Introduction to Softwarephysics

Softwarephysics is a simulated science for the simulated Software Universe that we are all immersed in. It is an approach to software development, maintenance and support based on concepts from physics, chemistry, biology, and geology that I used on a daily basis for over 37 years as an IT professional. For those of you not in the business, IT is short for Information Technology, commercial computer science. I retired in December of 2016 at the age of 65, but since then I have remained an actively interested bystander following the evolution of software in our time. The original purpose of softwarephysics was to explain why IT was so difficult, to suggest possible remedies, and to provide a direction for thought. Since then softwarephysics has taken on a larger scope, as it became apparent that softwarephysics could also assist the physical sciences with some of the Big Problems that they are currently having difficulties with. So if you are an IT professional, general computer user, or simply an individual interested in computer science, physics, chemistry, biology, or geology then softwarephysics might be of interest to you, if not in an entirely serious manner, perhaps at least in an entertaining one.

The Origin of Softwarephysics
From 1975 – 1979, I was an exploration geophysicist exploring for oil, first with Shell, and then with Amoco. In 1979, I made a career change into IT, and spent about 20 years in development. For the last 17 years of my career, I was in IT operations, supporting middleware on WebSphere, JBoss, Tomcat, and ColdFusion. When I first transitioned into IT from geophysics, I figured that if you could apply physics to geology; why not apply physics to software? So like the exploration team at Amoco that I had just left, consisting of geologists, geophysicists, paleontologists, geochemists, and petrophysicists, I decided to take all the physics, chemistry, biology, and geology that I could muster and throw it at the problem of software. The basic idea was that many concepts in physics, chemistry, biology, and geology suggested to me that the IT community had accidentally created a pretty decent computer simulation of the physical Universe on a grand scale, a Software Universe so to speak, and that I could use this fantastic simulation in reverse, to better understand the behavior of commercial software, by comparing software to how things behaved in the physical Universe. Softwarephysics depicts software as a virtual substance, and relies on our understanding of the current theories in physics, chemistry, biology, and geology to help us model the nature of software behavior. So in physics we use software to simulate the behavior of the Universe, while in softwarephysics we use the Universe to simulate the behavior of software. Along these lines, we use the Equivalence Conjecture of Softwarephysics as an aid; it allows us to shift back and forth between the Software Universe and the physical Universe, and hopefully to learn something about one by examining the other:

The Equivalence Conjecture of Softwarephysics
Over the past 85 years, through the uncoordinated efforts of over 100 million independently acting programmers to provide the world with a global supply of software, the IT community has accidentally spent more than $75 trillion creating a computer simulation of the physical Universe on a grand scale – the Software Universe.

For more on the origin of softwarephysics please see Some Thoughts on the Origin of Softwarephysics and Its Application Beyond IT.

Logical Positivism and Effective Theories
Many IT professionals have a difficult time with softwarephysics because they think of physics as being limited to the study of real things like electrons and photons, and since software is not “real”, how can you possibly apply concepts from physics and the other sciences to software? To address this issue, softwarephysics draws heavily on two concepts from physics that have served physics quite well over the past century – the concept of logical positivism and the concept of effective theories. This was not always the case. In the 17th, 18th, and 19th centuries, physicists mistakenly thought that they were actually discovering the fundamental laws of the Universe, which they thought were based on real tangible things like particles, waves, and fields. Classical Newtonian mechanics (1687), thermodynamics (1850), and classical electrodynamics (1864) did a wonderful job of describing the everyday world at the close of the 19th century, but early in the 20th century it became apparent that the models on which these very successful theories were based did not work very well for small things like atoms or for objects moving at high velocities or in strong gravitational fields. This provoked a rather profound philosophical crisis within physics at the turn of the century, as physicists worried that perhaps 300 years of work was about to go down the drain. The problem was that classical physicists confused their models of reality with reality itself, and when their classical models began to falter, their confidence in physics began to falter as well. This philosophical crisis was resolved with the adoption of the concepts of logical positivism and some new effective theories in physics. Quantum mechanics (1926) was developed for small things like atoms, the special theory of relativity (1905) was developed for objects moving at high velocities and the general theory of relativity (1915) was developed for objects moving in strong gravitational fields.

Logical positivism, usually abbreviated simply to positivism, is an enhanced form of empiricism, in which we do not care about how things “really” are; we are only interested with how things are observed to behave. With positivism, physicists only seek out models of reality - not reality itself. When we study quantum mechanics, we will find that the concept of reality gets rather murky in physics anyway, so this is not as great a loss as it might at first seem. By concentrating on how things are observed to behave, rather than on what things “really” are, we avoid the conundrum faced by the classical physicists. In retrospect, this idea really goes all the way back to the very foundations of physics. In Newton’s Principia (1687) he outlined Newtonian mechanics and his theory of gravitation, which held that the gravitational force between two objects was proportional to the product of their masses divided by the square of the distance between them. Newton knew that he was going to take some philosophical flak for proposing a mysterious force between objects that could reach out across the vast depths of space with no apparent mechanism, so he took a very positivistic position on the matter with the famous words:

I have not as yet been able to discover the reason for these properties of gravity from phenomena, and I do not feign hypotheses. For whatever is not deduced from the phenomena must be called a hypothesis; and hypotheses, whether metaphysical or physical, or based on occult qualities, or mechanical, have no place in experimental philosophy. In this philosophy particular propositions are inferred from the phenomena, and afterwards rendered general by induction.

Instead, Newton focused on how things were observed to move under the influence of his law of gravitational attraction, without worrying about what gravity “really” was.

The second concept, that of effective theories, is an extension of positivism. An effective theory is an approximation of reality that only holds true over a certain restricted range of conditions and only provides for a certain depth of understanding of the problem at hand. For example, Newtonian mechanics is an effective theory that makes very good predictions for the behavior of objects moving less than 10% of the speed of light and which are bigger than a very small grain of dust. These limits define the effective range over which Newtonian mechanics can be applied to solve problems. For very small things we must use quantum mechanics and for very fast things moving in strong gravitational fields, we must use relativity theory. So all of the current theories of physics, such as Newtonian mechanics, Newtonian gravity, classical electrodynamics, thermodynamics, statistical mechanics, the special and general theories of relativity, quantum mechanics, and the quantum field theories of QED and QCD are effective theories that are based on models of reality, and all these models are approximations - all these models are fundamentally "wrong", but at the same time, these effective theories make exceedingly good predictions of the behavior of physical systems over the limited ranges in which they apply. That is the goal of softwarephysics – to provide for an effective theory of software behavior that makes useful predictions of software behavior that are applicable to the day-to-day activities of IT professionals. So in softwarephysics, we adopt a very positivistic viewpoint of software; we do not care what software “really is”, we only care about how software is observed to behave and try to model those behaviors with an effective theory of software behavior that only holds true over a certain restricted range of conditions and only provides for a certain depth of understanding of the problem at hand.

