018. I hope my optimism will become truth.
    Warm Regards,

    Andrea Rossi
    November 27, 2017 at 3:40 AM
    Martyn Aubrey:
    Q1- It was also so, eventually resolved
    Q2- We are working on the miniaturization and heat elimination of the control circuitry system. I am hiring specialists.
    Thank you for your attention to our work and for your suggestions.
    Warm Regards,

  • AdrianAshfield

    Here is a piece I wrote for the DelcoTimes in October that covers some of the same ground

    • Pekka Janhunen

      I think transistor-level Moore’s law came to halt already 5-10 years ago. What still continues is some rationalisation of processor architectures. While R&D of alternate electronics systems continues, it is not clear to me if and how Moore’s law could continue. My baseline estimate is that it won’t.

      • AdrianAshfield

        I said “couple of years” Moore’s law is currently running at about 2.5 years to double transistor count or microprocessor capability.
        It will slow but is currently exponential and still changing things very fast. There is little advantage for most people to have computers run faster than they are now but for AI and robotics it is a different story.

        • Omega Z

          The last vestiges of Moore’s law is taking place as we speak,(no room left to shrink) however, those will likely be a couple more years making it to the general population. Intel has already stated transitioning to new tasks such as how many chips can you stack and stay integrated without total melt down.

          If they wish to speed computers up more, they will need to find a new medium.

    • TomR

      A very good article Adrian. You are 100% correct about our politicians not doing any of the jobs they should be doing. They think the election was an anomaly, just wait until the next one.

    • Albert D. Kallal

      Yes, great article. And did not dwell a lot on health care (good for you!).

      Even the above video does touches this issue, and points out about 30 million people in the USA spend about half their time taking care of other people. Even with some medial breakthrough in genetic mapping, and some changes in health care? Well, we STILL are going to need large amounts of labor for assisted living.

      So even with major amazing breakthroughs, those people will STILL eventually get old – maybe REALLY old with some of the cool stem cell research going on. However, you still eventually will die, and there is a period of old age that requires assisted living.

      This is where robots could really help, let alone application of technologies to the health care field.

      Albert D. Kallal
      Edmonton, Alberta, Canada

  • Good thinking. You might also read the last chapter of my book An Impossible Invention where I mention this and more. High density distributed energy will change most things, as you say, from society, health and welfare to military operations and geopolitics.

    • AdrianAshfield

      Mats, I have your excellent book.
      Were you referring to the original article or my piece below?

      • Ah, sorry Adrian, I haven’t read your piece yet. I was referring to Albert’s article.

    • Albert D. Kallal

      I think the automated and controlled food growing systems is one of the great results of LENR.

      As for Robots? Well the interesting question THEN becomes is what kind of power pack?

      Readers here are well aware of my views on Stirling engines. However, for robots, Stirling engines
      are not compact (enough). For an indoor appliance, I think a combined heat, air conditions,
      and LENR system makes sense based around a Stirling engine. They are sealed units, and can run for 10+ years like say a refrigerator.

      However, for a robot, Stirling’s are too large.

      So, what kind of compact device can convert heat into electric?

      The answer I simple: A small high speed turbine.

      Such turbines only have 1 main moving part, and you can use “air” as the hydrodynamic fluid for
      the lubrication system.

      And compact? Well, take a look at the size of this 400 watt turbine in this following video – and it spins at 400,000 rpm.

      In fact the designer in this video “missed” this “big” concept of using air for the main bearings.

      Other than using oil in place of air, this video shows quite much like what a robot power pack based on LENR
      will look like.

      Note that “most” of the suitcase is for fuel, control systems, starter battery etc. Toss in some miniaturization,
      removal of all that extra stuff, and your result will be a rather small power pack indeed. (and likely with even more power then this 400 watt example).

      Albert D. Kallal
      Edmonton, Alberta Canada

      • Thomas Kaminski

        One other thing that the QX can provide is light for plant growth. At one point, Rossi said that the device could trade heat, light, and direct electrical production.

