[ExI] ET Emergence (Was Re: Uploads as a group of AI agents)

Tue Mar 31 23:11:47 UTC 2026

On Tue, Mar 31, 2026, 5:14 PM John Clark <johnkclark at gmail.com> wrote:

> On Tue, Mar 31, 2026 at 4:19 PM Jason Resch via extropy-chat <
> extropy-chat at lists.extropy.org> wrote:
>
>
>> *>  It's true that with a reversible computer you could theoretically
>>>>> complete any calculation using an arbitrarily small amount of energy,
>>>>> however the smaller your energy usage is the slower your calculation is,
>>>>> and as your energy usage approaches zero the time to complete your
>>>>> calculation approaches infinity.  *
>>>>>
>>>>
>>>> *>What's your source for this?*
>>>>
>>>
>>> *Reversibility and Adiabatic Computation: Trading Time and Space for
>>> Energy* <https://arxiv.org/pdf/quant-ph/9703022>
>>>
>>
>> *> This seems to confirm what I said. From the paper:*
>> *"Considerations of thermodynamics of computing started in the early *
>> *fifties. J. von Neumann reputedly thought that a computer operating at *
>> *temperature T must dissipate at least kT ln 2 Joule per elementary bit
>> op-eration"*
>>
>
> *That's how much energy you need to erase one bit of information and,
> unless you have infinite memory or are using reversible computation, you're
> going to need to erase information. *
>
> *I just had this conversation with Gemini:*
> *==*
>
> *Me: Is it true that with a reversible computer you could theoretically
> perform a calculation with an arbitrarily small amount of energy?*
>
> Gemini: It is theoretically true. In classical computing, we are used to
> the idea that running a processor generates heat, but physics suggests that
> this heat isn't necessarily a byproduct of the *calculation*
> itself—rather, it's a byproduct of erasing information.
>
> The Landauer Principle
>
> The fundamental link between information and energy is defined by *Landauer's
> Principle*. It states that any l*ogically irreversible* manipulation of
> information, such as *erasing a bi*t, must dissipate a minimum amount of
> heat to the surroundings.
>
> The theoretical lower limit for the energy cost of erasing one bit of
> information is:
>
> E=K*T*ln2
>
> Where K is the Boltzmann constant and T is the absolute temperature of
> the circuit. In standard "irreversible" computing, operations like *AND*
> or *OR* gates are inherently dissipative because they have two inputs but
> only one output. You cannot "undo" the operation to find the original
> inputs, meaning information has been lost (erased), and heat must be
> released.
> *The Practical Catch: Entropy and Speed*
>
> While the "energy per logic operation" can go to zero in theory, we face
> two massive engineering hurdles:
>
>    1.
>
>    *The Adiabatic Requirement:* To keep energy dissipation near zero, the
>    system must change states very slowly (adiabatically). *The faster you
>    want the answer, the more energy you inevitably dissipate* due to
>    non-equilibrium effects
>    2.
>
>    .*Error Correction:* In the real world, thermal noise causes bits to
>    flip. Correcting these errors usually involves measuring a state and
>    resetting it—which is an irreversible act of erasure that triggers
>    Landauer’s limit anyway.
>
> *Me: So the less energy you use the more time it would take to get an
> answer? *
>
> Gemini: *Exactly*. That is the fundamental trade-off in *adiabatic*
> (reversible) computing. In a perfect world, *if you wanted to use zero
> energy, you would have to take an infinite amount of time to complete the
> calculation*.
>
> In physics, this is often described as staying in *thermodynamic
> equilibrium*.
> *The Time-Energy Trade-off*
>
> To understand why speed costs energy, you can think of the computer's
> state like a ball in a landscape of hills and valleys (potential energy).
>
>    -
>
>    *Fast Calculation:* To flip a bit quickly, you have to "push" the
>    system hard. This creates turbulence, friction, and heat. The faster you
>    move the system from state A to state B, the more you deviate from
>    equilibrium, and that deviation results in energy being lost to the
>    environment as entropy.
>    -
>
>    *Slow (Reversible) Calculation:* If you move the system
>    "infinitesimally" slowly, you can keep it in a state of near-equilibrium.
>    By the time you reach the end of the calculation, you haven't "splashed"
>    any energy out of the system.
>
>
> The "Friction" of Information
>
> In a standard processor, we use a lot of energy to slam transistors on and
> off to reach gigahertz speeds (10^9 cycles per second). In a reversible
> system, we use *adiabatic switching*, where the energy used to charge a
> logic gate is actually recovered and "recycled" back into the power supply
> instead of being dumped as heat.
>
> However, the efficiency of this recycling depends entirely on the speed:
>
>    1.
>
>    *Energy dissipation* in these circuits is generally proportional to
>    the clock frequency .
>    2.
>
>    Therefore, *if you want to cut your energy loss in half, you have to
>    cut your processing speed in half*.
>
>
> *   John K Clark*
>