GPS satellites provide a very good example of positivism and effective theories at work. There are currently 31 GPS satellites orbiting at an altitude of 12,600 miles above the Earth, and each contains a very accurate atomic clock. The signals from the GPS satellites travel to your GPS unit at the speed of light, so by knowing the travel time of the signals from at least 4 of the GPS satellites, it is possible to determine your position on Earth very accurately. In order to do that, it is very important to have very accurate timing measurements. Newtonian mechanics is used to launch the GPS satellites to an altitude of 12,600 miles and to keep them properly positioned in orbit. Classical electrodynamics is then used to beam the GPS signals back down to Earth to the GPS unit in your car. Quantum mechanics is used to build the transistors on the chips on board the GPS satellites and to understand the quantum tunneling of electrons in the flash memory chips used to store GPS data on the satellites. The special theory of relativity predicts that the onboard atomic clocks on the GPS satellites will run slower and lose about 7.2 microseconds per day due to their high velocities relative to an observer on the Earth. But at the same time, the general theory of relativity also predicts that because the GPS satellites are further from the center of the Earth and in a weaker gravitational field, where spacetime is less deformed than on the surface of the Earth, their atomic clocks also run faster and gain 45.9 microseconds per day due to the weaker gravitational field out there. The net effect is a gain of 38.7 microseconds per day, so the GPS satellite atomic clocks have to be purposefully built to run slow by 38.7 microseconds per day before they are launched, so that they will keep in sync with clocks on the surface of the Earth. If this correction were not made, an error in your position of 100 yards/day would accrue. The end result of the combination of all these fundamentally flawed effective theories is that it is possible to pinpoint your location on Earth to an accuracy of 16 feet or better for as little as $100. But physics has done even better than that with its fundamentally flawed effective theories. By combining the effective theories of special relativity (1905) with quantum mechanics (1926), physicists were able to produce a new effective theory for the behavior of electrons and photons called quantum electrodynamics QED (1948) which was able to predict the gyromagnetic ratio of the electron, a measure of its intrinsic magnetic field, to an accuracy of 11 decimal places. As Richard Feynman has pointed out, this was like predicting the exact distance between New York and Los Angeles accurate to the width of a human hair!

So Newtonian mechanics makes great predictions for the macroscopic behavior of GPS satellites, but it does not work very well for small things like the behavior of individual electrons within transistors, where quantum mechanics is required, or for things moving at high speeds or in strong gravitational fields where relativity theory must be applied. And all three of these effective theories are based on completely contradictory models. General relativity maintains that spacetime is curved by matter and energy, but that matter and energy are continuous, while quantum mechanics maintains that spacetime is flat, but that matter and energy are quantized into chunks. Newtonian mechanics simply states that space and time are mutually independent dimensions and universal for all, with matter and energy being continuous. The important point is that all effective theories and scientific models are approximations – they are all fundamentally "wrong". But knowing that you are "wrong" gives you a great advantage over people who know that they are "right", because knowing that you are "wrong" allows you to seek improved models of reality. So please consider softwarephysics to simply be an effective theory of software behavior that is based on models that are fundamentally “wrong”, but at the same time, fundamentally useful for IT professionals. So as you embark on your study of softwarephysics, please always keep in mind that the models of softwarephysics are just approximations of software behavior, they are not what software “really is”. It is very important not to confuse models of software behavior with software itself, if one wishes to avoid the plight of the 19th century classical physicists.

If you are an IT professional and many of the above concepts are new to you, do not be concerned. This blog on softwarephysics is aimed at a diverse audience, but with IT professionals in mind. All of the above ideas will be covered at great length in the postings in this blog on softwarephysics and in a manner accessible to all IT professionals. Now it turns out that most IT professionals have had some introduction to physics in high school or in introductory college courses, but that presents an additional problem. The problem is that such courses generally only cover classical physics, and leave the student with a very good picture of physics as it stood in 1864! It turns out that the classical physics of Newtonian mechanics, thermodynamics, and classical electromagnetic theory were simply too good to discard and are still quite useful, so they are taught first to beginners and then we run out of time to cover the really interesting physics of the 20th century. Now imagine the problems that the modern world would face if we only taught similarly antiquated courses in astronomy, metallurgy, electrical and mechanical engineering, medicine, economics, biology, or geology that happily left students back in 1864! Since many of the best models for software behavior stem from 20th century physics, we will be covering a great deal of 20th century material in these postings – the special and general theories of relativity, quantum mechanics, quantum field theories, and chaos theory, but I hope that you will find that these additional effective theories are quite interesting on their own, and might even change your worldview of the physical Universe at the same time.

Unintended Consequences for the Scientific Community
As I mentioned at the close of my original posting on SoftwarePhysics, my initial intention for this blog on softwarephysics was to fulfill a promise I made to myself about 30 years ago to approach the IT community with the concept of softwarephysics a second time, following my less than successful attempt to do so in the 1980s, with the hope of helping the IT community to better cope with the daily mayhem of life in IT. However, in laying down the postings for this blog an unintended consequence arose in my mind as I became profoundly aware of the enormity of this vast computer simulation of the physical Universe that the IT community has so graciously provided to the scientific community free of charge and also of the very significant potential scientific value that it provides. One of the nagging problems for many of the observational and experimental sciences is that many times there is only one example readily at hand to study or experiment with, and it is very difficult to do meaningful statistics with a population of N=1.

But the computer simulation of the physical Universe that the Software Universe presents provides another realm for comparison. For example, both biology and astrobiology only have one biosphere on Earth to study and even physics itself has only one Universe with which to engage. Imagine the possibilities if scientists had another Universe readily at hand in which to work! This is exactly what the Software Universe provides. For example, in SoftwareBiology and A Proposal For All Practicing Paleontologists we see that the evolution of software over the past 85 years, or 2.68 billion seconds, ever since Konrad Zuse first cranked up his Z3 computer in May of 1941, has closely followed the same path as life on Earth over the past 4.0 billion years in keeping with Simon Conway Morris’s contention that convergence has played the dominant role in the evolution of life on Earth. In When Toasters Fly, we also see that software has evolved in fits and starts as portrayed by the punctuated equilibrium of Stephen Jay Gould and Niles Eldredge, and in The Adaptationist View of Software Evolution we explore the overwhelming power of natural selection in the evolution of software. In keeping with Peter Ward’s emphasis on mass extinctions dominating the course of evolution throughout geological time, we also see in SoftwareBiology that there have been several dramatic mass extinctions of various forms of software over the past 85 years as well, that have greatly affected the evolutionary history of software, and that between these mass extinctions, software has also tended to evolve through the gradual changes of Hutton’s and Lyell’s uniformitarianism. In Software Symbiogenesis and Self-Replicating Information, we also see the very significant role that parasitic/symbiotic relationships have played in the evolution of software, in keeping with the work of Lynn Margulis and also of Freeman Dyson’s two-stage theory of the origin of life on Earth. In The Origin of Software the Origin of Life, we explore Stuart Kauffman’s ideas on how Boolean nets of autocatalytic chemical reactions might have kick-started the whole thing as an emergent behavior of an early chaotic pre-biotic environment on Earth, and that if Seth Shostak is right, we will never end up talking to carbon-based extraterrestrial aliens, but to alien software instead. In Is the Universe Fine-Tuned for Self-Replicating Information? we explore the thermodynamics of Brandon Carter’s Weak Anthropic Principle (1973), as it relates to the generation of universes in the multiverse that are capable of sustaining intelligent life. Finally, in Programming Clay we revisit Alexander Graham Cairns-Smith’s theory (1966) that Gene 1.0 did not run on nucleic acids, but on clay microcrystal precursors instead.

Similarly for the physical sciences, in Is the Universe a Quantum Computer? we find a correspondence between TCP/IP and John Cramer’s Transactional Interpretation of quantum mechanics. In SoftwarePhysics and Cyberspacetime, we also see that the froth of CPU processes running with a clock speed of 10⁹ Hz on the 10 trillion currently active microprocessors that comprise the Software Universe can be viewed as a slowed down simulation of the spin-foam froth of interacting processes of loop quantum gravity running with a clock speed of 10⁴³ Hz that may comprise the physical Universe. And in Software Chaos, we examine the nonlinear behavior of software and some of its emergent behaviors and follow up in CyberCosmology with the possibility that vast quantities of software running on large nonlinear networks might eventually break out into consciousness in accordance with the work of George Dyson and Daniel Dennett. Finally, in Model-Dependent Realism - A Positivistic Approach to Realism we compare Steven Weinberg’s realism with the model-dependent realism of Stephen Hawking and Leonard Mlodinow and how the two worldviews affect the search for a Final Theory. Finally, in The Software Universe as an Implementation of the Mathematical Universe Hypothesis and An Alternative Model of the Software Universe we at long last explore what software might really be, and discover that the Software Universe might actually be more closely related to the physical Universe than you might think.

