• Things I never thought would appear

    From Joy Beeson@21:1/5 to All on Thu Oct 10 18:55:14 2024
    A computer farm that occupies an entire building and has its own
    nuclear generator.

    I knew the mini-skirt would come back -- styles always come back --
    but this time around it's old, wrinkled-up women who are the early
    adopters.

    I thought that we were supposed to be a conservative bunch; old women
    were the last to give up pantsuits when they went out of style.

    --
    Joy Beeson
    joy beeson at centurylink dot net
    http://wlweather.net/PAGEJOY/

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Dorothy J Heydt@21:1/5 to jbeeson@invalid.net.invalid on Fri Oct 11 01:07:56 2024
    In article <lomggjdgr9uo147ijjh6kr3bhikg6hgfis@4ax.com>,
    Joy Beeson <jbeeson@invalid.net.invalid> wrote:

    A computer farm that occupies an entire building and has its own
    nuclear generator.

    [Hal Heydt]
    At the other end of the scale is a computer with a circuit board
    that meets the ISO size spec for a credit card. Look up the
    latest one, Raspberry Pi Pi5.

    I knew the mini-skirt would come back -- styles always come back --
    but this time around it's old, wrinkled-up women who are the early
    adopters.

    I thought that we were supposed to be a conservative bunch; old women
    were the last to give up pantsuits when they went out of style.

    I was under the impression that the revolt over the midi-skirt
    killed fashion design over hem length once and for all. (The
    grounds being the NO ONE looked good in a midi.)

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Keith F. Lynch@21:1/5 to Joy Beeson on Fri Oct 11 02:31:32 2024
    Joy Beeson <jbeeson@invalid.net.invalid> wrote:
    A computer farm that occupies an entire building and has its own
    nuclear generator.

    I've never heard of a computer with its own nuclear generator (not
    counting space probes powered by RTGs). But there's nothing new about computers filling a building and consuming vast amounts of electricity.

    What is new is that computers are enormously more powerful. One of the
    best illustrations of that is this list of years:

    1400 1706 1949 1958 1961 1973 1983 1987 1989 1997 1999 2002 2011 2022

    The Nth term is when pi was first calculated to 10^N places. The
    third term is 1949 because that's when pi was calculated to 1000
    decimal places (using a desktop electromechanical calculator). The
    sixth term is 1973 because that's when a computer calculated in to a
    million places. A billion places in 1989, a trillion in 2002, and
    currently the record is 202 trillion. Print that many digits in
    books, they would fill a very large library.

    I knew the mini-skirt would come back -- styles always come back --
    but this time around it's old, wrinkled-up women who are the early
    adopters.

    Styles are associated with the generations who were young when they
    were popular. As such, people tend to look old in old movies, even
    if they were young at the time.
    --
    Keith F. Lynch - http://keithlynch.net/
    Please see http://keithlynch.net/email.html before emailing me.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Jay E. Morris@21:1/5 to Keith F. Lynch on Thu Oct 10 21:42:15 2024
    On 10/10/2024 9:31 PM, Keith F. Lynch wrote:
    Joy Beeson<jbeeson@invalid.net.invalid> wrote:
    A computer farm that occupies an entire building and has its own
    nuclear generator.
    I've never heard of a computer with its own nuclear generator (not
    counting space probes powered by RTGs). But there's nothing new about computers filling a building and consuming vast amounts of electricity.

    Microsoft wants Three Mile Island to fuel its AI power needs / Microsoft
    has signed a 20-year deal to exclusively access 835 megawatts of energy
    from a nuclear plant.

    https://www.theverge.com/2024/9/20/24249770/microsoft-three-mile-island-nuclear-power-plant-deal-ai-data-centers

    Oracle considers nuclear energy to power its AI
    The tech company races to harness nuclear power for its data centers, as
    AI's insatiable energy demands surge; while new technologies emerge to
    meet this need, nuclear energy's risks and rewards create an unsettling
    modern tale

    https://www.ynetnews.com/business/article/rjkyhtepc

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Keith F. Lynch@21:1/5 to Jay E. Morris on Fri Oct 11 04:27:01 2024
    Jay E. Morris <morrisj@epsilon3.comcon> wrote:
    Microsoft wants Three Mile Island to fuel its AI power needs

    Yes, I read that. And I read that they want taxpayers to pay for it,
    which I am of course totally against. Microsoft should pay its own
    way. If it can't afford to do so, it should go bankrupt.

