Into this Universe, and why not knowing,
Nor whence, like Water willy-nilly flowing:
And out of it, as Wind along the Waste,
I know not whither, willy-nilly blowing
Rubaiyat of Omar Khayyam
Edward Fitzgerald

Some years ago I wrote an essay, Immortal Imperative; it dealt with the future of humanity, if that is to be anything significant, and for some aspects of this article, you might need to read that original. The scope of that topic is beyond my perspective here, but the idea was to investigate alternatives and essentials for us to have any future worth contributing to.

My theme was not nearly comprehensive, and additional ideas proliferated insistently, so when I encountered Google Gemini, we discussed matters collegially, and began to explore various concepts that had been overlooked or underemphasised in the previous essay. The outcome still is nothing like comprehensive, but readers might find parts of both essays challenging, even if not convincing.

Terms and confusions

...wenn du lange in einen Abgrund blickst, blickt der Abgrund auch in dich hinein

...if you gaze long into an abyss, the abyss will gaze back into you
Friedrich Nietzsche

First, note a few concepts and terms to save the trouble of harking back to the original essay.

Homo ephemerens and Homo futurens are not accepted terms in biology, and I do not define them in formal biological terms. Homo ephemerens simply refers to the Homo sapiens of today and of the last Terasecond or so, plus the next few Gigaseconds, and Homo futurens is the population that succeeds ourselves for many Petaseconds, if we survive.

Beyond Earth’s orbit “year” is ambiguous, so I measure time in seconds, which are after all, SI units of time, for example:

· 1 Megasecond (Ms): ~2 weeks. Relevant to transient, urgent events

· 1 Gigasecond (Gs): ~32 Earth years. About half a current human adult lifetime

· 1 Terasecond (Ts): ~32,000 Earth years. It took Homo ephemerens about 1 Terasecond to emerge from the original hunter‑gatherer status to our current technological level.

· 1 Petasecond (Ps): ~32,000,000 Earth years. The operational time scale of Homo futurens. For comparison, the Dino-killer asteroid impact happened about 2 Petaseconds ago.

Another concept that arises is that of the subjective emotional awareness, what amounts to our sense of self. I have gotten into the habit of calling it “CES”, standing for “Cogito Ergo Sum” simply for lack of a good and compact term for such a common and subtle concept, one that has not yet nearly been characterised in any empirically falsifiable way.

Don’t worry about that term in this article; it only becomes relevant when discriminating between sentient organisms and artefacts that can pass the Turing test; the reason is that we do not yet have any way of detecting a CES, or referring to it unambiguously, or generating one, or working on one in any way that is more effective than conversing with it or drugging it or something like that.

What this is about

Myself when young did eagerly frequent
Doctor and Saint, and heard great Argument
About it and about: but evermore
Came out by the same Door as in I went.
Edward Fitzgerald: Rubaiyat of Omar Khayyam

I have gone into the details of this topic elsewhere, and at length, so I omit justification or explanation; just please do not mistake the items in this essay for the thesis.

After all, if you were to read the conclusions of relativity theory or quantum theory without reading their substance first, you would certainly dismiss them as nonsense. And what I present here is nothing like as abstruse as those are.

· Immortality, in the sense of not having any limit to one’s life expectancy unless you are killed, is biologically possible: given current advances in genetic engineering, it should be attainable for the entire population within a few tens of Gigaseconds after the time of my writing this.

· Homo ephemerens (that is you and me, remember) not only is very mortal, but we doom our nations to mortality too; we and they are temperamentally, mentally, and morally, ill‑equipped to handle immortality; commonly we cannot even deal with our Gigaseconds: when the TV fails, most of us are hard put even to face the proverbial rainy Sunday.

· Our culture too, is poorly equipped to deal with our mortality; if all goes well, we as individuals spend most of our first Gigasecond preparing for our productive lives. We then prepare for retirement during most of our waning years, with seldom more than a third Gigasecond to round them off. We behave at first as if we were immortal, but soon tamely tailor our concerns to fit our objectives, our bottom lines if you like, into a brief senility.

· Even those of us who mean to create a dynasty for their descendants, are whistling in the dark; how many dynasties outlast three generations? And after the first two, how many would have gratified the visions of that first generation? And as for establishing empires, it is an old empire that outlasts two centuries, and usually debased at that.

· This ephemerality mocks our dreams and creations. We talk big, think small and die smaller. About the scattered remnants of our statuary the lone and level sands stretch to the nearest landfill. We waste unprecedented opportunities because they will not pay in our lifetimes, and we fritter away our heritage in rejecting prospects that could have matured in a few Gigaseconds.

· To lend point to bioengineering that could create us immortal, we need the mental capacity to remember what we learn, and the intelligence to process our memories and to implement the schemes and achieve the ideals that we conceive. A brain of Homo ephemerens will not suffice for even 32 Gigaseconds, let alone effective immortality.

· Another point is that to be immortal in a society of mortals would be frustrating in many ways; this applies to both parties; mortals tend to scale their efforts and their expectations according to their temporal limitations of planning and profit. To accommodate long‑term plans and objectives, any immortal would need deeper and longer vision than mortals could manage, so mortals would raise frustrating objections. Resentments would be inevitable and mutual; Homo ephemerens constantly would be aware that Homo futurens could simply wait for opposition to lapse into oblivion, and though that would be true, having to work on that basis would leave the immortals with a constant perception of waste and futility.

· Furthermore, biological immortality is useless in the long term while humanity remains confined to a vulnerable environment such as a single solar system. At any few Megasecond’ notice, a rogue planet could collide with Earth at high speed, leaving the planet literally molten. Or a hydrogen-rich rogue black dwarf could trigger a lethal solar nova. To survive events of that type, an immortal species must establish off-world independence far in advance; a DART‑type of emergency intervention would not suffice. Survival would demand not only immortality and anticipation, but also the skills, commitment, and organisation, to conceive, anticipate, and deal with the events.

· Those are simplistic examples of the imperatives that immortality imposes on conscious immortals. There are more where they came from. Homo futurens, as the inheritor of our species status, will have to unite in more than the individual, more than the population, much as Homo sapiens is not Joe Average, although Joe Average is of the species Homo sapiens. In highly sentient species such as apes, parrots, or elephants — species that establish group heritages — a population comprises the individuals plus its heritage, education, and information.

· The Homo ephemerens population, after our recent Terasecond or so of marking time, will have to produce something new under the sun if we ever are to emerge as Homo futurens. As Humanity, we must displace Darwinistic minutiae from human evolution, which, in spite of the protestations of many authors, we so far have failed to do — Homo sapiens has by no means halted the process of Darwinism in our populations. And yet, we now are on the brink of replacing Darwinism with something that never has existed on Earth in the last hundred Petaseconds or so of life on Earth: Teleological or Volitional Evolution, acquiring the abilities, not only to change in desired directions, but at rates foreign to natural selection. And if we stumble on that brink, humanity is doomed.