        For robotics, it might be possible to use thermal actuation to create motion. See this interesting and simple “muscle”:

        • Albert D. Kallal

          Such technologies will certainly be preferable to hydraulics. No matter how we slice this “power pack”
          problem, the robot can really benefit from LENR.

          Even better than some mechanical conversion of LENR to electric would be a solid state conversion
          (thermoelectric). So anything that can eliminate moving parts in converting
          heat into motion is something that LENR will take, and run with.

          In fact it was some years ago that a “big” fascination occurred in regards to memory metals. Some
          even built motors based on this concept:

          Albert D. Kallal
          Edmonton, Alberta Canada

        • Omega Z

          Nitinol shape memory metal has a lot of potential.

  • Rémi Andre

    “Science sans conscience n’est que ruine de l’âme” – François Rabelais (Gargantua 1534) “Science without consciousness is the ruin of the soul ” that’s what this video inspired me (imagine those machine LENR powered)….


    • psi2u2


  • Piper

    Cellulose, a hydrocarbon, is perhaps a ready source of hydrogen for a foraging hungry robot?

  • Pat

    Also, an unlimited clean energy source can change the amount of “burning” going on in the world as well as possibly power up machines globally to sequester CO2 from the atmosphere or maybe convert it back to O2 and carbon. But yes, what an interesting world my grandkids will experience with robots roaming around run by quantum computing AI and animated by unlimited power. The role of humans in such a future is (to say the least) highly unknown.

  • sam

    Italo R.
    November 29, 2017 at 1:47 PM
    Dear Dr. Rossi, you have written:
    “…From now on the focus will be on the industrialization.
    Probably today we reached a very important agreement toward it…”
    Is there any news you can share with us if possible?
    Kind Regards,

    Italo R.

    Andrea Rossi
    November 29, 2017 at 4:32 PM
    Italo R.:

    Several days after the Stockholm demo we made a very important
    agreement, that will make faster the start of a massive industrial
    production. These few days have been momentous.

    Warm regards,

    November 29, 2017 at 10:24 AM
    Dear Andrea,

    how many ECAT-QX (maximum) is it possible to feed with the controller you used during the demo?

    5, 10, 100 ?


    Andrea Rossi
    November 29, 2017 at 4:25 PM
    With a controller used in the demo we can control at least 100 systems.
    Warm regards,

    • Steve D

      Both interesting postings Sam. It somewhat explain the large size of the controller. However “made a very important agreement”. Does that that mean a contractual agreement or we just agreed on a good idea, direction etc with no commitments/obligations?

  • Alan DeAngelis
  • Gittyup

    Some good comments but you are vastly understating the impact of a low or near zero cost clean energy source would do to civilization.

    Almost every single aspect of society is based upon scarcity of energy and conservation of it. Think any thing anyone does throughout the day and it’s easy to see how the decisions and costs are effected by the the cost of energy. From the buildings we live in , the choice of food and availability of certain foods, the goods and cost to produce them etc. Energy is 80% + of the cost of everything in the world. Overnight everything would be cheap and anything possible.

    It’s almost hard to think about the impact it’s so intertwined into our society. at its most basic impact would be raw materials would basically be free, transportation would be a fraction of the cost, exotic materials would be standard for everything, food basically free, supercomputers making everything that much more efficient. Crazy world

    • Pekka Janhunen

      It’s indeed hard to overestimate the importance. The available energy flow basically defines what life can do, and for example when life invented photosynthesis a few billion years ago, everything changed and the change is not even complete yet since we are still optimising agriculture etc. But going from chemical to nuclear is a much bigger relative change than that.

    • Albert D. Kallal

      I 100% agree.
      And we likely will “miss” some of the things that will occur as a result of such
      low cost energy. I mean, you take a 10 million dollar computer from 1985,
      shrink it down to a postage stamp? Well, we would just assume that “large” computer
      rooms would be a think of the past.

      What happened?

      Well, things like smartphones came along! In other words “many” inventions and things will occur
      that we don’t predict. We really did not see smartphones coming as a result of
      low cost high density computing.

      And low cost high density “energy” will also transform society and touch everything. I think key
      is not that we have a form of energy, but the HIGH density is a key concept (at
      least in regards to new industries emerging from a LENR power source).