    Is anyone reseaching how to make computers more energy-efficient?
    I recall Feynman writing on that topic.

    I know that computers are a long way from their theoretical maximum
    possible efficiency.

    Am I the only one here who is annoyed by the use of "AI" to mean
    anything other than human-level (or better) machine intelligence?
    --
    Keith F. Lynch - http://keithlynch.net/
    Please see http://keithlynch.net/email.html before emailing me.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Gary McGath@21:1/5 to Keith F. Lynch on Fri Oct 11 06:55:15 2024
    On 10/11/24 12:27 AM, Keith F. Lynch wrote:


    Am I the only one here who is annoyed by the use of "AI" to mean
    anything other than human-level (or better) machine intelligence?

    My definition of "artificial intelligence" is something a computer does
    which a lot of people think only humans should be able to do. This
    criterion shifts over time.

    I don't think "human-level machine intelligence" is meaningful, because computation to carry out tasks and cognition to further the existence of
    a living organism aren't commensurable. The measure of human
    intelligence has little to do with, for example, how many digits of pi
    we can calculate in an hour.

    --
    Gary McGath http://www.mcgath.com

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Keith F. Lynch@21:1/5 to Gary McGath on Fri Oct 11 11:48:07 2024
    Gary McGath <garym@mcgath.com> wrote:
    I don't think "human-level machine intelligence" is meaningful,
    because computation to carry out tasks and cognition to further the
    existence of a living organism aren't commensurable.

    I'm not convinced of that. Nobody has come up with a task that a
    person could do but that no computer could ever do. There are tasks
    that it's been proven computers can't do, but no person has proven
    able to do them either.

    So I think it's not at all unlikely that computers, some of them in
    humanoid robot bodies, will someday be able to do every task that
    a person can do, better and less expensively. At that point the
    unemployment rate will increase to 100%. Those who don't own stock
    in the AI companies will have financial problems.

    Or, of course, the intelligent machines may decide we're a nuisance,
    and wipe us out.

    A third category is uploaded people, i.e. human consciousnesses copied
    into a computer. Simply completely map a human brain and then emulate
    it in software. It would of course have the same memories and
    personality as the original person. By cranking up the clock
    frequency, it could work must faster than us flesh people, get a
    full night's sleep in a few seconds, or a lengthy vacation in a few
    minutes. And it could work for much less income, since it could
    enjoy realistic virtual entertainments. It would be potentially
    immortal. Or at least last as long as our civilization.

    But present-day computers and programs are nowhere close to
    human-level intelligence yet.

    I understand there is a complete brain map of a fruit fly. I haven't
    heard whether they've emulated it in software and gotten fly-like
    behavior out of it.
    --
    Keith F. Lynch - http://keithlynch.net/
    Please see http://keithlynch.net/email.html before emailing me.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Scott Dorsey@21:1/5 to jbeeson@invalid.net.invalid on Fri Oct 11 13:17:55 2024
    Joy Beeson <jbeeson@invalid.net.invalid> wrote:
    A computer farm that occupies an entire building and has its own
    nuclear generator.

    That sounds like Multivac to me. Except that Multivac always made
    correct and wise decisions.
    --scott

    --
    "C'est un Nagra. C'est suisse, et tres, tres precis."

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Gary McGath@21:1/5 to Keith F. Lynch on Fri Oct 11 19:01:25 2024
    On 10/11/24 7:48 AM, Keith F. Lynch wrote:
    Gary McGath <garym@mcgath.com> wrote:
    I don't think "human-level machine intelligence" is meaningful,
    because computation to carry out tasks and cognition to further the
    existence of a living organism aren't commensurable.