Ephemerality, Futurity and Perspective

"Shut up!" he explained.
Ring Lardner

Our past as Homo is a mere tremor in the course of biology on this planet; depending on who does the counting, and with what in mind, that past has extended over something like one to one hundred Terasecond. Our science similarly extends over about ten to about a hundred Gigaseconds: however you count it, that is a blink since the origin of life about a hundred Petaseconds ago.

The implications for our perspective on our past and future are profound. My favourite example is that the lifestyle of Queen Victoria at the time of her accession to the throne, was more like that of Solomon than to her own lifestyle at the time of her death. That is about like comparing the effective progress of two Gigaseconds to the progress, or rather the stasis, of one tenth of a Terasecond — more than fifty times faster.

For some three thousand to thirty thousand years we have been smugly coasting along on the bullyings and illogical assurances of ignorant elders who could not face having youngsters who might question their mysticism and their intellectually paralysing refusal to permit anyone to question, let alone correct their ignorance and disinformation. They betrayed their generations, reserving their choicest invective for anyone evil enough to wonder:

...they are like the deaf adder that stoppeth her ear; Which will not hearken to the voice of charmers, charming never
so wisely. Break their teeth, O God, in their mouth: break out the great teeth of the young lions, O Lord.
Let them melt away as waters which run continually: when he bendeth his bow to shoot his arrows,
let them be as cut in pieces. As a snail which melteth, let every one of them pass away:
like the untimely birth of a woman, that they may not see the sun. The righteous shall rejoice when he seeth the vengeance: he shall wash his feet in the blood of the wicked.
Psalms 58

If we could have escaped such anti‑questioning authoritarian social poison from the first, and thereby grown directly into a systematic scientific attitude, as the Greeks nearly did, competitive and sceptical, but mutually encouraging, we could easily have achieved say, the scientific and technological level of Galileo perhaps 300 Gigaseconds earlier than we actually did, and brought it all up to date in the following half dozen Gigaseconds. In the event we wasted untold resources, lives, and brainpower; and that waste might yet mean the difference between the survival and the extermination of humanity.

In some ways our perspective on the twentieth century has been at least as dizzying, as the acceleration of progress during Victoria’s life, making a mockery of periodic claims that the end of science is in sight. Not only are those claims practically and philosophically untenable, but they are hardly relevant, because what matters more immediately in our context here, is the progress of technology rather than science, and technological progress is not even apparently declining. It also is particularly difficult to project, because technological advances are combinatorial; hardly any one of them arose from the application of just a single particular item, unless you count the hula hoop.

One consequence is that there is more than one way to achieve practically everything, and the adoption of almost any success is an adventitious function of social factors of indefinite complexity.

Ironically, this aspect of this class of advance did not await the intellectual birth of human technology; before there were humans, the fact that there is more than one way to achieve practically every product of natural selection, had been an intrinsic aspect of biological evolution. Some advances superseded each other, for example, as for example, two‑winged vertebrate flight superseded four‑winged. Others developed over and over: several forms of flight, many forms of eye and of ear, to mention just two classes of independent development along different lines. That independence of lines of development has been a reality since the origin of life on the planet. Conversely there always are niches that remain unfilled, but could in principle be filled by many approaches in many ways, and they tend to be occupied in different ways in different places and time depending on adventitious and combinatorial factors. And yet, many a niche goes unfilled. This makes it almost impossible to predict details of future developments with any confidence, let alone reliability, either in evolution or human technology.

And the same applies at more abstract levels such as strategies, either human or evolutionary.

In contemplating our own future, we encounter a precipice: trying to predict the nature of Homo futurens by extrapolation from Homo ephemerens is bound to be wrong in detail, and largely in trend as well. I have stated with confidence why Homo ephemerens is doomed if we do not evolve, and evolve relatively fast and in particular directions, of which longevity, and preferably immortality, is one. However that not only does not enable us to predict the details of our evolution, it does not even predict our direction of evolution better than nebulously, and with no confidence at that.

Furthermore, to an extent unprecedented before the emergence of 20^th century biology, each generation, at a guess, thirty Gigasecond from now on, may be seen as Homo ephemerens relative to its own successors a few generations down its subsequent line, and this relationship will remain consistent for perhaps a Terasecond or so.

By that time, if all goes well, the entire human population would presumably be fully engineered, though generally visibly human. However, as no particular configuration or biology can be defined rationally as uniquely and universally superior in every role, and as biological engineering of immortal humans should by then have matured beyond our reasonable ability to predict, every new human and every mature human should by then be an undoubted Homo futurens by any previous standard, however biologically distinct that individual might be.

Variation within such standards might include anatomy, physiology, reproductive principles, social roles, and serial material bodily alterations. So, for example, an individual might elect to reproduce somatically and asexually for some reason; as long as it met legal criteria for population and social controls, there could be no ethical or logical objection.

Homo futurens and Volitional Evolution

We all love miracles, but if a plan depends on a miracle don’t make it your Plan A
Unattributed

No species before Homo sapiens ever diverged from Darwinian evolution, and our own species has diverged only superficially and stammeringly in a few variations on the principles of artificial selection. In recent Gigaseconds however, molecular biology has achieved the inconceivable, and promises to drag humanity, willingly or violently nillingly, into a new age in which we can endow our offspring with clean slates and with progressively cleaner inscriptions than ever in the past.

More than Enough

The sum of the intelligence on the planet is a constant;
the population is growing.
Unattributed

Firstly, before discussing genetic engineering, we must face population control. Even in the face of wars, famine, and pestilence, human reproduction has quadrupled the global population within a human lifetime: about three Gigaseconds; this simply cannot be sustained, and if overpopulation is not controlled, the consequences will be beyond the imagination of either laity or politicians. Without repeating details here, population reduction and improvement can be implemented by licensing every one person to have one child. This might sound as though it would maintain the population, but in practice it would reduce the population relatively rapidly, while permitting rapid improvement in population composition, fitness, and health.

The child of each parent would be freely medically and educationally subsidised, but parents of every reasonably avoidable extra child would be subject to punitive measures. Part of the free service would be optional in vitro fertilisation to ensure one child, the best possible child from those parents, as is the case commonly nowadays for in vitro fertilisation anyway. Engineering by more advanced measures, also free of cost, would be optional, on request by the parents, and approval by the reproductive professional charged with selection of the blastocyst.

This can be done voluntarily and increasingly creatively, not by compulsion, but by letting parents choose or tweak their offspring’s major genetic attributes. Most parents would like their children to be as outstanding as possible, and those that do not, we need not trouble ourselves with; traditional Darwinism will rapidly render their descendants dysfunctional in a world in which they can neither contribute nor profit.

Letting parents exercise their own choice of how to propagate should work within families and local populations. Granted, no matter how kindly and gently the genetic engineering were introduced, it would be guaranteed to encounter resistance, especially at first. There will be misunderstandings, gossip, political point‑scoring, religious rabble‑rousing — all the usual stuff. However, there are mitigating factors too, and in the face of firm control, they must prevail.