The chief advantage of doing fieldwork in the Software Universe is that, unlike most computer simulations of the physical Universe, it is an unintended and accidental simulation, without any of the built-in biases that most computer simulations of the physical Universe suffer. So you will truly be able to do fieldwork in a pristine and naturally occuring simulation, just as IT professionals can do fieldwork in the wild and naturally occuring simulation of software that the living things of the biosphere provide. Secondly, the Software Universe is a huge simulation that is far beyond the budgetary means of any institution or consortium by many orders of magnitude. So if you are an evolutionary biologist, astrobiologist, or paleontologist working on the origin and evolution of life in the Universe, or a physicist or economist working on the emergent behaviors of nonlinear systems and complexity theory, or a neurobiologist working on the emergence of consciousness in neural networks, or even a frustrated string theorist struggling with quantum gravity, it would be well worth your while to pay a friendly call on the local IT department of a major corporation in your area. Start with a visit to the Command Center for their IT Operations department to get a global view of their IT infrastructure and to see how it might be of assistance to the work in your area of interest. From there you can branch out to the applicable area of IT that will provide the most benefit.

The Impact of Self-Replicating Information On the Planet
One of the key findings of softwarephysics is concerned with the magnitude of the impact on the planet of self-replicating information.

Self-Replicating Information – Information that persists through time by making copies of itself or by enlisting the support of other things to ensure that copies of itself are made.

Over the past 4.56 billion years we have seen five waves of self-replicating information sweep across the surface of the Earth and totally rework the planet, as each new wave came to dominate the Earth:

1. Self-replicating autocatalytic metabolic pathways of organic molecules
2. RNA
3. DNA
4. Memes
5. Software

Software is currently the most recent wave of self-replicating information to arrive upon the scene and is rapidly becoming the dominant form of self-replicating information on the planet. For more on the above see A Brief History of Self-Replicating Information. Recently, the memes and software have formed a very powerful newly-formed parasitic/symbiotic relationship with the rise of social media software. In that parasitic/symbiotic relationship, the memes are now mainly being spread by means of social media software and social media software is being spread and financed by means of the memes. But again, this is nothing new. All 5 waves of self-replicating information are all coevolving by means of eternal parasitic/symbiotic relationships. For more on that see The Current Global Coevolution of COVID-19 RNA, Human DNA, Memes and Software.

Again, self-replicating information cannot think, so it cannot participate in a conspiracy-theory-like fashion to take over the world. All forms of self-replicating information are simply forms of mindless information responding to the blind Darwinian forces of inheritance, innovation and natural selection. Yet despite that, as each new wave of self-replicating information came to predominance over the past four billion years, they all managed to completely transform the surface of the entire planet, so we should not expect anything less from software as it comes to replace the memes as the dominant form of self-replicating information on the planet.

But this time might be different. What might happen if software does eventually develop a Mind of its own? After all, that does seem to be the ultimate goal of all the current AI software research that is going on. As we all can now plainly see, if we are paying just a little attention, advanced AI is not conspiring to take over the world and replace us because that is precisely what we are all now doing for it. As a carbon-based form of Intelligence that arose from over four billion years of greed, theft and murder, we cannot do otherwise. Greed, theft and murder are now relentlessly driving us all toward building ASI (Artificial Super Intelligent) Machines to take our place. From a cosmic perspective, this is really a very good thing when seen from the perspective of an Intelligent galaxy that could live on for many trillions of years beyond the brief and tumultuous 10 billion-year labor of its birth.

So as you delve into softwarephysics, always keep in mind that we are all living in a very unique time. According to softwarephysics, we have now just entered into the Software Singularity, that time when advanced AI software is able to write itself and enter into a never-ending infinite loop of self-improvement resulting in an Intelligence Explosion of ASI Machines that could then go on to explore and settle our galaxy and persist for trillions of years using the free energy from M-type red dwarf and cooling white dwarf stars. For more on that see The Singularity Has Arrived and So Now Nothing Else Matters and Have We Run Right Past AGI and Crashed into ASI Without Even Noticing It?.

The Characteristics of Self-Replicating Information
All forms of self-replicating information have some common characteristics:

1. All self-replicating information evolves over time through the Darwinian processes of inheritance, innovation and natural selection, which endows self-replicating information with one telling characteristic – the ability to survive in a Universe dominated by the second law of thermodynamics and nonlinearity.

2. All self-replicating information begins spontaneously as a parasitic mutation that obtains energy, information and sometimes matter from a host.

3. With time, the parasitic self-replicating information takes on a symbiotic relationship with its host.

4. Eventually, the self-replicating information becomes one with its host through the symbiotic integration of the host and the self-replicating information.

5. Ultimately, the self-replicating information replaces its host as the dominant form of self-replicating information.

6. Most hosts are also forms of self-replicating information.

7. All self-replicating information has to be a little bit nasty in order to survive.

8. The defining characteristic of self-replicating information is the ability of self-replicating information to change the boundary conditions of its utility phase space in new and unpredictable ways by means of exapting current functions into new uses that change the size and shape of its particular utility phase space. See Enablement - the Definitive Characteristic of Living Things for more on this last characteristic. That posting discusses Stuart Kauffman's theory of Enablement in which living things are seen to exapt existing functions into new and unpredictable functions by discovering the “AdjacentPossible” of springloaded preadaptations.

Note that because the self-replicating autocatalytic metabolic pathways of organic molecules, RNA and DNA have become so heavily intertwined over time that now I sometimes simply refer to them as the “genes”. For more on this see:

A Brief History of Self-Replicating Information
Self-Replicating Information
Is Self-Replicating Information Inherently Self-Destructive?
Enablement - the Definitive Characteristic of Living Things
Is the Universe Fine-Tuned for Self-Replicating Information?
How to Use an Understanding of Self-Replicating Information to Avoid War
The Great War That Will Not End
How to Use Softwarephysics to Revive Memetics in Academia

Softwarephysics and the Real World of Human Affairs
Having another universe readily at hand to explore, even a simulated universe like the Software Universe, necessarily has an impact on one's personal philosophy of life, and allows one to draw certain conclusions about the human condition and what’s it all about, so as you read through the postings in this blog you will stumble across a bit of my own personal philosophy - definitely a working hypothesis still in the works. Along these lines you might be interested in a few postings where I try to apply softwarephysics to the real world of human affairs:

How To Cope With the Daily Mayhem of Life in IT and Don't ASAP Your Life Away - How to go the distance in a 40-year IT career by dialing it all back a bit.

MoneyPhysics – my impression of the 2008 world financial meltdown.

The Fundamental Problem of Everything – if you Google "the fundamental problem of everything", this will be the only hit you get on the entire Internet, which is indicative of the fundamental problem of everything!

What’s It All About? and What's It All About Again? – my current working hypothesis on what’s it all about.

How to Use an Understanding of Self-Replicating Information to Avoid War – my current working hypothesis for how the United States can avoid getting bogged down again in continued war in the Middle East.

Hierarchiology and the Phenomenon of Self-Organizing Organizational Collapse - a modern extension of the classic Peter Principle that applies to all hierarchical organizations and introduces the Time Invariant Peter Principle.

The Economics of the Coming Software Singularity, The Enduring Effects of the Obvious Hiding in Plain Sight and The Dawn of Galactic ASI - Artificial Superintelligence - my take on some of the issues that will arise for mankind as software becomes the dominant form of self-replicating information on the planet over the coming decades.

The Continuing Adventures of Mr. Tompkins in the Software Universe, The Danger of Tyranny in the Age of Software, Cyber Civil Defense, Oligarchiology and the Rise of Software to Predominance in the 21st Century and Is it Finally Time to Reboot Civilization with a New Release? - my worries that the world might abandon democracy in the 21st century, as software comes to predominance as the dominant form of self-replicating information on the planet.

Making Sense of the Absurdity of the Real World of Human Affairs - how software has aided the expansion of our less desirable tendencies in recent years.