    I'm not convinced of that. Nobody has come up with a task that a
    person could do but that no computer could ever do. There are tasks
    that it's been proven computers can't do, but no person has proven
    able to do them either.

    You left out the part of my post where I said that carrying out tasks
    isn't the point. Human (and animal) intelligence is a faculty for
    maintaining and enhancing the life of which it is a part. We might be
    able to create machines whose prime directive is to survive, reproduce,
    and maximize their satisfaction (though I don't know what that would
    mean in a machine designed and created by humans), but it would be a bad
    idea.

    So I think it's not at all unlikely that computers, some of them in
    humanoid robot bodies, will someday be able to do every task that
    a person can do, better and less expensively. At that point the
    unemployment rate will increase to 100%. Those who don't own stock
    in the AI companies will have financial problems.

    You're overlooking the principle of comparative advantage. People in
    such a world wouldn't sit around and wait for the machines to feed them.
    They'd do the things at which they're relatively best, while machines
    would do the tasks which they're relatively best at.

    Or, of course, the intelligent machines may decide we're a nuisance,
    and wipe us out.

    A third category is uploaded people, i.e. human consciousnesses copied
    into a computer. Simply completely map a human brain and then emulate
    it in software. It would of course have the same memories and
    personality as the original person. By cranking up the clock
    frequency, it could work must faster than us flesh people, get a
    full night's sleep in a few seconds, or a lengthy vacation in a few
    minutes. And it could work for much less income, since it could
    enjoy realistic virtual entertainments. It would be potentially
    immortal. Or at least last as long as our civilization.

    The key word there is "emulate." They wouldn't be people. At a minimum,
    they'd need to have human-equivalent bodies to keep the same
    personalities; otherwise they'd have different needs and different ways
    of interacting with the world, and so would diverge from human attitudes.

    --
    Gary McGath http://www.mcgath.com

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Charles Packer@21:1/5 to Keith F. Lynch on Sat Oct 12 07:52:18 2024
    On Fri, 11 Oct 2024 11:48:07 -0000 (UTC), Keith F. Lynch wrote:

    Gary McGath <garym@mcgath.com> wrote:
    I don't think "human-level machine intelligence" is meaningful, because
    computation to carry out tasks and cognition to further the existence
    of a living organism aren't commensurable.

    I'm not convinced of that. Nobody has come up with a task that a person could do but that no computer could ever do. There are tasks that it's
    been proven computers can't do, but no person has proven able to do them either.

    So I think it's not at all unlikely that computers, some of them in
    humanoid robot bodies, will someday be able to do every task that a
    person can do, better and less expensively. At that point the
    unemployment rate will increase to 100%. Those who don't own stock in
    the AI companies will have financial problems.

    Or, of course, the intelligent machines may decide we're a nuisance, and
    wipe us out.

    A third category is uploaded people, i.e. human consciousnesses copied
    into a computer. Simply completely map a human brain and then emulate
    it in software. It would of course have the same memories and
    personality as the original person. By cranking up the clock frequency,
    it could work must faster than us flesh people, get a full night's sleep
    in a few seconds, or a lengthy vacation in a few minutes. And it could
    work for much less income, since it could enjoy realistic virtual entertainments. It would be potentially immortal. Or at least last as
    long as our civilization.

    But present-day computers and programs are nowhere close to human-level intelligence yet.

    I understand there is a complete brain map of a fruit fly. I haven't
    heard whether they've emulated it in software and gotten fly-like
    behavior out of it.

    They have. One of the series of papers published in Nature says that
    they presented the model with simulated sugar and got the fly to stick out
    its simulated tongue.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Keith F. Lynch@21:1/5 to Charles Packer on Sat Oct 12 12:43:30 2024
    Charles Packer <mailbox@cpacker.org> wrote:
    Keith F. Lynch wrote:
    I understand there is a complete brain map of a fruit fly. I
    haven't heard whether they've emulated it in software and gotten
    fly-like behavior out of it.