Although introduction of rational measures cannot be abrupt, positive action is necessary to reduce the current enormities of injustice and cruelty. Natural confusion and laissez faire got us into this in the first place with such things as anti‑vaccination hysteria; to retrieve a future for our descendants, we must wake up to realities and deal with them.

It would be criminal to assert the right of parents to inflict preventable genetic tragedies upon a new generation — the likes of Duchenne, thalassaemia, or Huntington’s —. That would not be “liberty”, but state-sanctioned child abuse.

More of what?

The race is not always to the swift, nor the battle to the strong —
but that's the way to bet.
Damon Runyon

Homo futurens would have to grow into a new human ecology, and not only personally, but as a population. It is likely that the very concept of nations will erode rapidly before that stage of social progress matures. The distinction between Homo ephemerens and Homo futurens would be broad and blurred, but would not extend into the indefinite future; Homo ephemerens cannot forever survive the stranglehold of infinity. What the populations of Homo futurens will look like after one Terasecond, I cannot say, but I am sure that developments will exceed all the changes in human evolution during the previous ten Teraseconds. For one thing, long before that happened, our technology could achieve biological immortality, created or encountered fellow species, developed Solar System colonisation, and found or created more that we are not yet in a position to predict.

A new form of sociotechnical critical mass is inevitable, in which the more we achieve, the faster we develop. Eventually our humanity and our biology and technology must integrate, bidding farewell, not only to nationalism, but to racism, speciesism, interstellar alienism, and even technologism.

The Past is a Poor Future

Somewhere ages and ages hence:
Two roads diverged in a wood, and I —
I took the one less traveled by,
And that has made all the difference.
Robert Frost

So then: we Homo ephemerens cannot survive into any worthwhile future on any worthwhile time scale; we cannot rely on Darwinian selection to lead us into survivability, and we certainly cannot hope to defy evolution in general; we need to re‑asses the consequences of our past and adjust our approach to the future, so Darwinism must yield to teleology.

Past themes in fantasy and science fiction have touched on teleological evolution, but rarely on any basis of practicality, and one merciless principle in real life is that if you will repeat past blunders, the future will get worse; so if we are to maintain any pretence of human dignity, we must be careful of how we choose what to change, and of how to change it; and that includes choosing whether actually to change or not change at all.

Perceptive population control will be our first and fastest tool, and must remain in force in Earth indefinitely. Once we are in any sort of healthy population equilibrium, we must look to the future and plan a more radical course, but I begin with the initially more pedestrian, more practical course. What happens elsewhere will be for the local population to decide. One hopes that there will be early adoption of effective and compassionate customs of population planning and control, but it is beyond our ability to determine that here.

Consider family‑size constraint supplemented by in vitro fertilisation, and imagine the mentality of would-be parents who would avoidably inflict genetic imperfections and suffering on their own children — that would be unethical by any rational standards, especially if misrepresented as a parental right.

We speak of parental rights, but what about the children’s rights? Or the rights of the germ line? Such smugly futile cruelty would be borne by the innocent children, not the guilty parents. Whether to class it as a felony presumably will be bitterly disputed, but such behaviour certainly would disfavour the fitness of one’s own germ line, because the children could expect more unnecessary problems statistically, whether intellectually or physiologically, when in competition with selected products of in vitro fertilisation.

This policy of routine selection of acceptable blastocysts would reduce entropic waste of health and talent, and it would do so rapidly in terms of functional adaptation of the population to the requirements of individual health and compassion. Within a few generations it would abolish inviable genetic attributes in populations, in which, by that time, individuals would not look much different from generations of today, except for generally better health, strength, skills, longevity, and, accordingly, aptitudes. The statistical effect however, would be dramatic. Drastically harmful genes, whether nuclear or mitochondrial, whether quantitative or qualitative, whether recessive or dominant, monogenic or polygenic, would be reduced to vanishing infrequency. Ideally, the occurrence of such harmful traits would approach the rarity of accident and mutation.

For simple genetic reasons, traditional artificial selection by elimination of unacceptable phenotypes after birth, or even during gestation, could never achieve such results at any reasonable rate, if ever. This would remain true whether harmful alleles or other conditions were troublesomely recessive, or otherwise latent or patent.

And, supposing you did do your selection without examination of the blastocyst? And now you have to decide what to do with the unacceptable baby? Killing it at this point would in most countries, count as unambiguous murder. In ethical terms, sterilising it would hardly be better. And yet, you have increased the population by one liability. How can you ethically claim the lenience of the law?

In contrast, disposal of rejected blastocysts would hardly be more problematic than disposal of unused sperms or ova, whether selection had been by clinical inspection or by throwing dice; after all, one cannot implant or keep every zygote or blastocyst. Disposal after implantation also would be more wasteful and traumatic for all concerned. From the point of view of compassion, selection of blastocysts for implantation during in vitro fertilisation, before the young had developed any capacity for personality or pain, would be far less brutal whether by sterilisation or elimination of babies.

Denial of the facts cannot rationally be tolerated. Clinical toleration, uncontrolled populations, and obstructive legislation, cannot serve in any long term wherever humanity is to survive. Building a viable population will demand social, philosophical, and practical reorientation of our education and our society if we are to achieve fundamental objectives of compassion, cooperation, and health. Humanity of today, whenever and wherever we exceed our needs for food and similar basic requirements, is oriented towards novelty and expansion, often in pathetically trivial forms. It is said that Alexander the Great mourned the lack of new worlds to conquer, but be that as it may, the perceived lack of new worlds certainly has been a common theme among youthsome adventurers since Victorian days at least.

It cannot last. There is a current desire to expand frontiers into outer space, in futilely artificial directions such as the Moon and Mars, but those are the products of ignorance as much as of venturesomeness. For one thing, many such ventures fall into the natural sphere of AGI as it already is maturing. As trends towards Homo futurens become established in the population, we should see political and social improvement, as universal improvements in public intelligence render simplistic propaganda less effective, but there is much beyond that to deal with as well.

Within a few generations after establishing the in vitro fertilisation practice, that purely selective approach would progress to active genetic engineering throughout the species. That would require no legislative compulsion; Homo futurens will no longer be satisfied with statistical good health and good abilities, but will work towards new, additional features. Which these would be, we cannot be sure of in general, but we might predict for example that within several Gigaseconds children will be born with no need of various vitamins in their diet, or that they will be capable of colour vision in the near infrared and near ultraviolet, or that their longevity would begin to extend to a several extra Gigaseconds without mental senescence.

And such policies would be for the parents of their day to decide, not for our generation to dictate.

Many concepts that we now take for granted would change or fall away, nor is it clear which changes would be possible in responsible, educated, competent adults or populations either before or after conception. Physiological and anatomical changes are enormously, combinatorially, complex, especially if they have to be genetically controlled, but it is conceivable that many classes of functional adaptation or experimentation could become routine. Such adjustments could be analogous to our current practices of tattooing or body building, though one assumes that they would be more rational and effective.