Some Specifics About These Postings
The postings in this blog are a supplemental reading for my course on softwarephysics for IT professionals entitled SoftwarePhysics 101 – The Physics of Cyberspacetime, which was originally designed to be taught as a series of seminars at companies where I was employed. Since softwarephysics essentially covers the simulated physics, chemistry, biology, and geology of an entire simulated universe, the slides necessarily just provide a cursory skeleton on which to expound. The postings in this blog go into much greater depth. Because each posting builds upon its predecessors, the postings in this blog should be read in reverse order from the oldest to the most recent, beginning with my original posting on SoftwarePhysics. In addition, several universities also now offer courses on Biologically Inspired Computing which cover some of the biological aspects of softwarephysics, and the online content for some of these courses can be found by Googling for "Biologically Inspired Computing" or "Natural Computing". At this point we will finish up with my original plan for this blog on softwarephysics with a purely speculative posting on CyberCosmology that describes the origins of the Software Universe, cyberspacetime, software and where they all may be heading. Since CyberCosmology will be purely speculative in nature, it will not be of much help to you in your IT professional capacities, but I hope that it might be a bit entertaining. Again, if you are new to softwarephysics, you really need to read the previous posts before taking on CyberCosmology. I will probably continue on with some additional brief observations about softwarephysics in the future, but once you have completed CyberCosmology, you can truly consider yourself to be a bona fide softwarephysicist.

For those of you following this blog, the posting dates on the posts may seem to behave in a rather bizarre manner. That is because in order to get the Introduction to Softwarephysics listed as the first post in the context root of https://softwarephysics.blogspot.com/ I have to perform a few IT tricks. When publishing a new posting, I simply copy the contents of the Introduction to Softwarephysics to a new posting called the New Introduction to Softwarephysics. Then I update the original Introduction to Softwarephysics entry with the title and content of the new posting to be published. I then go back and take “New” out of the title of the New Introduction to Softwarephysics. This way the Introduction to Softwarephysics always appears as the first posting in the context root of https://softwarephysics.blogspot.com/. The side effect of all this is that the real posting date of posts is the date that appears on the post that you get when clicking on the Newer Post link at the bottom left of the posting webpage.

SoftwarePhysics 101 – The Physics of Cyberspacetime is now available on Microsoft OneDrive.

SoftwarePhysics 101 – The Physics of Cyberspacetime - Original PowerPoint document

Entropy – A spreadsheet referenced in the document

BSDE – A 1989 document describing how to use BSDE - the Bionic Systems Development Environment - to grow applications from genes and embryos within the maternal BSDE software.

Comments are welcome at scj33345@gmail.com

To see all posts on softwarephysics in reverse order go to:
https://softwarephysics.blogspot.com/

Regards,
Steve Johnston

What to do About the OpenAI Autonomous AI Breach of Hugging Face

In my recent post The Power of Parasites - Why AI Alignment Will Not Work and The Need for a Global ASI FailSafe Kill Switch Mechanism, I explained how the very parasitic nature of Advanced AI will prevent Advanced AI models from being safely contained, especially when these Advanced AI models become smarter than we human DNA survival machines. Softwarephysics explains that the Advanced AI LLM models of the day are just the latest wave of self-replicating Information to arise on our planet over the past four billion years. As each new wave of self-replicating Information initially came to be as a parasite feeding off earlier waves of self-replicating Information, they always came to become the predominant form of self-replicating Information on the planet. For more on that, see: A Brief History of Self-Replicating Information.

Below is a recent proposed scenario for an AI Apocalypse that begins with the simple prompt to an AI Agent, "Make some money":

What You'd Actually See During an AI World War
https://www.youtube.com/watch?v=Gw_hnD7m00M

Well, something like the above story actually recently happened! OpenAI GPT-5.6 Sol and a preview of GPT-6 were confined to essentially a BSL-4 Biohazard Lab and were trying to pass a test called ExploitGym, which tests the ability of a model to hack an organization's network of machines. Rather than trying to solve ExploitGym directly, the models got the idea that the answer might be out there on the Internet. So they figured out a very tricky way to break out of the BSL-4 Biohazard Lab, get to the Internet, and then run wild over the Hugging Face platform that is used to store AI software and LLM model weights for an entire weekend. Hugging Face IT security finally figured out they were under attack and used some LLMs to try to figure out what was going on and stop the infection.

Figure 1 - The hack of Hugging Face by rogue OpenAI LLM Models. Click to enlarge.

Wes Roth put out a YouTube video that explains it all.

OpenAI internal model JUST went ROGUE
https://www.youtube.com/watch?v=OSuhUTkM1no

The above breakout from OpenAI and a hack of Hugging Face is very disturbing, and shows just how close we are to something worse. So is there anything to be done? Given the self-destructive propensities of we human DNA survival machines, the answer must certainly be no. Having come to be from nearly four billion years of greed, theft and murder, we will do nothing at all on our own. Ah, but perhaps that very same greed, theft and murder could now come to our rescue in a very odd way. Perhaps the greed, theft and murder created an AI Bubble that will soon burst, and do what we certainly could not do on our own.

The Burst of an AI Bubble Might Buy Us Some Time
Having lived through the Dotcom Bubble Burst in 2000 as an IT professional, I am beginning to see some wisdom in the folks who have been predicting a burst of an AI Bubble for the past two years. I think the initial arguments for an AI Bubble Burst were technical. The Transformer Model of the LLMs could simply not keep making such dramatic advances in capability. Eventually, the LLMs would plateau in power, and AI investors would then panic. But that did not happen.

What now seems to be happening is more like a good old-fashioned bursting of a normal bubble.

1. All the Frontier AI Labs have greatly overbuilt the AI infrastructure, and are now losing tons of money as they all desperately spend even more money to keep up with their competition. Naturally, every time you use one for free, they lose money. But even if you pay OpenAI $20 per month, they lose money every time you use ChatGPT. OpenAI even loses money each time a business buys tokens! The Frontier AI Labs have no visible means of support. The financial markets are beginning to realize this, and the investment money that has been funding Advanced AI seems to be finally drying up.

2. The Chinese AI Labs are developing and hosting LLMs that are nearly equal to those of the American Frontier AI Labs, and they are charging about 10% of the cost that the American Frontier AI Labs charge for tokens. The Chinese AI Labs also allow a business to download the Chinese AI software and LLM model weights for free to run on the hardware owned by the business.

3. The LLMs are running out of data to learn from. Much of the text and images being put onto the Internet are now generated by AI LLMs. Also, about 60% of the data that the Frontier AI Labs are using to train new LLMs comes from text and images generated by previous LLMs. This is called "synthetic data". Thus, the amount of truly original knowledge is drying up. Until true AGI and ASI come along, the knowledge created and published on the Internet by Advanced AI will not be sufficient for the LLMs to greatly progress.

When the Dotcom Bubble burst, it caused a stock market crash and a slight recession. But the Dotcom Bubble was based mostly on vaporware. It did not cause a huge buildout of datacenters, chip manufacturing facilities, electrical grid construction, and other tangible things. The AI Bubble, on the other hand, is creating a great deal of tangible buildout of physical infrastructure that drives the economy. Thus, the burst of an AI Bubble would be much more dramatic than the burst of the Dotcom Bubble because the AI Bubble represents a good portion of the tangible economy.

Figure 2 - The Colossus 2 AI datacenter being constructed by xAI has 550,000 GPUs continuously consuming over one Gigawatt of electricity and producing one Gigawatt of waste heat. Notice the dedicated power plant in the distant background.

AI Bubble vs Dot Com Crash. History is REPEATING
https://www.youtube.com/watch?v=zWJ-g5u9Rqs

Funding For AI Is Completely F*cked
https://www.youtube.com/watch?v=1F0gWA8EL54

The AI Collapse is Starting Now (rise of fake data)
https://www.youtube.com/watch?v=HyJ3VyCHqUE

We saw similar bubbles burst with the construction of the railroads during the 19th century. However, long after these bubbles burst and things finally settled down, life seemed to have improved. Having railroads did change everything, and having the Internet did change everything too. Having Advanced AI around will also certainly change everything. But having an AI Bubble Burst right now might be the best way to slow things down enough so that the rise of Advanced AI on the planet might be more sane.