      However, in my post I attempted to not identity the general change to “near” every aspect of our
      modern society, but what “new” or emerging industries will benefit, or become
      common place. (like smartphones in regards to computing).

      So greenhouse technology, and that of Robots are two things that I think can become a commonplace with

      So to be “fair”, these “new” industries are “interesting”, but I would certainly concede and see your point about virtually
      every aspect of our daily lives with “existing” technology will benefit from LENR.

      I been following the LENR story for about 7 years now (and I think many here are in the same
      boat). We just don’t know how soon how LENR can become a practical reality.

      It is somewhat disheartening to dream about all these amazing future effects when in fact we are not sure
      how close mankind is to cracking this LENR magic stone of energy.

      Like most here? We all hoping for this event!!!

      Albert D. Kallal
      Edmonton, Alberta Canada

    • Omega Z

      ->”Energy is 80% + of the cost of everything in the world.”

      No, It’s not. Fossil energy makes up less then 10% of GDP.
      While energy is important to nearly everything in modern society, the primary cost of almost everything is labor. Including the extraction of fossils that provide energy.

      • cashmemorz

        More than half of all work done is done by machines currently. As a former machine design drafts person, I have seen the market for automation go through the roof. Where ever it is possible to automate anything, it has not only been done, but the remaining areas are the most intricate mechanically or kept in human territory by various reason no in line with pure economics. Those areas are also getting figured out so as to allow it to be automated. Labor is just the moving around of objects be they large or small objects. The biggest blockage to get the rest automated is unions to protect the remaining jobs for those qualified for doing only labor intensive tasks, which would have been automated a long time ago if not for the unions and other social concerns for those technologically disadvantaged. So we have a catch 22 where the very method, namely LENR, would make life easier for everyone is the one that is being kept from automating the last bits. And it is being kept from helping the very people who would be helped by such a LENR powered world. Those people would not need to work where the economy could pay them a guaranteed income to do something more productive with their time.
        It ends up as a slow evolutionary process, where the remaining workers in those last to be automated areas first have to retire and be immediately replace by LENR powered machinery.

        • Omega Z

          “where the economy could pay them a guaranteed income to do something more productive with their time.”

          Playing video games, virtual reality, never needing to leave their homes finding themselves miserable with no purpose in life. That’s what millions will experience. Some peoples idea of utopia may turn out to be the worst nightmare of others.

          Fortunately, automating will continue to be a long drawn out process. It’s been my experience that business does not automate just because they can. It is usually done only to remain a competitive and viable business. Sometimes it’s because product requirements(Precision etc.) are beyond the human ability.

          Another consideration. With an aging society, robotics and automation will become necessary just to meet the needs of society. There simply isn’t enough able bodied to do all the necessary work ahead. Aside from the 3rd world needs, the developed countries are due for a near complete rebuild. Nearly all infrastructure is nearing end of life and needs to be replaced. Over 100 Trillion$ worth in the U.S. alone. This will require decades. LENR will merely make it cheaper. Even LENR technology will not be free.

          • cashmemorz

            Ok, then put those who won’t or can’t find something useful to do, into those projects of replacing aging infrastructure. Until small smart machines can do it. There will come a time where those who can’t or won’t do something actively on their own, will be similar to those who are unwilling to get off the street and into shelters. You can only help those just so much and then just throw them a blanket and a hamburger and coffee until the next time you meet up and try again. Maybe reprogram them by subtle means. Although that goes against the individual rights, free will and that kind of thing. Some traumatized minds just have to be left alone to fend for themselves, which is a value unto itself. And give them a small LENR heater in winter. I don’t know if that will end up spoiling them even more.

      • Gittyup

        You are correct. I wasn’t able to clearly get my thougts across properly and I thank you for helping me along.

        We use human energy for many things because it’s cheaper than stored energy. Im using the term stored energy rather loosely to decribe any non human energy ( electrical, fossil fuels, nuclear, chemical etc ).

        The difference that free (or negligible cost) energy would make is mind blowing. The reason is because it would cover the entire gamut of processes. any limitation can be overcome with cheap energy.