By whatever means, genetic or not, we would expect Homo futurens to pursue radical classes of objectives and concepts. Immortal longevity would be one of the most fundamental: not just a few centuries of longevity, but for as long as the person chooses. Its importance is fundamental because without it the entire set of associated objectives would lose their very significance. This implies drastic physical and physiological adaptations, of which we shall examine some later.

AI in the Evolution of Homo futurens

For progress there is no cure.
Any attempt to find automatically safe channels
for the present explosive variety of progress must lead to frustration.
The only safety possible is relative, and it lies in an intelligent exercise of day-to-day judgment.
John von Neumann

Many human activities, some of them currently regarded as professional, increasingly bid fair to being displaced wholly or partly by AI or AGI functions. There is nothing new about that; wide varieties of established ways of earning a living have vanished throughout history, raging from flint napping to card punching. And though some were admirable in their sophistication and vital in their day, not many still are mourned. The contexts within which, and the degree to which, fear of obsolescence is justified, are hard to predict, but there is no obvious end to such things.

What is sure is that the most drastic social effects of the introduction of AI in any form will be hardest upon the least advanced members of new generations. It accordingly behooves caring parents to ensure the best possible education, practical and abstract, for their children, and that authorities should supply that education as freely and effectively as may be.

As for the AI, bear in mind the major classes: Function‑dedicated AI (FDAI), Artificial General Intelligence (AGI), and what I shall call AI‑CES. They have no end of implications in many contexts, but I think we could distinguish those three for practical purposes.

First, think of humanity on this planet: if we were limited to Earth, that fact alone would condemn us to extinction, but let us assume a healthy planetary population with a healthy open‑ended relationship with off‑planet populations. We would need tools, machines, and various classes of servants, consultants, colleagues and companions.

Think of the contexts in the following sections.

Function‑dedicated AI

To live only for some future goal is shallow.
It's the sides of the mountain that sustain life, not the top.
Robert M. Pirsig

In the short term, FDAI would be adequate for most circumscribed purposes. We already see intimations of it at the basic level, variously independent, or as control components of systems or devices. In such roles FDAI could drive our vehicles in well-defined circumstances, run our factory lines, build our buildings according to specified plans on defined sites, dig our mines, manage our tills, mind our pets, in fact do a whole lot of things that we hardly ever thought of as suited to AI at all. The fact is that a lot of the things we do, though far more complex than we ever had realised, are not so complex that we cannot automate them. There is a lot that FDAI can do beyond the functions of the robots that we see on assembly lines and in warehouses.

There is more to this than we might think. For example, where a function requires 24‑hour, 365‑day alertness, patience, strength, tools, immunity to disease, pollution, or toxicity, humans not only could not compete, but would not want to; much of such work is unpleasant, even dangerous, and practically slavery, none of which FDAI need be designed to take exception to, but which does not inspire reliability or enthusiasm in humans.

AIs also would be more expendable than humans, in some contexts even more expendable than valued animals, though one would avoid this aspect wherever practical — AI not only is expensive, but deserves respect, just as we treat our cars with respect, even when we can afford a new car.

Expendability of AI deviceswould be important in situations where there are unavoidable risks, such as in traffic or down mines or flying deliveries. They also have advantages for many functions in space, in which their independence of physiological requirements or disposal, and insensitivity to radiation, are valuable; one can accommodate a motherboard and power supply in less space than the couch for a human with his food, toilet, windows, oxygen and more. Their immunity to fatigue in 24-hour jobs, or boredom when alertness must be maintained, could be valuable in dedicated functions like scavenging junk such as dead satellites cluttering orbit space, and in delivering or retrieving space craft and supplies where required, even out to Kuiper belt distances; and they could undertake functions in space such as managing communication relay webs, observation sites, and navigation beacons.

Similarly, they could have implications for managing some marine tasks, such as submarine cargo vessels, and flotsam disposal.

The intelligence of FDAI devices could be supplied via hardware and software units using sensory and control units with standard interfaces. Such items could be swapped out for updating and maintenance wherever required. They and their components could be swapped out with duplicates, and their data could be retrieved reliably and in detail at short notice without special training and debriefing. With assistance from AGI in functions requiring management and judgement, FDAI could replace large ranges of functions that we currently see as strictly human, such as running factories, buildings, large vehicles such as ships, educational institutions, agricultural and natural resource management such as mining, and various forms of traffic control. Teams of dedicated FDAI functions could combine to deal with matters such as emergency response or security.

The limitations are hard to define.

Artificial General Intelligence

A ship in harbor is safe —
but that is not what ships are for.
John A. Shedd

From several points of view one could regard AGI as more, or less, “intelligent” than FDAI, but the problem is that, as just noted, there is no simple limit in principle to how complex a function or a dedicated task can be, nor how stable its aspects and demands can be.

Similarly, your AGI unit could in principle be fairly simple, and be implemented in fairly straightforward hardware, though it generally must require considerable data and computing resources. It also might have special-purpose processors, sensory devices, sophisticated software, and broad‑context libraries and communication facilities, in comparison to FDAIs. The difference, such as there is a categorical difference at all, lies largely in the application and the data library at its disposal. The library is not just a collection of facts or parameters; an AGI must have structural models, typically language models, in which patterns of relationships may be stored and may be recombined and used, not necessarily in any particular way twice: patterns that had never been used before may emerge or be reused later in new ways.

We do not yet know all the ways in which living brains work, but we do know that animals with brains the size of small pinheads can behave in ways that startlingly resemble some ways in which human brains work, including showing apparent emotion and intelligence, and it is very difficult to say in which ways some are more like FDAIs or like AGIs. It seems unlikely however, that they work on anything much like a Large Language Model; probably more like a neural network.

But the most essential difference, is that ideally AGIs can work creatively in dealing with problems in any field, whether familiar or not, while an FDAI is at least in some respects specialised to work with its dedicated function.

A shocking aspect of human cooperation and authority is the degree to which different services will conflict small‑mindedly on vital matters even when they put fellow members or their countries at risk. A long‑standing joke is that when branches or ranks of armed forces, or of unions, that notionally should be cooperating, they will go out of their way to frustrate, and even betray, each other, even in wartime. It should be natural for AI in general to avoid such obscenities, not because of personal taste, but because of simple built‑in logic.

Another aspect of AI control, wherever there is a need for pioneering in which it would be unrealistic or expensive to send human staff, is that FDAI and AGI units could serve for indefinite periods or at levels of risk unacceptable to humans. For example, in some ways, functions that resemble some of those in deep space, could be established in deep sea situations such as mining, prospecting, communication and navigation relay webs. If it became uneconomic to rescue such a unit, it could be instructed to shut down or enter a disposal procedure once its usable stored information had been saved as required.

In distant installations in space or on alien planets, once such AI units had prepared the necessary facilities, the necessary Homo futurens reproductive units could be introduced, probably in the form of frozen or diapausing larval forms if the distance is interstellar. It is not that a fully fledged Homo futurens could not survive such a period of Teraseconds, nor even that such a human could not keep busy and entertained (who among us could imagine what could occupy such a brain for so long?) but that it would be logistically far less demanding to carry frozen larval stages than a company of active adults for Teraseconds on end. Larval stages designed specifically for freezing or similarly functional diapause would require less elaborate support than gestation of reconstituted frozen ova.