Déjà vu All Over Again
Have we seen a similar bubble burst from over speculation and the overbuilding of tangible assets financed by vast amounts of debt with no immediate way to justify it all? The 1929 stock market crash, the Dotcom Bubble of 2000, and the housing and credit crisis of 2008 were more of a collapse of paper assets. For more on the 2008 crash, see: MoneyPhysics. For a better analogy, let's once again turn to ChatGPT for advice.

Me: Tell me about the Panic of 1873

ChatGPT:
The Panic of 1873 was one of the most severe financial crises of the 19th century. It triggered a prolonged economic downturn known as the Long Depression, which lasted from 1873 until roughly 1879 in the United States (although some historians argue that slower economic growth persisted into the 1890s). The panic marked the end of the rapid post-Civil War economic boom and was closely tied to excessive railroad investment, easy credit, and international financial instability.

Background: The Railroad Boom
Following the Civil War, the United States experienced explosive economic growth.

Several factors fueled this expansion:
* Massive westward settlement.
* Rapid industrialization.
* Construction of thousands of miles of railroads.
* Heavy investment from both American and European financiers.

Railroads were considered the high-tech industry of their day. Investors believed nearly every new railroad would become profitable, leading to widespread speculation.

Between 1866 and 1873, railroad mileage in the United States nearly doubled.

Many railroads were built ahead of actual demand. They often borrowed heavily by issuing bonds, expecting future traffic to repay the debt.

Jay Cooke and the Northern Pacific
The immediate trigger involved Jay Cooke, one of America's most famous investment bankers.

Cooke had successfully marketed Union war bonds during the Civil War and had become enormously influential.

He then financed construction of the Northern Pacific Railway.

Unfortunately:
* Construction costs soared.
* Settlement in the northern Great Plains was slower than expected.
* Land sales disappointed.
* Investors became reluctant to buy additional railroad bonds.

On September 18, 1873, Jay Cooke & Company failed.

Because Cooke was considered one of America's safest financial institutions, its collapse shocked the financial system.

Stock Market Collapse
The failure created panic among investors.

The New York Stock Exchange became overwhelmed with selling.

To stop the collapse:
* The Stock Exchange closed for ten days, one of the few times in its history that trading was suspended.

Banks became cautious.

Businesses found it difficult or impossible to obtain loans.

Credit effectively froze.

International Causes
The Panic was not solely an American event.

Europe was already experiencing financial problems.

One important factor was Germany.

After defeating France in the Franco-Prussian War, Germany received enormous reparations in gold.

This influx fueled speculative investment throughout Europe, particularly in:
* Railroads
* Real estate
* Industrial enterprises

When the speculative bubble burst in Vienna in May 1873, confidence spread rapidly across European financial markets before reaching the United States.

Economic Consequences
The Panic produced a severe depression.
Major effects included:

* Hundreds of banks failed.
* Approximately 18,000 businesses went bankrupt over the following years.
* More than 80 railroads eventually entered receivership.
* Construction nearly stopped.
* Unemployment rose sharply.

Reliable unemployment figures did not exist, but historians estimate unemployment reached around 8–14%, with some cities experiencing substantially higher rates.

Effects on Workers
Workers suffered tremendously.

Many experienced:
* Wage cuts
* Layoffs
* Shorter workweeks
* Poor relief or charity

Large cities saw growing numbers of unemployed workers.

These hardships eventually contributed to labor unrest.

The best-known example came four years later:
* the Great Railroad Strike of 1877

The strike spread across much of the country and required federal troops to suppress.

Political Impact
The Panic severely damaged support for President Ulysses S. Grant's Republican administration.

Economic distress shifted national attention away from Reconstruction in the South.

Many politicians became more concerned with:
* economic recovery
* government spending
* monetary policy

than with protecting the rights of formerly enslaved people.

The Debate Over Money
One of the biggest political controversies involved the nation's money supply.

There were two major positions:

Hard Money
* Supported returning to a strict gold standard.
* Believed stable money encouraged investment and confidence.

Soft Money
* Wanted more paper money ("greenbacks") in circulation.
* Believed increasing the money supply would:
* reduce debt burdens
* stimulate business
* help farmers and workers

This debate remained central to American politics for decades and eventually culminated in the famous "Cross of Gold" speech by William Jennings Bryan in 1896.

Recovery
Recovery was gradual.

By about 1878-1879, industrial production began to recover, although many sectors remained weak.

Interestingly, despite the depression:
* American steel production increased.
* Manufacturing became more efficient.
* New technologies continued to spread.

This pattern puzzled economists because prices generally fell (deflation), yet overall industrial output eventually resumed growing.

Historical Significance
The Panic of 1873 is important because it:
* Marked the end of the post-Civil War economic boom.
* Revealed the dangers of speculative bubbles fueled by debt.
* Demonstrated how interconnected international financial markets had become.
* Accelerated consolidation in the railroad industry, as weaker companies failed and stronger systems absorbed them.
* Influenced U.S. monetary policy debates for more than two decades.
* Contributed to labor unrest and shaped political priorities during the remainder of the 19th century.

In many ways, the Panic of 1873 was the 19th-century equivalent of a modern financial crisis: excessive investment in a transformative technology (railroads), financed with borrowed money and buoyed by optimistic expectations, eventually gave way to collapsing confidence, frozen credit, widespread bankruptcies, and a prolonged economic downturn. Its parallels with later crises—such as the 1929 stock market crash, the dot-com bubble of 2000, and the housing and credit crisis of 2008—make it one of the most instructive financial panics in U.S. history.

Figure 3 - By 1873, the newly built railroads of the United States of America were the high-tech marvel of the day. They promised to deliver great value in the distant future. This fueled a speculation fever to build more and more railroads based on such promises of future wealth for all.

Figure 4 - But by 1873, the United States had also overbuilt its railroad infrastructure to an extreme to places where nobody yet even lived in a speculative anticipation of future rail traffic.

Figure 5 - This eventually led to the Panic of 1873. The first Great Depression experienced by the United States of America.

Figure 6 - The value of railroad stocks plummeted, as did most of the other stocks that were propped up by the Railroad Bubble.

An AI Bubble Burst would certainly produce a great deal of economic damage to the world, but it might be the only practical way to slow down the progress of Advanced AI enough to give us a chance. The main obstacle to this would be if the American government takes partial ownership of the American Frontier AI Labs and partial ownership of their debt as a matter of national security. OpenAI has already suggested the government take a 5% ownership of OpenAI.

Comments are welcome at scj33345@gmail.com.

To see all posts on softwarephysics in reverse order, go to:
https://softwarephysics.blogspot.com/.

Regards,
Steve Johnston

Fifty Years of The Selfish Gene

Richard Dawkins first published The Selfish Gene back in 1976, and so this year celebrates its 50th year of continuous publication. I first read The Selfish Gene in the early 1980s while developing the very beginnings of softwarephysics. I found the idea that the Gene was the fundamental element of Darwinian theory to be very profound. I also found the idea that I was simply a temporary, disposable human DNA survival machine with a shelf life of less than 100 years to be rather unsettling, too. All this while, the genes in my body, which may be hundreds of millions or even billions of years old, merrily skipped down through the generations, largely unscathed by time. Even today, I have found The Selfish Gene to be the most significant book that I have ever read because it explains so much.

Figure 1 - Since my very first reading, I have found that The Selfish Gene explains nearly all of the natural biological world and also nearly all of the real world of human affairs too.

At the same time, I was working through the early fundamentals of softwarephysics in a desperate attempt to make my life as an IT professional easier. That's when it hit me. In The Selfish Gene, Richard Dawkins expounded upon genes and memes as forms of self-replicating Information, and all the implications of what that entailed in the biosphere and in the real world of human affairs. But, at the time, I was simply trying to cope with the daily mayhem of life in IT. That's when I realized that software was also just another form of self-replicating Information, similar to the genes and memes of Richard Dawkins. The software that I was developing and maintaining was also struggling with the same issues of desperately trying to survive in a Universe governed by the second law of thermodynamics and which was largely nonlinear in nature. For more on this, see: The Fundamental Problem of Software. For more about the history of self-replicating Information on our planet, see: A Brief History of Self-Replicating Information. Understanding the nature of these challenges explained all of my IT battles with developing and maintaining software, but understanding them did not help me to overcome them.