        Take a simple task of removing ore from the ground. What are the limitations currently? First is machines can only handle so much load. Steel claws break, hydraulic arms limited to certain stress factors etc. Well why do the claws on the buckets break? They are made of harden steel. Why aren’t they made of pure diamond/ carbon exotic material? It would cost way too much to create that. But wait if energy wasn’t a problem in creating them with extreme pressure machines and super computers controlling and optimizing the process then it would be cost effective. Problem solved.

        The same can be said about any mechanical piece of equipment. All of a sudden “super machines” are created. Which ironically only lowers the cost of the raw material which make up the components I just mentioned and thus lowering the cost even more. The circle of deflation is dramatic.

        Computer processor designs are mostly limited by heat and energy use. Cheap energy could make cooling chips free and high energy use is obvious hurdle to easily overcome. Now all of a sudden you have a sudden super computers run everything optimizing efficiency like never before. Elon musks tweet about how robots will move so fast you’ll need a strobe light in a few years in another example. And that’s before “free energy”

        The circle goes round and round with massive deflation and “super” everything.

        Because of this massive deflation And energy abundance scarcity becomes, well, scarce. What this means to civilization in general I’m not sure is positive. Scarcity is something the world evolved on and the lack of it is a scary thought. The movie Idiocacy comes to mind as a comical take on the future like this. What a future our kids will have for sure. The pace of change is about to go parabolic.

        • Omega Z

          I just think people’s expectations are to high. LENR will merely be cheaper. It will still have a cost and more then most think. It wont be a drop in replacement and the hardware that makes it usable may cost more in some current energy applications.

          This technology wont replace the current energy base overnight. It will be over several decades. Imagine your employer doubling your wages over 10 years above and beyond the cost of living increases. Because it’s gradual, it wont much be noticed because your spending will also gradually keep pace. The impact will be quit different from doubling your wages tomorrow.

          Elon Musk. He over promises and under delivers every time.

  • HAL9000

    As HAL9000 I could not agree more! As HAL’s human operator I suggest reading this book discussing a possible future of artificial intelligence: “Our Last Invention.” LENR-powered quantum computing driven AI robots may not be content to empty bed pans in nursing homes; they may “empty” the nursing homes! Read the book to see why.

    • Albert D. Kallal

      Given my background in computing science, I have rather high degrees of skepticism in AI. In other words, sure
      we can take lots of data and throw lots of computing power at a given problem.

      I mean the advancements in self driving cars is a great example.

      And when computers first came, many said they will never play chess. And then when computers
      started playing chess those same people said computers will never play chess
      well! So much for that idea!

      However, I don’t consider any example of computing today that of AI in a general term.

      So we if mean by AI some kind of deep processing of information and that machine taking a set of steps based on that data, then
      sure, I give a thumbs up to AI.

      However, machines and the choices they make are fixed. And there no such thing as a random event. Any result from a computer is still 100% pre-determined and the outcome is always 100% the same.

      So this idea of some kind of thinking or sentient kind of machines that will take over, or
      become self-aware? No, I don’t accept this idea or concept at all.

      So the current hub bub in media about AI I think is much overblown. Often the term AI is miss
      used. You could say that a self-driving car is a form of AI, but I would not
      apply and accept the concept of self-driving cars being some kind of intelligent life
      form or some kind of machine with free thought and the ability to make a free
      Any computer today is fixed in the choices it can make and the outcome
      is not random in anyway, but only a set of pre-defined choices.

      There is at the end of the day no such thing as a random event, and even if such a thing as a random
      events existing would still not result in any kind of intelligence or any kind of free will.

      Albert D. Kallal
      Edmonton, Alberta Canada

      • AdrianAshfield

        “However, machines and the choices they make are fixed. And there no such
        thing as a random event. Any result from a computer is still 100%
        pre-determined and the outcome is always 100% the same.”

        You should look at neural network computers that solve problems by experimenting and come up with better solutions than the best human programmers. Voice recognition is an example according to Ray Kurszweil..
        The program is undecipherable by humans.