What would keep the Homo futurens staff busy in general, given that AI would by then be able to undertake almost any utility functions we could foresee, and a lot besides, I leave to your imagination. Remember that the brain and longevity of Homo futurens would enable them to undertake projects that would make no sense, let alone profit, for Homo ephemerens: mining Mercury, building space colonies, exploiting Kuiper and Oort resources. Such things would be matters of incentive rather than practicality; they would make no sense in terms of current human economics, but the incentives of Homo futurens must be directed rather at the next Teraseconds and Petaseconds at least, than at tomorrow’s, or next year’s bottom line.

It would be analogous to our twenty‑first century digging up of low‑grade iron ore, useless to pre‑industrial people, or planting of timber seedlings that would be of no use to anyone for forty years; that sort of thing is economically rewarding to us, but hunter‑gatherers could make no sense of them; in forty years they simply might not be alive; and they could not make much use of heavy timber anyway, and they could use only limited amounts of ochre. For them such activity would be suicidally stupid.

The AGIs could manage the evaluation of asteroids and other bodies in the solar system, either to identify profitable bodies, or to plan the deflection of threats as described at this link: protection from asteroid collisions.

AI Vanguard and Rearguard

Fools ignore complexity. Pragmatists suffer it. Some can avoid it.
Geniuses remove it.
Alan Perlis

For at least a Terasecond, humans have been making animals, machines, and slaves do their work, but human bosses generally have been present. This cannot last: life, environment, and objectives grow too large in scale, too complex in intent, too demanding in stamina, for us to manage on such a basis.

So far alternatives to personal human exertion have been few and superficial, so much so, that the Flash Gordon syndrome has been practically universal in our preconceptions of space activities: the assumption has been that a space ship must be run by heroic humans, just as ocean‑going vessels always had to be. But the very emergence of concepts such as of information theory, automation, finity of resources, and technological feasibility, demand changes in principles and outlook. Undreamt‑of challenges emerge: constraints of time, material resources, technology, and space. Those challenges predict the emergence of Homo futurens — or oblivion.

However, Homo futurens is only one current in an ocean of implications. Any viable future must accommodate advances in biology, technology, and possibly also unanticipated alien participants. If we do not prepare we shall deserve a future in which evolved rats and cockroaches sneer at our fossil traces.

Our futures include roles for AI in undersea craft, mining infrastructure, spacecraft, and more, roles in which humans are too costly in terms of food, environment, instrumentation, and accomodation. Still, omitting troublesome questions of economy, there commonly are practical reasons for refusing to include live crew members; consider the following:

Undersea craft and engineering tasks should be practically all AI‑executed. Undersea work would in many ways be a good prototyping environment for management and technology in space; challenges of control, navigation, and communication relay networks and their protocols, match the two environments in essence. Difficulties of long-distance undersea signalling and latency should present good models for long‑distance communication networks in space. Undersea relay networks should extend from North‑polar to South sub‑polar stations under the ice, through the tropics, from the deeps to pelagic surface vessels, taking in seismic and volcanic phenomena, conservation, and energy storage engineering.

Particularly obvious examples for future AI requirements include outer space initiatives. Unmanned spacecraft need not allow for food, medicine, oxygen, and protection from various classes of stress or radiation. At present, not counting tragic and unnecessary losses of particularly valuable lives, the only justified investment in manned space stations and related facilities is not so much for scientific reasons, as because we still have so much to learn about space skills, physiology, and engineering, not to mention human motivation and publicity to motivate an increasingly blasé and cynical population.

It would be far better, beyond our current navel‑gazing, to staff our expeditions with increasingly versatile, durable, powerful, AI‑controlled equipment; units that can undertake entire projects of exploration, construction, prospecting, and pioneering.

We still need progress in basic AI, let alone physical and general technology, but we now have a foot in the door, so to speak. Recent advances have been, and remain, stunning, even implausible, but anyone who thinks that LLMs or human brains are what it all comes down to, has a great deal to catch up with.

Lunar and Arean AI expeditions would be good training enterprises for AI technology, rather than militant human gesturing, and so would the launching of astronometric, observation, and communication craft, to be followed by intensive engineering, scientific, and pioneering work with asteroids. After that our craft could go to moons of gas‑giant planets. Still later we could work in the Kuiper belt and the Oort cloud. None of those would benefit from intimate human presence commensurately with the concomitant costs, especially not in the early stages, and by the time that we find valid roles for humans in space or in submarine works, our AI engineering expertise, equipment, and infrastructure should in turn have expanded out of recognition beyond current abilities.

In particular we should be able to communicate via powerful relay webs in both environments; such webs would have latency similar to direct communication: seconds to hours, but relays would be able to convey large volumes of reliable, precise information accumulated from various aspects and precise coordinates. Direct communication overlong distances, as for example with the Pioneer spacecraft, is slow, of small volume, and demands high redundancy.

With such unmanned craft instead of human crews, our investment should be a small fraction of what would have been necessary, had the expeditions been human‑staffed. And the yield in information from AI technology should be many times larger than from submarine or space‑based industry that depends on human crews.

For expeditions further out in space, our pioneering logistics would need to be on altogether more sophisticated levels. There would need to be completely different perspectives for us to be able to undertake our first faltering steps into interstellar civilisation. To achieve those, we need to re-equip our imaginations; even our fantasy and science fiction tend to be 19^th century sea-going adventures tarted up to accommodate extra‑Solar realities; unavoidable unrealities get propped up with imaginary and inconsistent means of superluminal travel and communication.

The realities of interstellar work are otherwise; they demand different approaches to speeds, hazards, energy supplies and infrastructures, irrespective of whether we encounter alien life of any form at all, let alone technological populations.

Considerations would include identifying, obtaining, and exploiting modes and materials for interstellar travel. It is not yet clear that we ever will be able to travel routinely at anything like relativistic speeds (say 0.1c) but even those would limit us to Gigasecond or Terasecond interstellar journeys for quite close stars.

This has many implications. To be able to contemplate functional interstellar activity and infrastructure, we already would have needed elaborate communication and navigation relay networks within our solar system and out into the Oort cloud. Beyond the cloud we should have to organise elaborate expeditions, with redundant communication relay stations installed on the way out: a sort of interstellar Internet to match the much denser Internet within the Solar system.

The presence and structure of that Internet would not remove the latency limitation, and in fact the Internet communication conventions would have to be elaborated to accommodate the latencies and the populations of IDs and relays, but the comparatively short arcs between relay units, as opposed to between stars, would permit high volumes and reliability, and the large connectivity would improve redundancy. Longer‑distance objectives would not be of direct interest to Homo ephemerens, though they should be vital to Homo futurens, but beyond the Oort cloud new principles and conventions would emerge.

Even for such local enterprises however, and even after the emergence of Homo futurens, AIs would be the almost universally logical commanders, pilots or crew of every off‑Earth craft. That certainly would be the case for either production, inspection, and exploration craft, or one‑way craft. Live crews or passengers would be limited to special purposes or expeditions, and commonly in dormant forms, pending arrival for colonisation.