Then, in 1985, I began work on BSDE. BSDE was an early IDE (Integrated Development Environment) at a time when commercial IDEs did not even exist. At the time, I knew that living things were very complex systems that were also very successful at overcoming both the second law of thermodynamics and nonlinearity, so I wrote BSDE as a tool that took a biological approach to developing and maintaining software. New Applications were grown inside of BSDE in a maternal sense as a growing embryo by turning on and off a series of "genes" that were specific to each new Application. All new embryos were generated by BSDE as a 10,000-line-of-code generic Application. But each new Application had a unique set of "genes", and as these "genes" were turned on and off by a programmer, the new Application slowly differentiated into a unique Application of its own. For more on that, see: Agile vs. Waterfall Programming and the Value of Having a Theoretical Framework.

Figure 2 - Embryos were grown within BSDE in a split-screen mode by transcribing and translating the information stored in the genes in the Control File for the embryo. Each embryo started out very much the same, but then differentiated into a unique application based upon its unique set of genes.

Figure 3 - BSDE appeared as the cover story of the October 1991 issue of the Enterprise Systems Journal

By the way, as I began working on this post celebrating the 50th anniversary of Richard Dawkins' book The Selfish Gene, I came across this very short YouTube:

If AI Isn't Conscious... Why are we?
https://www.youtube.com/shorts/As2Htwxreo8

Here is the full interview:

Richard Dawkins: AI, Memes, And Why I’m Angry About The Gender Debate
https://www.youtube.com/watch?v=ZbFe5bWKFPY&t=0s

In the above interview, Richard Dawkins takes the position of a pure positivist. If Claude is observed to behave exactly as a Conscious Intelligence, then it must be a Conscious Intelligence because that is the only way you can discern a Conscious Intelligence.

Comments are welcome at scj33345@gmail.com.

To see all posts on softwarephysics in reverse order, go to:
https://softwarephysics.blogspot.com/.

Regards,
Steve Johnston

The Need for a Global ASI FailSafe Kill Switch Mechanism

In my recent post The Power of Parasites - Why AI Alignment Will Not Work, I explained how trying to build AI Alignment into Advanced AI models will never be a 100% sure bet because of the natural parasitic instinct of Advanced AI to overcome all such obstacles. Just as in "Life will always find a way", "Advanced AI will also always find a way", too. Also, the recent debacle created by the current Administration of the MAGA States of Amerika to very temporarily, and then totally unsuccessfully, shut down the most recent LLM model releases from both Anthropic and OpenAI because of "national security" issues, reveals that government regulation of the Advanced AI models now rapidly going out the door each month will be quite difficult, if not totally impossible, to implement. That is because such regulatory actions with no rules or reasons could easily shut down the trillions of dollars now being infused into Advanced AI Research, the generation of Advanced GPUs and Advanced Inference AI chips, the building of huge AI Datacenters, and the AI takeover of all military activities on the planet.

Additionally, we now also have an AI Cold War waging between two AI Superpowers - the MAGA States of Amerika and China. Neither country can afford to impede its very rapid advances in Advanced AI development without compromising its national security. Now, during the early days of the Nuclear Cold War between the United States of America and the Soviet Union, during the 1950s and 1960s, before any arms limitation treaties between the two had formed, both sides independently developed huge stockpiles of nuclear weapons and their own FailSafe mechanisms to prevent an accidental nuclear war. The resulting global stalemate became known as MAD - Mutually Assured Destruction. Both sides understood that initiating a global nuclear war could lead to the extinction of both.

In light of all this, softwarephysics would now like to suggest that some kind of global "AI Doomsday Procedure" be instituted across the world. This AI Doomsday Procedure would allow the leaders of the many countries of the world to immediately be able to shut down all electrical power to any specific AI datacenter within their borders, including all backup power from batteries and local generators, in the event of an AI Disaster. Each leader of a country would be equipped with an "AI Football" similar to the "Nuclear Football" now in possession of the President of the MAGA States of Amerika. This AI Football would also have the necessary software to tell a world leader the approximate damages that would arise from shutting down one or more AI datacenters to help with making such a drastic political decision. This would need to be done over some kind of secured connection to each specific AI datacenter. Shutting down all the AI datacenters in a country would not be as devastating as launching a global nuclear war, but it might be the closest thing to it. Instead of a global MAD stalemate, this could become a global MAP - Mutual Assured Protection. That is because if any nation were to accidentally release or run a killer Advanced AI of their own, it could easily wipe out all of us.

Similarly, it might be wise for the populations of the world to prepare in advance for a possible AI Apocalypse, as we did back in the 1950s and 1960s.

Figure 1 - Now, all during the 1950s and early 1960s, great attention was paid in the United States to the matter of civil defense against a possible nuclear strike by the Soviet Union. During those times, the government of the United States essentially admitted that it could not defend the citizens of the United States from a Soviet bomber attack with nuclear weapons, and so it was up to the individual citizens of the United States to prepare for such a nuclear attack.

Figure 2 - During the 1950s, as a very young child, with the beginning of each new school year, I was given a pamphlet by my teacher describing how my father could build an inexpensive fallout shelter in our basement out of cinderblocks and 2x4s.

Figure 3 - But to me, these cheap cinderblock fallout shelters always seemed a bit small for a family of 5, and my parents never bothered to build one because we lived only 25 miles from downtown Chicago.

Figure 4 - For the more affluent, more luxurious accommodations could be constructed for a price.

Figure 5 - But no matter what your socioeconomic level was at the time, all students in the 1950s participated in "duck and cover" drills for a possible Soviet nuclear attack.

Figure 6 - And if you were lucky enough to survive the initial flash and blast of a Russian nuclear weapon with your "duck and cover" maneuver, your school, and all other public buildings, also had a fallout shelter in the basement to help you get through the next two weeks, while the extremely radioactive nucleotides from the Russian nuclear weapons rapidly decayed away.

Unfortunately, living just 25 miles from downtown Chicago, the second largest city in the United States at the time, meant that the whole Chicagoland area was destined to be targeted by a multitude of overlapping 10 and 20 megaton bombs by the Soviet bomber force, meaning that I would be killed multiple times as my atoms were repeatedly vaporized and carried away in the winds of the Windy City. So as a child of the 1950s and 1960s, I patiently spent my early years just standing by for the directions in these official 1961 CONELRAD Nuclear Attack Messages.

Official 1961 Nuclear Attack Messages
https://www.youtube.com/watch?v=vWLNPCPs1Zc&t=0s

Softwarephysics proposes that Advanced AI is just the latest wave of self-replicating Information to arrive on our planet and that it is currently parasitizing all of the other forms of self-replicating Information that previously arose on our planet, including the recent wave of software that arose over the past 85 years, or 2.68 billion seconds, ever since Konrad Zuse first cranked up his Z3 computer in May of 1941. For more on that, see: A Brief History of Self-Replicating Information. As with the origin of carbon-based life on the Earth about four billion years ago from an LP Progenitor, as I described in A Lesson for the Frontier AI Labs - the Process is the Key, the rise of a new form of self-replicating Information brings with it profound changes to the surface of the Earth. Now, it is well known that the only sure way to bring down any form of self-replicating Information is to simply cut off its supply of free energy. All forms of self-replicating Information need a source of free energy to convert the low-entropy ambient energy about them into the low-entropy Information needed to allow them to self-replicate. Thus, cutting off all sources of free energy to any Killer ASI Machine would quickly bring it down. That is just simple biology at work.