I had this conversation with GPT, which went into much greater detail:

https://chatgpt.com/share/69cc534f-65bc-8329-81b2-5b0c83b47143

Which adds:
"However, the proportionality constant is technology-dependent, not
universal, and the literature still treats the fully general,
technology-independent lower bound on reversible-computing dissipation as a
function of speed as an open problem. Some analyses even suggest
better-than-1/�� suppression may be possible in certain quantum-adiabatic
regimes."

My reading of this is that energy loss is implementation dependent, not a
result of any physical lower bound on energy dissipation.

Jason

>
>>
>>
>>> *>> For great intelligence to be useful an animal needs hands with
>>>>> opposable thumbs or some other organ that can delicately manipulate matter,*
>>>>>
>>>>
>>>> *> When there are social dynamics and kne must kit think others of your
>>>> same species to win mates, then there's no upper bound on selection
>>>> pressure for intelligence. That may explain what happens with whales.*
>>>>
>>>
>>> *Perhaps so, and echolocation requires a great deal of data processing,
>>> but bigger animals require bigger brains than smaller animals do, and I
>>> think the intelligence of whales has been overestimated. Whales were hunted
>>> almost to extinction and water is an excellent conductor of sound so the
>>> sounds of whaling boats must have been audible for hundreds of miles, and
>>> yet they never learn to avoid them. And there is the phenomenon of mass
>>> whale beachings which doesn't exactly enhance their reputation for being
>>> bright. Whales certainly never made a radio telescope, or made anything at
>>> all for that matter except for other whales.  *
>>>
>>>
>>>> *> Their bodies are so large that it costs them very metabolically
>>>> little to have a much larger brain. The relative benefits even if minor,
>>>> can be justified. I think this explains why larger animals tend to have
>>>> larger brains. Not because so many larger brains are needed to control a
>>>> larger body, but because a larger brain can be supported more easily*
>>>>
>>>
>>> *A whale's brain is about 5 times as massive as a human brain but a
>>> whale's body is about 500 times as massive as a human body, and pound for
>>> pound a brain uses about 10 times as much energy as any other parts of the
>>> body.  *
>>>
>>> * >> **If a zebra on the African Savanna had an IQ of 200 that wouldn't
>>>>> help get its genes into the next generation very much, and that's why it
>>>>> never evolved to get that smart. *
>>>>>
>>>>
>>>> *> Why then are crows so smart?*
>>>>
>>>
>>> *Crows are not smart enough to build a radio telescope nor are they
>>> likely to evolve into something that could because, although they can learn
>>> to open milk bottles with their beaks, they have no way to manipulate
>>> matter delicate enough to repair a watch in the way that a human can. It's
>>> unclear what environmental factors caused our hominid ancestors to walk
>>> bipedally, but the first one to do so had a brain no larger than that of a
>>> chimpanzee; but after that and it had 2 limbs that could be used for things
>>> other than locomotion the brain size of its descendants grew at an
>>> extremely rapid rate. *
>>>
>>> *>> And yet none of those species have even come close to building a
>>>>> radio telescope, in the last 3.8 billion years only one species has managed
>>>>> to do so. *
>>>>
>>>>
>>>> *> You could have said the same about us only a few tens of thousands
>>>> of years ago.*
>>>
>>>
>>> *It's much worse than that, you could say the same thing about us just a
>>> century ago. And that is exactly why it's so bizarre that we have seen no
>>> evidence that this galaxy, or any other galaxy, has been engineered. The
>>> most obvious explanation for that anomaly is that we are the first.
>>> Somebody has to be. *
>>>
>>> * John K Clark  *
>>>
>>>
>>>
>>>
>>>>
>>>> *Perhaps a brilliant zebra would have a few minor advantages but
>>>>> unless it had opposable thumbs or something equivalent it would not
>>>>> be worth the price it would have to pay for being smart. The human brain
>>>>> only amounts to  2% of the body weight of a human but it consumes 20% of
>>>>> the body's energy.  *
>>>>>
>>>>> *And there are other disadvantages in having a large brain, a baby
>>>>> must get through a female's birth canal, and that means most of the growth
>>>>> of the brain must occur after birth, and that means for many years after
>>>>> birth the young are completely helpless, and that places a huge burden on
>>>>> the parents that can last for over a decade.  *
>>>>>
>>>>
>>>>
>>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20260331/b74da7cc/attachment.htm>