    They have. One of the series of papers published in Nature says
    that they presented the model with simulated sugar and got the fly
    to stick out its simulated tongue.

    Cool! We're only a few short steps away from being able to present
    the model of a Usenet poster's brain with a simulated Usenet post and
    get a post in a reply.
    --
    Keith F. Lynch - http://keithlynch.net/
    Please see http://keithlynch.net/email.html before emailing me.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Keith F. Lynch@21:1/5 to Evelyn C. Leeper on Sat Oct 12 12:56:46 2024
    Evelyn C. Leeper <evelynchimelisleeper@gmail.com> wrote:
    Keith F. Lynch wrote:
    Nobody has come up with a task that a person could do but that no
    computer could ever do.

    Make a baby?

    I'm not convinced that computers couldn't do that.

    To reach other solar systems may require trips of tens of thousands of
    years. Robot space probes could be programmed to grow human colonists
    if and when they get to a habitable planet.

    And even if computers couldn't do that, my point is that it's not
    something someone is typically willing to pay someone to do.
    --
    Keith F. Lynch - http://keithlynch.net/
    Please see http://keithlynch.net/email.html before emailing me.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Keith F. Lynch@21:1/5 to Gary McGath on Sat Oct 12 12:34:06 2024
    Gary McGath <garym@mcgath.com> wrote:
    You left out the part of my post where I said that carrying out
    tasks isn't the point. Human (and animal) intelligence is a faculty
    for maintaining and enhancing the life of which it is a part.
    We might be able to create machines whose prime directive is to
    survive, reproduce, and maximize their satisfaction (though I don't
    know what that would mean in a machine designed and created by
    humans), but it would be a bad idea.

    What would it be like to be an AI is a different question from what
    would it be like to share the world with AIs.

    Also, not everyone chooses to use their intelligence to maintain or
    enhance their life.

    You're overlooking the principle of comparative advantage. People
    in such a world wouldn't sit around and wait for the machines to
    feed them. They'd do the things at which they're relatively best,
    while machines would do the tasks which they're relatively best at.

    Comparative advantage means it makes sense for a doctor to hire a
    receptionist even if he'd be better receptionist than the person he
    hires. But that's only because his high income suffices to pay for
    the receptionist.

    Today, nobody wonders whether it would pay better to compete with a hydroelectric dam by turning a hand-cranked generator or to compete
    with a computer by doing arithmetic by hand. Obviously neither one
    would give anything close to a living wage.

    I'm suggesting that, given true AI, people would be hopelessly
    outcompeted by AIs in literally *every* field. Ten years after that
    doctor saves money by replacing his human receptionist with a robot receptionist, his patients save money by replacing him with a robot
    doctor.

    The key word there is "emulate." They wouldn't be people. At a
    minimum, they'd need to have human-equivalent bodies to keep the
    same personalities; otherwise they'd have different needs and
    different ways of interacting with the world, and so would diverge
    from human attitudes.

    To a degree, that has already happened. The personality of a person
    with a car differs from that of a person without one. The personality
    of a person with a cell phone differs from that of a person without
    one. The personality of a person with a disability differs from that
    of a person without one. But they're still people.

    There are already people with artificial hearts and artificial
    kidneys. To the extent that those work as well as the original, their
    life should be unchanged. If those work better than the original,
    their life should be improved. The same should be true of artificial
    bodies or artificial brains.
    --
    Keith F. Lynch - http://keithlynch.net/
    Please see http://keithlynch.net/email.html before emailing me.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Gary McGath@21:1/5 to Keith F. Lynch on Sun Oct 13 19:50:26 2024
    On 10/12/24 8:34 AM, Keith F. Lynch wrote:
    Gary McGath <garym@mcgath.com> wrote:
    You left out the part of my post where I said that carrying out
    tasks isn't the point. Human (and animal) intelligence is a faculty
    for maintaining and enhancing the life of which it is a part.
    We might be able to create machines whose prime directive is to
    survive, reproduce, and maximize their satisfaction (though I don't
    know what that would mean in a machine designed and created by
    humans), but it would be a bad idea.