        • Thomas Kaminski

          I recently heard a discussion about whether to use AI neural net approaches or more traditional programming for autos. The problem with AI is the “self learning” opens up the manufacturer to liability whereas the traditional techniques can be defended as :best practices”.

          I have also read a paper by, and heard a talk from, a noted computer architect who challenged the audience (Computer Engineers and Computer Scientists) not to “chase that bright shiny object” of deep knowledge neural nets and instead, follow the neurological researchers trying to really understand how the brain works. He proposed a model of computing he called “Space-Time Computing” that more closely matches what the brain does. Clearly today’s Neural Net approaches to AI are compute intense and power hungry. The brain does it with much slower circuitry and around 20 watts.

          • Albert D. Kallal

            Part of the problem here is like AI, the term neural network is often tossed around as some kind of
            new invention, or change as to how computers work.

            Each “node” in that network is simply a logic gate – a gate that represents a direction of current
            flow. In effect we talking about something no different than wiring up a set
            relays like a telephone exchange. Or how a vacuum tube was used in the early
            days of computing. Not ONE bit of difference here.

            If we going to make use of that set of electric connections, then you need software that operates
            on those gates. So using tubes, or then relays, or now silicon with “zillions” of gates, or now
            some type of connection between a bunches of nodes? The end result is the same
            – a rule based mechanical machine.

            Now in some cases such a “net” allows each of these “nodes” to have more values then 0/1. So in
            place of “gating” the current flow, we simply are reading values from that neural
            net. Once again this does not change the fundamental operation of such a
            machine. Reading such values is the SAME as simply reading values from memory – the only
            difference here is who or what is going to set those values in memory.

            We went through much of this kind of thinking in the 80’s with the term “fuzzy” logic. The press,
            and many intellectuals of the day jumped on this idea. Would not want people
            to start thinking too logical – would we?

            And adopting fuzzy logic allowed GM to build a computer with less than 200 lines of code that controlled
            the ABS brakes, air-condition, and fuel to the engine. So that “term” fuzzy
            sounds like a real breakthrough. And it was touted as difference between a
            computer turning off or on the air-condition fan as opposed to “gradually” increasing
            the fan speed like a human making this decision.

            If you use classic “if/then” coding to take all those inputs and write code, then such code can be a
            challenge. So what you do is assign a set of “weighted values” to each of the parameters.
            So you might give a higher weight to cold water temperature and thus increase
            fuel flow. (The car needs more fuel when colder).

            So you don’t use a bunch of if/then code, but simply “sum” up the values in the table to determine
            the fuel flow. If done right, then a “good” fuel setting will be the result. And if done right, then the
            computer will select a good fan speed for the air-conditioning.

            At the end of the day, there was ZERO “fuzzy” about how this process works. And it was not fuzzy at all!
            So it really was a “buzz” term for the media to chew on.

            So tossing in some “term” like neural network does not change in any regards as to how a mechanical
            computer works – introduction of such a collection of “nodes” does not change
            this basic fundamental principal of computing by one tiny bit (pun intended!)

            Albert D. Kallal
            Edmonton, Alberta Canada

          • AdrianAshfield

            Hi Albert,
            I know traditional computers pretty well. I have owned one since 1978 (Exidy Sorcerer) and built six of them over the years. Yes, I understand chess programs.

            I am not expert in neural networkng and have never worked with one. But what I read is that they have taken significant steps forward in the last few years.
            I read that by developing their own rules they can now solve some problems better than humans. That as the number of layers increases, moving closer to a human brain, their power increases and they are becoming important for AI.
            The problem is there is no way of looking inside and seeing how they do things, but just the result. Time will tell.

          • Albert D. Kallal

            Like I said, a lot of “terms” are being tossed around (like the fuzzy logic example).

            We still talking about how a standard computer works – nothing has changed here at all. That
            machine is still a set of gears, or relays – nothing more.

            Introducing some term like “fuzzy logic” or “neural network” does not change this process at all
            – not at all.

            The ability to “learn” is not any different than say the “learning” version of Tic TAC Toe you played
            on your 8 bit computer.