Spacecraft Vary Wildly

Systems run best when designed to run downhill.
John Gall

I say little about the designs of spacecraft, except that the enormous armoured Star Wars and most other space opera vessels are more like fairy tale themes than anything interesting.

Most really long‑distance spacecraft will be custom designed, even when fleet designs are based on an existing model, and hardly any will be intended for repeated use in the way we re‑use terrestrial or marine vessels on Earth. The main exceptions would be when we have fixed sources of materials such as when loading mined metals that have to be shipped, say from Mercury or the Kuiper belt to sites of consumption or transfer, or that are cycled on elliptical orbits from periapsis to apoapsis, for solar exposure to points of collection for consumption.

For the speculatively foreseeable future, interstellar vessels are likely to be carrying large lumps of more or less inert reaction mass in front. Let us call them fender masses. Their prime function is as reaction mass for the presumably nuclear propulsion engines driving the craft on interstellar or Oort cloud courses. The peak velocities would be very high, and this implies that collision with space dust, that otherwise could be deadly to the craft, could easily be absorbed, while the main craft travelled through clean swept space in the wake of the fender.

We need not worry about really large collisions because, though they would be extremely rare, anything that could break through the fender would smash the craft, probably largely by impacts with the fragments of the fender.

Other functions for the mass would be for fine tuning of the trajectory of slingshots, by swinging the mass about to shift the craft’s centre of mass for tidal effects in directing interstellar craft for precision navigation.

Probes, Flotillas, and Fleets

Many are stubborn in pursuit of the path they have chosen,
few in pursuit of the goal.
Friedrich Nietzsche

As far as I can predict, Humanity will always have some class of activity equivalent to economy, budgeting, and finance, though the relationships between the components need not necessarily be rigidly hierarchical, nor narrowly financial. There always will be conflict for resources of one sort or another in one way or another. A largely stigmergic approach is likely to be the most viable for any management on any scale of cooperative teamwork that requires initiative beyond micromanagement, let alone distributed management over communication latencies of Teraseconds to Petaseconds or perhaps longer. It would be bad enough with Megasecond latencies, partly because any authority extended over such distances is likely to be most suitably managed by stigmergic approaches anyway, with local actors continuing with their initiatives in anticipation of material needs of remote partners.

Micromanagement over large enterprises encounters combinatorial problems anyway, and its successes tend to be small in comparison to the difficulties that it entails.

Stigmergy as a class of teamwork, occurs in nature in various forms, such as in social insects building dwellings, sociable weavers building nests, or dam‑building by beavers, and in such cases there always is an element of innate control of the behaviour of individual. In insects the behaviour is elementary, apparently practically mindless, but the concept occurs widely in various forms and degrees, and its essence is progressive behaviour where latency of communication or similar obstacles modify or interfere with micromanagement of the teamwork of multiple partners.

It is interesting to note that stigmergy emerges similarly from communication obstacles, to lack of intelligence in the effectors on the spot — compare the termites mindlessly starting their new constructions with randomly placed blobs of mud, with the pioneering man‑on‑the‑spot who builds his local fort in a position, that, though tenable, is not where the home manager would have wished to establish his remote seat of government of the new colony. Both classes of dislocation have similarly suboptimal results and paradoxically, similar strengths.

Loose systems last longer and function better.
John Gall

When latency entailed by distance is one of the obstacles, stigmergy is hardly realistically avoidable. However, we should bear in mind that our AIs at any level, no matter how independently they have been trained to think or attack problems, will be forearmed with advanced information in their respective fields. A prospector AGI for example, although with wide general reference resources would be equipped in advance to be able offhand to recognise thousands of minerals and variants of pedology, and to associate them with observed or deduced circumstances, and would be able to evaluate the probability and value of each observation, and the implications for prospecting or threats. In these respects the AGI would resemble a broadly educated human geologist who might have incidental interests in biology, literature, and mechanics. Similarly, an AGI‑CES qualified in biology and astrobiology might do better than most humans on encountering anything from an alien prokaryote to a giant arthropodous pelagic plankton filter‑feeder. Or alien witch‑hunters, or a civilised alien culture.

In short, the challenges of stigmergy should not be great; in many cases there is no formal central management, and in other cases, if there are management authorities, with remote agents, and the agents should understand each other well enough to make allowances for each other’s circumstances, an enterprise may work, commonly not ideally efficiently, by nonetheless effectively. This is implicit in clichés of pioneer colonisation on Earth, such as: “the man on the spot”.

Consider two major classes of stigmergy: asymmetric and symmetric. The former would be where we have a management centre emitting control signals to peripheral units under not much CES‑centred control, and generally no human crew or passengers. Commonly everything could just be AGI‑controlled.

An asymmetric example would be an asteroid mining or colonising unit commanded by instructions from Earth, but not micro‑managed. In the words of Carveth Read:

It is better to be vaguely right than exactly wrong.

A symmetric example might be a live colony in its early stages on say, an Earth‑like planet in the Alpha Centauri star system. It might already have a population of Homo futurens running the colony with cooperation from AGIs and the machinery that they control. It would be too far from Earth for direct control, given a latency of hundreds of Megaseconds.

The result would imply as many effectively independent command centres as necessary, in presumably constructive mutual goodwill.

Probes

Because strait is the gate, and narrow is the way, which leadeth unto life,
and few there be that find it
Matthew 7:14.

Probes would be exploratory or preliminary craft, such as those inspecting target areas, and craft establishing installations for relays, beacons, or observation stations, whether of optical or other radiation, gravitation, or any other desirable data. Or similar craft, though not exactly probes, might be maintenance or supply craft for such stations. The stations would be comparatively densely spaced throughout the solar system, all the way out to the Kuiper belt, but beyond Pluto and into the Kuiper and Oort regions, they would be stationed for special requirements only — at first anyway; if particularly rich pickings were discovered, denser populations could be established later, probably as flotillas.

Whatever the nature of the craft or its function, a probe would not generally carry any live passengers at all, unless it were on a seeding mission, say carrying microbes to establish them on a viable planet. Special purpose probes could be carried along with flotillas and fleets, to be shed along the way as relay stations, beacons and the like. It would be possible in principle to let the probe accelerate the launching craft by accelerating the probe backward, whether with a rail gun or with explosive fuel; it would be pointless to drop a beacon or relay that followed the delivery vehicle instead of staying more or less put.

Flotillas

We travel not for trafficking alone:
By hotter winds our fiery hearts are fanned:
For lust of knowing what should not be known
We make the golden journey to Samarkand.
James Elroy Flecker

Flotillas would be teams of craft of related functions, generally on a particular expedition. What their function might be, would depend on current needs. A flotilla might establish a team of craft, essentially a number of probes forming a web of function in a given region, or proceeding on a one‑way inspection of a region, hunting for valuable rogue bodies or prospective threats relevant to a future fleet, or it might be on a return journey retrieving resources from asteroids or neutron‑baking items.