    What would it be like to be an AI is a different question from what
    would it be like to share the world with AIs.

    Also, not everyone chooses to use their intelligence to maintain or
    enhance their life.

    Some do it badly, but if they didn't do it at all, they wouldn't live long.

    You're overlooking the principle of comparative advantage. People
    in such a world wouldn't sit around and wait for the machines to
    feed them. They'd do the things at which they're relatively best,
    while machines would do the tasks which they're relatively best at.

    Comparative advantage means it makes sense for a doctor to hire a receptionist even if he'd be better receptionist than the person he
    hires. But that's only because his high income suffices to pay for
    the receptionist.

    Today, nobody wonders whether it would pay better to compete with a hydroelectric dam by turning a hand-cranked generator or to compete
    with a computer by doing arithmetic by hand. Obviously neither one
    would give anything close to a living wage.

    I'm suggesting that, given true AI, people would be hopelessly
    outcompeted by AIs in literally *every* field. Ten years after that
    doctor saves money by replacing his human receptionist with a robot receptionist, his patients save money by replacing him with a robot
    doctor.

    What money? People have to earn it somehow. Comparative advantage still applies, or else people simply wouldn't be part of the economy and hence couldn't pay for the services of machines.

    Unless, perhaps, the humans became the pets of the machines, maintained
    because the machines are programmed to. This would be like the final
    page of "With Folded Hands."


    The key word there is "emulate." They wouldn't be people. At a
    minimum, they'd need to have human-equivalent bodies to keep the
    same personalities; otherwise they'd have different needs and
    different ways of interacting with the world, and so would diverge
    from human attitudes.

    To a degree, that has already happened. The personality of a person
    with a car differs from that of a person without one. The personality
    of a person with a cell phone differs from that of a person without
    one. The personality of a person with a disability differs from that
    of a person without one. But they're still people.

    This demonstrates a different point from the one you intended. A
    computer with a human voice generator differs from a computer without
    one, but it's still a computer.

    There are already people with artificial hearts and artificial
    kidneys. To the extent that those work as well as the original, their
    life should be unchanged. If those work better than the original,
    their life should be improved. The same should be true of artificial
    bodies or artificial brains.

    An artificial brain, if we're talking about replacing rather than
    supplementing the original, is different in kind from an artificial
    heart. Consciousness resides in the brain. Stephen Hawking was still
    Stephen Hawking even when almost nothing else in his body functioned
    without the aid of machines. Once his brain stopped, he ceased to be
    Stephen Hawking. A computer that emulated him, no matter how convincing
    it was, wouldn't resurrect him as far as he was concerned.

    --
    Gary McGath http://www.mcgath.com

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Mike Van Pelt@21:1/5 to Keith F. Lynch on Mon Oct 14 04:58:26 2024
    In article <vea9el$hvu$1@reader1.panix.com>,
    Keith F. Lynch <kfl@KeithLynch.net> wrote:
    Am I the only one here who is annoyed by the use of "AI" to mean
    anything other than human-level (or better) machine intelligence?

    (raises hand)

    I'd settle for 70-ish IQ intelligence.

    What the marketroids are pleased to call "Artificial
    Intelligence" is not intelligence of any kind, just pattern
    matching. (With amusing results where it breaks down
    and includes nonexistent scientific references or legal
    citations.) I've been lobbying for everyone to call it
    "Simulated Intelligence", but not gotten much traction.
    --
    Mike Van Pelt | "I don't advise it unless you're nuts."
    mvp at calweb.com | -- Ray Wilkinson, after riding out Hurricane
    KE6BVH | Ike on Surfside Beach in Galveston

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Dorothy J Heydt@21:1/5 to usenet@mikevanpelt.com on Mon Oct 14 16:21:38 2024
    In article <vei8di$12ono$1@dont-email.me>,
    Mike Van Pelt <usenet@mikevanpelt.com> wrote:
    In article <vea9el$hvu$1@reader1.panix.com>,
    Keith F. Lynch <kfl@KeithLynch.net> wrote:
    Am I the only one here who is annoyed by the use of "AI" to mean
    anything other than human-level (or better) machine intelligence?