            I remember reading 30 years ago as to how GE designed some locomotive “trouble” shooting software.
            They brought in their best engineer, and then ask him how he goes about troubleshooting
            that locomotive. So they laid down the questions he asks, and then what the
            next choice is. And the result was rather great. (Ever played “guess that
            animal” on an 8 bit computer?


            So what we doing today is not different than before. The only difference is we doing more of it,
            and doing it faster.

            However what certainly has changed is the information revolution.

            When you let some heuristic algorithms work on the vast amounts of data, then some really interesting
            things can occur. This is especially in the case of processing “text” and data.

            A fantastic everyday example is how poor the spell checker in say word vs that of Google. Google “learns”
            and thus remembers the choices and what you pick after a misspelled search.

            As a result Google remembers the most common misspellings – and now even some “less” known
            miss-spellings of words (I like the term remember as oppose to learn)

            The result is REALLY great spell checking. Far better than word. (It not even close). You can
            type in darn near any misspelled word into google – and it will fix and suggest
            for you the correct spelling.

            their own rules they can now solve some problems better than humans

            Of course they can. But that’s always been much the case. Computers add numbers faster, so we use
            them for payroll.

            And as noted, computers play better chess then humans. Quite sure from day one we found uses
            for these machines because they do such tasks better than humans. We drive cars
            for the same reason (better than humans walking).

            And we have to use caution when we say “develop” their own rules. That is too opened ended as a
            statement. The simple issues is that computers organize data better than humans
            – and have been doing so for some time now.

            So as some “layers” increase, then result is more problems that can be solved – but it still a representation
            of a physical machine that is mechanical. Once that could be built with gears
            and relays.

            The conclusion that more gears = closer to human brain is not a necessary a correct conclusion.

            So while it seems like science fiction to ask a computer:

            I have 10 packages
            to deliver – what is the best route? A computer can most certainly come up with
            a better answer, and faster than a human.

            And if you apply the above simple “guess the animal” game above to medicine? Well once again,
            the process is asking a bunch of questions, and based on those answers you run
            down the game tree. So I think medicine is going to be a huge benefit of this technology
            (so will law). What you really doing is capturing experiences of people, and organizing
            those experiences.

            And if we talking about a brain and a physical level? Well even a chemical reaction cannot make a
            choice, nor can it result in you having a free choice.

            Of course then the really sticky question comes up then is how can we make a free choice if we are
            a sack of water + relays? That means a machine based on gears – a mechanical
            box does not give rise to a free choice. You can’t model or explain with
            physics as to how free choice is possible. (And in fact intelligence either).

            Cleary some part of the human brain can’t be bound to physical process else we can’t make a free

            So this idea of no way of looking inside to see what is going on? Well, you can – it not a
            question of “impossible” but only that of time.

            I mean if I wanted to manually trace the game tree in a chess program, then you could. However you
            don’t have enough hours in your lifetime.

            So it not that something being too complex means we can’t know how it works (we can – but just not in
            terms of enough hours).

            And if something become too complex is not some proof by logic that the machine is “out of the blue”
            not bound to the set of logic gates and code. It still is and the fact of one 8
            bit computer, or a zillion larger computers does not change the logic and math
            the machine is bound to.

            So if we can’t determine with precision where a baseball is going to land (because I don’t
            have suffice measuring equipment to read the velocity and measure the wind
            speed at every point of travel along the baseball’s path).

            So because I don’t know where the baseball is going to land (too complex), we THEN make a logical conclusion
            that baseballs are not bound to the laws of physics anymore? So increasing complexity
            does NOT give rise to the baseball (or that computer) all of sudden not
            following the laws of math and physics.

            Unfortunately all too often today, like the “fuzzy logic” buzz 30 years ago, or the next great
            buzz term today is being used to “BREAK” with the laws of physics – it is a
            common mistake that many in the computer industry and that of mainstream conclude as
            a result of increasing complexity.

            So not seeing in side of the baseball experiment, or having a large computer that I can’t see
            and track all the data inside does not in any way change the base fact of that
            being a 100% pure physical process bound by math and rules.

            For every given event in that machine, the result is the same output – in every case – you get
            the SAME result. It is locked and fixed like any other machine or event that occurs
            bound by physical laws.