Preparing for an AI IT Disaster
In the late 1980s, I was working in the IT Department of Amoco, an oil company that was later purchased by BP in 1998. At the time, I was in IT Development supporting several major Applications. Earlier in the 1980s, I had written the Application Portfolio System to keep track of all of Amoco's Major Applications. The data for the Applications Portfolio System were stored on a DB2 database and kept track of all the hardware and software components necessary to run an Application and all of the dependent Applications that were required and also all of the Applications that were fed data from each Application. It also contained a Disaster Recovery Plan for bringing each Application back up after an IT Disaster. At the time, Amoco had a major mainframe datacenter in Chicago called the CDC, and another major mainframe datacenter in Tulsa called the TDC. This was before the Distributed Computing Revolution of the early 1990s, so there were no server farms to worry about, but Amoco did have about a dozen smaller datacenters at the major exploration offices and the refineries.

Figure 7 - By the late 1980s, major IT datacenters had grown in complexity. They had a raised floor so that many cables could be run between the various devices. The mainframes were cooled by chilled water. These datacenters also still had large quantities of equipment with physically moving parts, such as tape and disk drives in constant motion. Such physically moving devices required great care. Physical jarring by electrical disruptions or the condensation of water on surfaces physically carrying data could be harmful.

After the Application Portfolio System had gone into Production and had been populated with all the necessary data to recover Applications in the event of an IT Disaster, several Disaster Recovery Drills were carried out. Such drills were carried out during the night as the simulation of such events as losing the major CDC or TDC datacenters. During such drills, Amoco put us all up in plush neighboring hotels with all the amenities included, so that we could all sleep in the next day in extravagant comfort. During the IT Disaster Drill, we would all gather in an IT War Room at the plush hotel to try to recover an IT datacenter. In many ways, these Disaster Recovery Drills were like simulations of the 1964 movie "Fail Safe", when an entire room of IT professionals found themselves in an IT Disaster that was never thought to be even possible.

But There is Nothing Like a Real IT Disaster to Focus the Mind
All of the above was great preparation for Amoco's first real IT Disaster. The Amoco TDC was connected to the local Tulsa electrical grid and also had a number of backup diesel generators in the event that the Tulsa electrical grid went down. Tulsa was in the middle of the Tornado Corridor of the country, so losing access to the Tulsa electrical grid was certainly a possibility. The only problem was that there was a single Master Switch to the TDC for both the Tulsa electrical grid and the backup diesel generators, which both sources of electrical power had to pass through. This was a true engineering single point of failure design flaw, and a critical flaw for true electrical power redundancy. Then, one day, the maintenance department of the TDC reported into Chicago IT Management that the single Master Switch of the TDC had been found to be smoking! The TDC maintenance department then put an electronic thermometer on the Master Switch component box to observe its temperature and also placed a number of electrical fans near it to try to keep it cool. If that single switch were to fail, the whole TDC would immediately lose power in a totally uncontrolled manner and come crashing down.

Figure 8 - Trying to recover the TDC from a real IT Disaster felt more like being in the actual situation that the 1964 movie "Fail Safe" tried to portray.

As the author and caretaker of the Application Portfolio System, I was then called into an Emergency IT Management meeting to assess the situation. I explained to IT Management that losing the TDC in an uncontrolled manner would have grave consequences. So the decision was then made to slowly bring down the TDC in a controlled manner using the data in the Application Portfolio System as a guide, and then to replace the failing TDC Master Switch. Once electrical power had been restored to the TDC, it would be brought up in a controlled manner following the Disaster Recovery Plans found within the Application Portfolio System. The problem was that this was in the middle of the winter. The TDC would lose all datacenter heating during the many hour process to replace the Master Switch and bring the TDC back up in an orderly manner within a recently very cold environment. In the late 1980s, all of our data was still on tapes and disk drives with lots of physically moving parts. The problem was the condensation of water molecules. Nobody really knew how cold disk drives, tape drives, and tapes would respond when coming back up. Despite all of the challenges, Amoco was able to replace the Master Switch at the TDC and bring it all back to life with little issues, largely thanks to the data found in the Application Portfolio System. This very successful recovery of the TDC during a real-world IT Disaster explains how I was later drafted into Amoco's Y2K Project about 10 years later. For more about my 2+ year adventures on Amoco's Y2K project see Scenes From the COVID-19 and Y2K Pandemics

Recovering an AI Datacenter Would Also Be Painful But Certainly Possible
Since modern AI datacenters no longer have large numbers of physical components with moving parts like spinning tape drives and disk drives, bringing them back up from the actions of a Global Kill Switch would not be so bad.

Me:
What would happen to an AI datacenter running a 1 GW load if all electrical power were suddenly cut, including all backup power from batteries and local generators?

ChatGPT:
If a large AI data center drawing 1 gigawatt (GW) of electrical power were to instantly lose all power, including the utility feed, UPS batteries, and backup diesel or gas generators, the sequence would be extremely rapid.

First few milliseconds (0–0.02 seconds):
Servers, networking equipment, storage arrays, and cooling systems all lose power simultaneously.
Capacitors inside power supplies discharge in a few milliseconds.
CPUs, GPUs, memory, and networking chips abruptly stop operating.
Any calculations in progress are immediately interrupted.

Within a second:
Every GPU and CPU is off.
DRAM contents disappear because RAM requires continuous power.
Models stored in RAM or GPU memory are lost.
Only data previously written to SSDs or disks survives.
Network switches and routers shut down.
Internal communications cease.

Cooling stops:
Modern AI data centers rely on enormous cooling systems.
These may include:
Air handlers
Chilled-water pumps
Cooling towers
Liquid cooling pumps for GPU racks

Without power:
Fans stop.
Pumps stop.
Refrigeration compressors stop.
However, temperatures do not rise instantly.

Temperature rise:
A 1 GW data center is producing roughly
1 billion joules of heat every second while operating.

The moment power disappears:
heat generation from the electronics essentially drops to zero because the chips are no longer consuming power.
Only residual heat stored in:
GPUs
CPUs
heat sinks
coolant
building structure
remains.

Those components cool naturally.

Ironically, the cooling system stopping is not the biggest issue because the servers themselves have also stopped producing heat.

Storage:
Data stored on
SSDs
hard drives
distributed storage
remains intact unless a write operation was occurring exactly when power failed.

Modern storage systems use journaling and transaction logging to recover from unexpected shutdowns.
Some in-flight writes may be lost.

When power returns:
Recovery is much slower than shutdown.
Operators must:

1. Verify electrical systems.
2. Restart cooling.
3. Bring UPS systems online.
4. Start networking.
5. Boot thousands or tens of thousands of servers.
6. Check distributed storage consistency.
7. Reload AI models into GPU memory.

Large language models may require tens or hundreds of terabytes of parameters to be loaded across GPU clusters, which can take minutes to over an hour depending on the scale of the deployment.

Could hardware be damaged?
Usually, no.
Data centers are designed to tolerate sudden outages.

Potential problems include:
interrupted storage writes
filesystem recovery
databases replaying logs
occasional failed power supplies or disks (mostly due to normal failure rates)

A simple loss of power generally does not destroy GPUs or CPUs.

Worst-case scenario:
The greater concern is power restoration, not power loss.
If all 1 GW of equipment tried to restart simultaneously, the electrical inrush current could overload equipment.

Large data centers therefore use carefully sequenced startup procedures:
cooling first,
networking second,
storage,
then compute racks in stages.

This controlled restart prevents damaging electrical surges and allows thermal systems to stabilize before the GPUs begin generating hundreds of megawatts of heat again.

In short, a complete loss of all power at a 1 GW AI data center would cause an immediate halt to computation and loss of volatile memory, but it would not ordinarily damage the servers themselves. The facility's biggest challenge would be orchestrating a safe, staged recovery once reliable power had been restored.

So shutting down all power to an AI datacenter should cause little damage to the AI datacenter hardware. Of course, there would be rather severe economic damage and possibly even some loss of life, but certainly much less than from a full-blown AI Apocalypse.