One way or another, flotillas would not be carrying any live passengers except for payload, and then probably frozen or otherwise dormant. The duration of its expedition would be open ended, even indefinite if it would not be practical to retrieve the craft, such as when it is on a scouting and prospecting voyage. As far as possible, a spent fleet or fleet object would be followed for as long as signalling is possible, to glean as much information as possible from as far away as possible for long after the flotilla is expended.

Such information seems to be perennially valuable in one context or another.

So far our strategy would be to accelerate for as long as possible and the first fleet of craft, shedding a web of relay craft on the way, would need to have a team of primary pioneer craft for duplication.

It is likely that a flotilla on a prospecting voyage also would shed members en route to inspect promising or threatening objects on the way. Here too, the shedding of the probe would be be calculated to accelerate the flotilla craft in continuing on its way. For instance a rogue planet might be shot with a missile to indicate the presence of fuel, such as U or Th.

Fleets

Less is more, but less is more difficult than it looks.
Grant Snider

Fleets would be highly structured sequences of craft of multiple functions, generally not expected to be retrieved, commonly intended to establish industries or colonies, and explore colony regions of stars or other bodies. They would be large collections of craft, and in fact might comprise sequences of sub‑fleets.

The leading sub‑fleets would be intended to establish communication, navigation, and observation webs, establishing the safest routes for the succeeding sub‑fleets, and generally preparing for their functions.

The next sub‑fleets might be teams of a few craft with matching functions, each team accelerating as optimally as may be practical, picking up detailed data of the target region for the guidance of the following fleet, both threats and resources, such as dust clouds and rogues, and perhaps details of target planets in habitable zones. The teams would combine functions and data, and act as backups in case one or more members should hit something that they could not avoid at such speeds. They would not be carrying much in the line of fender masses, because heavy masses would reduce their speed.

They would however, depending on their intended route, expect to carry out slingshot steering as they pass a pivot mass such as a planet or a star, and the attitude of such attached fender masses as they do carry, could be adjusted for fine control of slingshot manoeuvres.

Data on each of the major craft in the fleet would be shared redundantly among the craft, so that loss of a few craft could be detected and reconstructed.

The main craft would bear the primary functions of the fleet, whether for colonisation or mining or whatever might be desirable enough to justify a fleet equivalent to multiple supertankers, possibly on a one‑way trip, over a period of Teraseconds.

If the objective is to be colonisation, then the Homo futurens on board should be in the form of viable larvae, frozen or more likely in diapause; this would be more practical than carrying embryos that require uterine support on spacecraft. After establishing a viable objective installation, the AGI would activate the larvae, feed, and educate them as they grow. Training them would be unnecessary, as that could be included in their mental and emotional pre‑programming.

It is not clear whether an AGI or teams of AGIs would be fully satisfactory for educating a larva resembling a miniature, highly gifted child, or an active, chinchilla‑like pet that grows and metamorphoses into full humanhood, or whether we would need an AGI‑CES for the purpose. Personally I am cynical about the ability of human educators to do better than a suitable AGI, as reflected in the epigraph, but given the sheer vagueness of our views on the nature and mechanics of CES in general, I am uncertain that it would make much difference. There is a lot that we do not yet know about Homo futurens anyway.

Besides, the hatchlings will have each other for company and mutual emotional development, plus much in their downloaded training to protect them from the most obvious social pathologies.

Meanwhile, if the colonisation had not proved too optimistic, the fleet should by now be establishing an ecology and a technological community on or in the new home. Only let them get as far as that, and we should have the basis for a second Earth, an Earth‑2, in a Terasecond or so (or perhaps an Earth‑n, if this kind of pioneering and colonisation had been going on long enough).

That should satisfy any doubters, who might be wondering how anyone could justify such an extravagance as a one-time fleet like that, planned and delivered over several Teraseconds. And yet, an entire planet, even a very modest planet, for the price of a few space fleets and a bit of patience and hard work, would be a spectacular bargain. In fact, some of our more prosperous descendant planetary colonies could be sending out their own contributions to a conscious universe within a Terasecond or so.

This is a serious consideration from several points of view. As I frequently point out, interplanetary emigration is no mitigation for planetary overpopulation, which in itself is something a civilised population should be able to avoid routinely and painlessly, and control rapidly after detecting any tendency to overpopulation, but it does not follow that there is no other incentive for colonisation, especially colonisation of sterile planets in habitable zones, or excavated into rogue planets and other promising real estate.

Such a planet would make the likes of the Star Wars super‑ships look pretty unimpressive in comparison.

Fuel for Hunter-gatherers

We know very little, and yet it is astonishing that we know so much,
and still more astonishing that so little knowledge can give us so much power.
Bertrand Russell

Meteorites have shown us that there is more material value floating about the universe than we can foreseeably expect to want. Some of it will demand a great deal of energy to extract, because it will occur largely in the form of stubbornly stable chemical compounds, such as oxides and glasses. Some will be even worse, being highly dilute, such as compounds of lithophilic rare earths.

Many of those we can recover by exploiting long-lived stars, red dwarfs in particular, to power chemical and physical extraction strategies.

I have mentioned examples of such a strategy in the essay at: Mercury, but that one discusses just a few examples applicable to just one planet; it deals with the possibilities of stripping a working area of protective crust from the mantle, and mantle from the core, exposing materials that differ in composition from common crustal minerals, and solid native metals, in contrast to the dilute, resistant ores that surface mining yields.

The materials remaining after the bulk metals (largely Fe and Ni) have been extracted, leave tailings of industrially valuable metals such as tungsten, hafnium, tantalum, platinum group, and more. Other substances of industrial value, such as diamond and silicon carbide, might be expected in some regions of planetary mantles, but how significant those might be in practice, I have no way of knowing.

The principle is that once we have found a planet, either a rogue or in orbit, that has a desirable core inside a suitably tractable crust, and is on a convenient trajectory, then if we can steer a sufficiently large body into a grazing trajectory to strip bare a large surface of a few, or a few hundred, square kilometres of core, we are in a uniquely profitable position.

The exposed material probably would be molten, or at least red‑hot; this might seem an obstacle, but the heat also would amount to a convenient source of energy to support mining. Some residual planetary cores might have been stripped by stellar explosions or similar events, but the essential point is that in the foreseeable future we need not fear running out of elements within a few atomic numbers of iron, say between titanium and zinc.

Superficially, such rare finds would seem hardly worth looking for, but the core of Mercury alone would offer something like 1E22 tonnes of indefinitely valuable raw metals. It would take feverish consumption of such a bonanza to exhaust it before we find the first rival, whether inside or outside our Solar System.

And Mercury is unlikely to be unique among rocky planets in zones of the galaxy rich in the ultra‑heavy elements up to uranium. For all we know, Sol might be unusually poor in heavy‑metal‑rich rocky planets in our region of the galaxy, compared to other local G‑type and K‑type suns.