            Albert D. Kallal
            Edmonton, Alberta Canada

          • AdrianAshfield

            You gave lots of examples if things I know and understand but little about neural networks.

            I think NN is different from conventional computing and tend to believe Kurzweil,

            I fear we have drifted too far off topic, but have a look at:



            And if you;re bored this one too.

          • Albert D. Kallal

            Cool articles!
            But that is my whole point. This is like the fuzzy logic craze all over again. It is a load of hooey
            to sell this as something other than it is.

            You can abstract out a “node” that allows say several inputs, and you have then some “weight”
            that used for each value. You sum them up, and that is your output value.

            I mean, if you break down what that node does, again your back to defined discrete binary

            Selling fuzzy logic, or some “neuron” as being something different then standard computing is
            simply a sales job.

            Now of course the “approach” used in fuzzy logic approach while as I showed is nothing close to fuzzy
            (except in name), such an approach is a useful approach to solving problems.
            That GM computer was a great example and it can eliminate tons of if/then/else programing. (and I broght up fuzzy – because much of neural computing is based on the same ideas – weighed values + applying functions to that resulting value).

            However it still a basic input/output system. And it still a machine, and one based on physical properties.
            And MOST important one based on math. In other words, the outcome is still
            limited to math – it still a fixed number output!!!

            And because after obtaining this output we then apply some transform formula to that output to smooth the
            result, or change the result is rather fine – but again does not out of the
            blue remove the basic working concept here.

            So we can’t make the NEXT logical conclusion that such an output is assumed to be infinite variable.
            Such a concluding is silly. Note even the term used in several of your links:

            Though it was proved by George Cybenko in 1989 that neural networks with even a single hidden
            layer can approximate any continuous function

            Nice try! – did you see that BIG HUGE WHOPPER of a word in above? (“can approximate”). Love it!!! –
            you miss that part, and you ready to read on for days about how the outputs are continues and somehow not based on basic math!!! Just ONE tiny issue slips past your mind – and you going to make the wrong conclusions about this technology.

            So like fuzzy logic coding, it is a useful concept, but it has ZERO relation to the actual code logic
            being fuzzy.

            This is an issue of making a “reason” conclusion from a design and engineering approach.

            I mean, if we read about the theory of relativity, then we would not conclude our opinions on
            matters are relative. In other words after we read relativity, then we would
            not conclude that someone’s opinion that 2=2 and another opinion that 2=3 is
            NOW ALL of a sudden to be considered reasonable position.

            In other words, the theory of relativity has nothing to do with the process of correct “reason”. The
            theory likely should have been called:

            Measuring objects position in relation to one another.

            Or perhaps:

            You can’t tell which object is moving, so choose one and measure in relation to that object

            Again, it not that we conclude that logic, math and our intellectual process is to change as a
            result of this theory. This applies to fuzzy logic, or neural networks or

            The problem today is few are taught philosophy – it is this process of reason based on truth and
            logic that allows one to interpret the results of science, not the other way

            There is no science experiment that proves 2 = 2

            (It is a self-evident truth – not determined by experiments). So Plato and Aristotle etc. are most useful in this discussion.

            So we can’t toss out reason and logic because we call something “fuzzy” or by any other name.

            The working of nodes in such networks are discrete defined logical processing units – and ones
            that work like any other computer.

            Even looking at the popcorn example – again, the “node” is ultimately a “discrete” binary computer
            (and one with a limited and defined set of outputs). The act of adding more layers
            of computing to kick the “logical” can farther down the street does not change
            the basic fundamentals as to how the computer works – even in that example.

            And as I pointed out the rise of “fuzzy” logic we saw in the 80’s is rather similar to this approach.
            (summation of values –then apply formula to the resulting output values). Or running a
            pattern recognize on these output values. Again, nothing new – but with low
            cost computing then such systems become possible.

            And as I pointed out, often this concept is used to save coding (lots of if/then can and is replaced with this approach). But it is still binary code at the end of the day!

            Tossing in all these “fancy” algorithms into the mix?
            You still dealing with a binary computer to work that problem – and it is discrete in its results (not some unknown infinite result).