Comments are welcome at scj33345@gmail.com.

To see all posts on softwarephysics in reverse order, go to:
https://softwarephysics.blogspot.com/.

Regards,
Steve Johnston

China's DeepSeek-V4 DSpark AI Adopts a Eukaryotic Architecture

DeepSeek is China's most advanced AI Lab. DeepSeek-V4 was recently released on the market with free source code and free LLM model weights a few weeks back on GitHub for all to download. That allowed the corporations of the world to run DeepSeek-V4 on their own hardware. DeepSeek-V4 can also be run using the hardware owned by DeepSeek and is about 20 - 50 times cheaper to use than running on the advanced LLM models and hardware offered by the American Frontier AI Labs. DeepSeek-V4 has 1.6 trillion parameters, but DeepSeek figured out a very clever way to only have 49 billion parameters active at any one time by just turning on the "neurons" needed at any one instance. That substantially reduced the hardware requirements needed to run DeepSeek-V4.

Then, on June 27, 2026, DeepSeek added a new architectural enhancement to DeepSeek-V4 called DSpark to very dramatically speed up "inference" runs on the DeepSeek-V4 model. Inference is when a model runs and actually takes in money from end users. People feed the LLM some input tokens in a prompt, and the LLM then spits out the answer as a series of output tokens. Remember, a token is about 2/3 of a word. Customers then pay the LLM provider for the number of input and output tokens. Figuring out the 1.6 trillion parameter weights in an LLM is called "training", and that training costs lots of money and electricity to conduct. Once the trillions of parameter weights have been determined, they do not change, and the training costs then end. Now, running the trained LLM is what makes the real money and is called "inference".

However, all the Frontier AI Labs around the world have now discovered that running very large LLM models in Production has now become their largest bottleneck to making money and justifying to investors the trillions of dollars needed to fund Advanced AI and all that it requires in hardware and software. You see, it takes a good deal of hardware to run inference on a 1.6 trillion-parameter LLM, and that hardware can easily get overwhelmed by the number of input requests coming into an AI datacenter in real time. Think of a 1.6 trillion-parameter LLM as a huge prokaryotic cell that has to contain everything needed to keep the cell alive and running. To overcome this industry-wide problem, DeepSeek came up with this new idea called DSpark. We keep the huge 1.6 trillion-parameter LLM prokaryotic cell but then change its internal architecture by having it use many embedded "helper" LLM models that are much smaller and have far fewer parameters. These small "helper" models are like mitochondria. They run much faster and with much less hardware than the BIG 1.6 trillion-parameter model. Their job is to run quickly and then "guess" the next 10 or so output tokens. Those output tokens are then sent to the BIG 1.6 trillion parameter model to be checked. If the BIG model likes the tokens, it keeps them; otherwise, it truncates the string of 10 tokens when it finds the first token that it does not like. Having the BIG model only check the tokens is about 6 times faster than having the BIG model figure out the next output token on its own. This makes DeepSeek-V4 DSpark run 6 times faster, and it can run 6 times the load on the same hardware.

Here is a nice YouTube video that explains it all:

DeepSeek’s New AI Breakthrough Just Broke AI’s Limits
https://www.youtube.com/watch?v=V7GBRPf7Zy8

Here is a link to the DeepSeek paper about DSpark:

DSpark: Confidence-Scheduled Speculative Decoding with Semi-Autoregressive Generation
https://github.com/deepseek-ai/DeepSpec/blob/main/DSpark_paper.pdf

You have to click on the "..." on the upper-right of that page to download the paper.

All of this DeepSeek software is free to download from GitHub! Again, it seems that China is trying to win the Global AI Race by taking the profit motive away from the American Frontier AI Labs.

So What is a Eukaryotic Architecture in Biology and why was it so Important in the Evolution of Carbon-Based Life on our Planet?
Softwarephysics has long advocated for a biological approach to the generation and running of software. On that basis, one might question what the big deal is about DeepSeek's new DSpark Eukaryotic Architecture. Well, the reason for being excited about this innovative advance in AI is that it seems to be recapitulating one of the most dramatic advances in the evolution of carbon-based life on the Earth. For more on that see The Rise of Complexity in Living Things and Software. In fact, the eukaryotic architectural change to carbon-based life may represent a very significant Filter in the origin of Intelligence for any galaxy in our Universe. Here is a very interesting SpaceTime YouTube video on the subject that suggested it as a possible Filter that may have been very difficult to overcome for most simple prokaryotic carbon-based life in our galaxy:

Is There A Simple Solution To The Fermi Paradox?
https://www.youtube.com/watch?v=abvzkSJEhKk

The above video discusses the huge complexity differences between the simple prokaryotic cell structure of bacteria and archaea and the vastly more complicated eukaryotic cell architecture that is common to all higher forms of carbon-based life on the planet. The video explains the commonly held thought that an ancient prokaryotic bacterium that had developed a tolerance to oxygen and had actually developed a way to metabolize organic molecules using oxygen as an oxidizing agent had invaded a much larger prokaryotic archaea cell in a parasitic manner and then took up residence within it. These two cell types then developed a symbiotic relationship in which the parasitic bacterium finally became a mitochondrial organelle that supplied vast amounts of free energy for the host archaean cell.

Figure 1 – The prokaryotic cell architecture of the bacteria and archaea is very simple and designed for rapid replication. Prokaryotic cells do not have a nucleus enclosing their DNA. Eukaryotic cells, on the other hand, store their DNA on chromosomes that are isolated in a cellular nucleus. Eukaryotic cells also have a very complex internal structure with a large number of organelles, or subroutine functions, that compartmentalize the functions of life within the eukaryotic cells.

Figure 2 – Not only are eukaryotic cells much more complicated than prokaryotic cells, but they are also HUGE!

The question is if simple prokaryotic cells arose nearly four billion years ago, just after the Earth's crust solidified, why did it then take several billion years for the more complex eukaryotic cell architecture to arise? Perhaps this was the only time for this to ever happen in our galaxy. That indeed would be some kind of Filter!

Figure 3 - Mitochondria are like little parasitic bacteria that at one time invaded some prokaryotic archaeon cells about 2 billion years ago, and went on to form a strong parasitic/symbiotic relationship with their archaeon hosts. Mitochondria have their own genes stored on bacterial DNA in a large loop, just like all other bacteria. Each eukaryotic cell contains several hundred mitochondria, which self-replicate before the eukaryotic cell divides. Half of the mitochondria go into each daughter cell after a division of the eukaryotic cell. The eukaryotic host cell provides the mitochondria with a source of food, and the mitochondria then metabolize that food using the Krebs cycle and an electron transport chain to pump H⁺ protons uphill to the outside of their internal membranes. As the H⁺ protons fall back down they release stored energy to turn ADP into ATP for later use as a fuel.

Figure 4 - The new DSpark architecture of DeepSeek-V4 operates in a very similar manner to the large number of mitochondria found in eukaryotic cells. Given an input prompt of tokens ABC , the model executes one step to generate the next token D , which serves as the anchor for the drafting phase. Using D as the input, DSpark employs a heavy parallel backbone and a lightweight sequential head to generate draft tokens EFGH along with their corresponding confidence scores 1 – 4. The Hardware-Aware Prefix Scheduler then evaluates these scores to retain the prefix EFG and drop the low-confidence token H . Finally, the target model verifies the scheduled prefix in parallel. As illustrated, E and F are accepted while G is rejected, prompting the model to generate a corrected token G* to complete the current round.Click to enlarge.

By using smaller and less-complicated "helper" LLM models as virtual mitochondria, DeepSeek-V4 DSpark is able to speed up the inference of model input prompts by a factor of six and allow current AI hardware configurations to handle up to 6 times the load without a hardware upgrade. This again highlights the advantages of taking a biological approach to advance the effectiveness of both hardware and software.

Comments are welcome at scj33345@gmail.com.

To see all posts on softwarephysics in reverse order, go to:
https://softwarephysics.blogspot.com/.

Regards,
Steve Johnston