    (raises hand)

    I'd settle for 70-ish IQ intelligence.

    What the marketroids are pleased to call "Artificial
    Intelligence" is not intelligence of any kind, just pattern
    matching. (With amusing results where it breaks down
    and includes nonexistent scientific references or legal
    citations.) I've been lobbying for everyone to call it
    "Simulated Intelligence", but not gotten much traction.

    [Hal Heydt]
    Dorothy's--and my--usual snark about anything claiming the be
    "AI" is that it was mostly "A" and very little "I".

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Keith F. Lynch@21:1/5 to Gary McGath on Tue Oct 15 03:07:40 2024
    Gary McGath <garym@mcgath.com> wrote:
    Keith F. Lynch wrote:
    I'm suggesting that, given true AI, people would be hopelessly
    outcompeted by AIs in literally *every* field. Ten years after
    that doctor saves money by replacing his human receptionist with a
    robot receptionist, his patients save money by replacing him with a
    robot doctor.

    What money? People have to earn it somehow. Comparative advantage
    still applies, or else people simply wouldn't be part of the economy
    and hence couldn't pay for the services of machines.

    Those without property or savings wouldn't be part of the economy and
    hence couldn't pay for the services of machines, since they couldn't
    gain money by selling their labor, since all human labor would have
    become worthless. Unfortunately, plenty of people are already in this position, as their labor is worthless due to their lack of skills.
    There's much less demand for unskilled labor than there used to be.

    It's claimed that the US unemployment rate is very low, but I suspect
    the government simply defines most unemployed people as not being part
    of the labor force. There seem to be more and more homeless people
    every year. The Washington Post keeps asking why so many people
    believe the economy is in bad shape when the government informs us
    that it's in great shape. My conjecture is that the government is
    lying to us.

    In the AI scenario, like today, plenty of people will own land, stock,
    or other property, and gain income from rents or dividends, not from
    their labor.

    Unless, perhaps, the humans became the pets of the machines,
    maintained because the machines are programmed to.

    Yes, that's a possibility. Presumably they'll spay and neuter us,
    so we won't use resources needed for the AI computations.

    Another possibility, even with today's technology, is a universal
    basic income, i.e. a negative income tax. Improvements in
    productivity imply that it ought to be possible. Candidates talk
    about extending Medicare to people of all ages. Why not do the same
    with Social Security?

    But instead we see that, despite high taxes, the US government has
    somehow accumulated the largest debt in world history, and it's
    increasing at an ever-increasing rate, and neither major candidate
    has any plan to change that.

    So perhaps the productivity improvements are either illusory or
    are going to just a few powerful people. As I've mentioned, my
    grandfather bought a large house free and clear while supporting
    five dependents and earning a three-digit annual salary. This was
    in the 1930s, when the economy was supposedly in much worse shape.

    An artificial brain, if we're talking about replacing rather than supplementing the original, is different in kind from an artificial
    heart. Consciousness resides in the brain.

    There's no sharp distinction between replacing and supplementing.
    Suppose your neurons were dying one by one, so surgeons were replacing
    each one, as it dies, by an electronic circuit which behaved exactly
    the same. Eventually, all your neurons will have been replaced by
    electronics, but your behavior would be unchanged. Your behavior
    includes your honest answers as to whether you felt you were still
    the same person.

    We already supplement. A forgetful person can write things down. I
    believe much of one's intelligence is outside one's brain. It resides
    in one's papers, surroundings, and friends. As I learned the hard way
    47 years ago, if I'm suddenly in a hostile environment, surrounded by
    liars and cut off from my records, family, and friends, I am much
    diminished, and quickly come to doubt everything.
    --
    Keith F. Lynch - http://keithlynch.net/
    Please see http://keithlynch.net/email.html before emailing me.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tim Merrigan@21:1/5 to Keith F. Lynch on Tue Oct 15 05:24:52 2024
    On 10/10/2024 7:31 PM, Keith F. Lynch wrote:
    I've never heard of a computer with its own nuclear generator

    Elon Musk (I think it was) recently got the remaining plant at Three
    Mile Island reactivated to power an AI server farm.