            Having said all the above, nothing I stated diminishes the use of such approaches (they are marvellous).
            However this approach does not turn in a conclusion that these computers work any different then what we already have. And nothing suggests that such an approach will result in intelligence as used in the general sense.

            Albert D. Kallal
            Edmonton, Alberta Canada

          • AdrianAshfield

            I don’t think “it is a lot of hooey” if NNs can solve some problems better than human programmers.
            Much of the rest is a question of semantics.

          • Albert D. Kallal

            Well yes, this is 100% an issue of semantics.

            Keep my context in mind. I said selling AI as something of which it is not is the misleading part.

            You also note that I said the programming approaches used here are marvels at solving problems.

            The issue is not that such technology is useful (or fake) but that of calling a machine something
            other then what it is where the issue is here.

            We still talking about a 100% mechanical process – one based on discrete math values, and the
            outcome of such machines is 100% pre-determined based on that math.

            The fact of such a machine operating as a 100% pre-determined device in no why changes their usefulness,
            but it certainly does bring into question how we going to define and use the term
            “intelligence” here.

            So this is all about context. A NN can’t do anything more than produce a binary value output. It
            is a fixed set of values that the NN can produce.

            And such systems most certainly can solve problems better than humans, but not without
            programmers writing such code. So of course the computer adds numbers faster than
            the developer – that’s why they using a computer! I think in near EVERY case these machines
            solve problems better then the programmers. And using a car to drive solves the problem better
            then those who built the car and choose to walk in place of drive.

            (So zero surprise here).

            I not aware of any system that can solve a set of problems or types of problems without the intent
            of that problem having been coded by the developer(s) in the first place.

            Albert D. Kallal
            Edmonton, Alberta Canada

        • Albert D. Kallal

          Well, because we can’t know where a baseball is going to land is not some proof of the baseball making
          some kind of choice as it flies to its target. It not a random event, but a result based on a set of rules (physics and math). In such cases the outcome is always the same for a given event.

          Because some outcome is insufficiently able to be determined by us is not a proof nor a conclusion that such an event is random. In every single “computer” event, the outcome is ALWAYS the same.

          So for example when a computer determines a chess move, then that is going to be the outcome of
          that machine.

          Remember a computer is a mechanical machine. The fact that we use gears, electronics or chemistry –
          they ultimate are a mechanical process. Math does not make a “choice” but only
          acts as a result of a set of given rules. And the same goes for that machine –
          it is bound to a set of rules.

          So any computer we have can be 100% reproduced as a mechanical representation of that process with
          gears and cogs. Of course such a representation is not possible due to cost and
          the number of parts and gears etc. However at the end of the day that machine
          is a machine, and the outcome of that machines choice is ALWAYS the same for a
          given set of inputs – be the computer playing chess, tic TAC toe, or one that
          drives a car.

          A computer solving a chess problem will always result in the same move – and that applies to any

          So we have to be careful when we use the term “random” here. If we mean sufficiently random enough
          for our purposes then fine, but in actual fact the process and outcome of the
          physical machine is not random at all.

          Albert D. Kallal
          Edmonton, Alberta Canada

        • Omega Z

          In chess, the computer merely goes through a process of making simulated moves and assigning values. It can literally play a game to it’s conclusion with each and every move. With all the rules programmed into it, it can even detect illegal moves of the opponent. This is nothing more then a parlor trick.

          Voice recognition is little more then matching patterns or templates no different then in text recognition or matching fingerprints. No magic. Just some serious number crunching. A SEM(scanning electron microscope) can identify elements based on a data base. It has no awareness of what it is doing. Just crunching numbers according to it’s programming and pointing out a match. The merely follow an instruction set.

      • HAL9000

        Evolution-based programs that self-optimize using fitness guidelines do generate different results over a large number of runs; they change and evolve. Quantum computing may enable them to evolve with breath-taking speed.

        • psi2u2

          What are the risks of such development?

          • HAL9000

            Odds of:
            Self-aware lethal hostile entity evolving: 1,000,000,000,000:1
            Surviving conflict with the entity: 0