    --

    Qualified immunity = virtual impunity.

    Tim Merrigan

    --
    This email has been checked for viruses by AVG antivirus software.
    www.avg.com

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tim Merrigan@21:1/5 to Evelyn C. Leeper on Tue Oct 15 05:31:05 2024
    On 10/11/2024 11:11 AM, Evelyn C. Leeper wrote:
    On 10/11/24 7:48 AM, Keith F. Lynch wrote:
    Gary McGath <garym@mcgath.com> wrote:
    I don't think "human-level machine intelligence" is meaningful,
    because computation to carry out tasks and cognition to further the
    existence of a living organism aren't commensurable.

    I'm not convinced of that.  Nobody has come up with a task that a
    person could do but that no computer could ever do.  There are tasks
    that it's been proven computers can't do, but no person has proven
    able to do them either.

    Make a baby?


    A robot computer manufacturing plant? Or even just computer chip plant?

    --

    Qualified immunity = virtual impunity.

    Tim Merrigan

    --
    This email has been checked for viruses by AVG antivirus software.
    www.avg.com

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tim Merrigan@21:1/5 to Tim Merrigan on Tue Oct 15 05:46:45 2024
    On 10/15/2024 5:31 AM, Tim Merrigan wrote:
    On 10/11/2024 11:11 AM, Evelyn C. Leeper wrote:
    On 10/11/24 7:48 AM, Keith F. Lynch wrote:
    Gary McGath <garym@mcgath.com> wrote:
    I don't think "human-level machine intelligence" is meaningful,
    because computation to carry out tasks and cognition to further the
    existence of a living organism aren't commensurable.

    I'm not convinced of that.  Nobody has come up with a task that a
    person could do but that no computer could ever do.  There are tasks
    that it's been proven computers can't do, but no person has proven
    able to do them either.

    Make a baby?


    A robot computer manufacturing plant?  Or even just computer chip plant?


    Some of the other respondents seemed to think you meant making baby
    humans, I interpreted it as making baby computers.

    --

    Qualified immunity = virtual impunity.

    Tim Merrigan

    --
    This email has been checked for viruses by AVG antivirus software.
    www.avg.com

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Tim Merrigan@21:1/5 to Keith F. Lynch on Tue Oct 15 05:33:20 2024
    On 10/12/2024 5:56 AM, Keith F. Lynch wrote:
    Evelyn C. Leeper <evelynchimelisleeper@gmail.com> wrote:
    Keith F. Lynch wrote:
    Nobody has come up with a task that a person could do but that no
    computer could ever do.

    Make a baby?

    I'm not convinced that computers couldn't do that.

    To reach other solar systems may require trips of tens of thousands of
    years. Robot space probes could be programmed to grow human colonists
    if and when they get to a habitable planet.

    And even if computers couldn't do that, my point is that it's not
    something someone is typically willing to pay someone to do.

    Surrogates?

    --

    Qualified immunity = virtual impunity.

    Tim Merrigan

    --
    This email has been checked for viruses by AVG antivirus software.
    www.avg.com

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Scott Dorsey@21:1/5 to Keith F. Lynch on Tue Oct 15 23:02:27 2024
    Keith F. Lynch <kfl@KeithLynch.net> wrote:
    It's claimed that the US unemployment rate is very low, but I suspect
    the government simply defines most unemployed people as not being part
    of the labor force. There seem to be more and more homeless people
    every year. The Washington Post keeps asking why so many people
    believe the economy is in bad shape when the government informs us
    that it's in great shape. My conjecture is that the government is
    lying to us.

    I hate to tell you this but a lot of the homeless people in the DC area actually have jobs.... just not jobs that pay enough for them to live
    off the streets. If anything, that is more damning.
    --scott

    --
    "C'est un Nagra. C'est suisse, et tres, tres precis."

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Scott Dorsey@21:1/5 to tppm@rr.ca.com on Tue Oct 15 23:03:45 2024
    Tim Merrigan <tppm@rr.ca.com> wrote:
    Some of the other respondents seemed to think you meant making baby
    humans, I interpreted it as making baby computers.

    And when two computers that love one another get together... that is how
    we get the 8051.
    --scott
    --
    "C'est un Nagra. C'est suisse, et tres, tres precis."

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Mike Van Pelt@21:1/5 to Keith F. Lynch on Tue Oct 15 23:47:48 2024
    In article <vekm9s$on2$1@reader1.panix.com>,
    Keith F. Lynch <kfl@KeithLynch.net> wrote:
    It's claimed that the US unemployment rate is very low, but I suspect
    the government simply defines most unemployed people as not being part
    of the labor force.

    This is true. If someone is not on record as actually seeking
    employment, they are "not part of the labor force", and are
    not counted towards the unemployment rate.


    --
    Mike Van Pelt | "I don't advise it unless you're nuts."
    mvp at calweb.com | -- Ray Wilkinson, after riding out Hurricane
    KE6BVH | Ike on Surfside Beach in Galveston

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Keith F. Lynch@21:1/5 to Scott Dorsey on Tue Oct 15 23:19:28 2024
    Scott Dorsey <kludge@panix.com> wrote:
    I hate to tell you this but a lot of the homeless people in the DC
    area actually have jobs.... just not jobs that pay enough for them
    to live off the streets. If anything, that is more damning.

    Yes, I believe that inflation is much higher than is claimed, and has
    been for a long time. For instance the new Yankee Stadium costs one
    thousand times more than the old one -- but seats fewer spectators.

    So the unemployment rate isn't the only thing the US government
    lies about.
    --
    Keith F. Lynch - http://keithlynch.net/
    Please see http://keithlynch.net/email.html before emailing me.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Jeff Jonas@21:1/5 to All on Sat Nov 9 08:14:41 2024
    Microsoft wants Three Mile Island to fuel its AI power needs

    Is anyone reseaching how to make computers more energy-efficient?

    Not anymore.
    "carbon neutral" advertising/green-washing is passe'.
    The only reason for energy efficiency is longer battery life
    but with high energy batteries, that's low on the priority list now.

    I recall Feynman writing on that topic.

    kinda:

    https://en.wikipedia.org/wiki/There%27s_Plenty_of_Room_at_the_Bottom "There's Plenty of Room at the Bottom: An Invitation to Enter a New Field of Physics"
    was a lecture given by physicist Richard Feynman at the annual American Physical Society meeting at Caltech on December 29, 1959.
    Feynman considered the possibility of direct manipulation of individual atoms as a more robust form of synthetic chemistry than those used at the time. Versions of the talk were reprinted in a few popular magazines, but it went largely unnoticed until the 1980s.


    https://www.zyvex.com/nanotech/feynman.html
    https://www2.cs.sfu.ca/~vaughan/teaching/415/papers/feynman-plenty.html There's Plenty of Room at the Bottom
    An Invitation to Enter a New Field of Physics
    by Richard P. Feynman


    https://www.americanscientist.org/article/computers-that-can-run-backwards Physicist Richard Feynman showed that it is theoretically possible to create an adiabatic reversible computer.
    Feynman took an interest in reversible computing in the 1970s
    because he wanted to know whether there is a fundamental lower limit to how much energy is needed to carry out a computation.
    What kinds of computers attain that limit?
    He knew of the Landauer limit, which puts a lower bound on the amount of energy lost to heat
    when a single bit of information is lost.
    He asked: If we use reversible circuits, which lose no bits,
    what is the minimum energy required to get the computation done?
    He eventually concluded that there is no theoretical minimum.

    --

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)