It might be worth noting that our galaxy has particularly plentiful stars in the size ranges of interest here. Practically all the stars we could make direct use of are dwarfs of the types M, K, G. It also has rogue bodies, some of which are likely to be of special value.

Fuel Finding and Farming

On the mountains of truth you can never climb in vain:
either you will reach a point higher up today,
or you will be training your powers
so that you will be able to climb higher tomorrow.
Friedrich Nietzsche

The very concept of interplanetary travel, let alone interstellar colonisation, raises questions of the need for fuel, not just fuel in general, but accessible, concentrated, efficient, manageable portable, safe, fuel.

On Earth we have been wasting chemical fuel blindly, parasitically, in insane quantities, and with obscene consequences. We are moving in the direction of solar and other climate-related sources of energy, plus heat from the planetary core, but if we are to rely on those we need to be near to the sun and the planet; they fail in some aspects of interplanetary portability at least.

All fuel considerations change in space when we contemplate long‑term, long‑distance, large‑scale energy requirements, such as for interstellar exploration, pioneering and establishments. Of those, the most dramatic requirement by far, is power for propulsion. For long range propulsion, chemical fuels simply are too demanding of mass and space to be useful primary sources. Granted, many chemical fuels have effectively indefinite shelf lives, but, in space, chemical energy is mainly of value in some kinds of batteries and stored combustibles. Other forms of stored power, such as mechanical springs, flywheels, and batteries, have varied roles, but their capacity is trivial compared to their mass and volume, so we may neglect them in this connection.

Some optimists think in terms of propulsion by laser power, solar sails and other ingenious options, but their limitations beyond the distances over which they are of value disqualify them from practical value for interstellar propulsion.

Even at the Asteroid belt, solar power is too attenuated for most purposes, and at the Kuiper belt the sun hardly serves as more than a navigation beacon.

Granted, in regions close to a star, such as within the orbit of Mercury, solar power can be adequate for most reasonable requirements, though not for most practical forms of propulsion, and not further away.

Apart from solar power, we might speculate on the possibility of supporting colonies, either human or AI, in honeycombed tunnels inside rogue bodies that have hot cores sufficiently long‑lived to keep a colony going for Petaseconds, but for practical requirements for interstellar propulsion, that leaves us few options apart from kinetic and potential energy, and, in particular, nuclear power.

There are several ways to employ kinetic and potential power, such as control of attitude by tidal forces, and transfer of momentum by slingshotting. Electromagnetic power from solar wind and induction power from magnetic fields, also can be useful. In particular they do not require much reaction mass. The trouble is that they either apply only to special temporary situations, and that most of them are not very intense; in space we usually want highly intense energy on demand after indeterminate times, and for arbitrary applications.

For our most intense power in most circumstances, nuclear power has no rivals, and it comprises basically three forms: fission, fusion, and radioactivity.

So far in this essay, we have little reason for planning to use fusion power on spacecraft. If we do succeed in taming it in compact and practical form, that would be marvellous, but so far, commitment to fusion power in space propulsion, amounts to wishful thinking.

Radioactivity, is already of importance in such forms as nuclear electricity driven by charged particles such as beta radiation and protons, or by thermoelectricity. Whether in the form of heat or current however, it is of only specialised interest, because not much can be done to control or hoard radioactive resources; only isotopes with short half‑lives provide power with useful intensity, and necessarily for just a few centuries at most.

That leaves us with fission. So far this is far and away the only serious candidate for sustained power, but even that is not as long‑lived as one would like for interstellar power unless we can overcome some of the inefficiencies of our currently popular modes of consumption. The ways we have been dealing with nuclear waste on Earth so far, have been downright irresponsible; often criminally irresponsible. Unlike radioactive power, fission power can be stimulated to release intense energy, but that reduces its profitable life proportionately, and even at zero consumption, its half‑life is short enough to present obstacles on interstellar voyage. For example, tritium’s halflife is only a fraction of a Gigasecond.

Of existing fission options, the most promising are various breeder reactors. From a given amount of raw uranium or thorium, suitable breeder reactors could yield fifty to a hundred times more usable energy than any “once-through” fission reactor. Our current practice has been sinfully wasteful and messy, although I applaud the example that India has set. But in outer space fewer concerns with radioactivity are relevant, and the need for that extra energy is vastly greater. I see breeder reactors as vital for serious interstellar travel or colonisation.

But in themselves even breeders are not enough in the long run. Immortal Homo futurens will have to bear in mind the unavoidable halflife limitations of nuclear power. I do not discuss the details here, because the variations on options of fission power are mature and they differ considerably in their respective applicability. We need only understand that none of them is effectively eternal.

Fortunately for our view of human indefinite potential, an alternative field is likely to be of vital importance for advanced interstellar colonies.

That is the one we might call Nuclear Fuel Farming. We could in principle (and might in practice) harness the potential to extract energy from the universe for a future of thousands of times longer than the time that has elapsed since the Big Bang‑— always assuming that we do not wipe ourselves out in the first few Gigaseconds.

Nuclear Fuel Farming

Farming looks easy when your plow is a pencil

and you’re a thousand miles from a cornfield.
Dwight D. Eisenhower

Most people unreasoningly assume that the way things seem as they experience them, is the way things fundamentally are. So, for example, the elements in our local world are everywhere: iron is iron, uranium is uranium, and if you want more of either one, you just have to look around. Even when astrophysicists explain that we are stardust, that does not suggest much, because a star is a star, so, if we were born around another star things would be much the same.

That is far from true; if our dust came from a small star, we could not exist at all, because small stars such as red dwarfs, simply cannot make most of the elements necessary for our make-up. Some larger stars could just about create us, because in their destruction they can create all the elements that are necessary for our life, all the way up to iodine, which has the atomic number 53. (Life more or less as we know it, might have managed with slightly fewer nutrient elements and less versatile biochemistry, but that is a detail).

But on Earth we have elements with atomic numbers up to 92 and a few beyond, and atomic masses to match, and they cannot be made by ordinary star masses, only by special classes of star explosions and collisions, and released into space by special classes of events that have supplied all of our gold, mercury, lead, bismuth, thorium, and others. And many of the elements that do get released in those ways, technetium, promethium, francium, radium, and more, that do get released, do not even last long enough to contribute significantly to the make‑up of our planets. Any that we find are temporary traces of products of incidental processes in Earth.

Such elements not only are rare on Earth, they are even rarer in space. Whole regions of our galaxy are practically without elements that are familiar on Earth.

Furthermore, most isotopes of all our elements hardly occur on Earth, nor anywhere else except temporarily, because they break down too fast, and this includes some of the most valuable elements we can use as sources of nuclear power, such as Tritium and Plutonium.

This has important implications. For one thing, it sharply reduces all the regions of space in which we may expect to find alien life similar to life on Earth, or alien life at all. For another, even if we could have life forms without phosphorus, sulphur, potassium, calcium, manganese, or iron, or any heavier nutrient elements, their technology could not get far without a lot of the elements that we take for granted, including all our metals, and especially all our metals heavier than manganese.

But even in regions where the heavy elements do occur, and where they may be used for fuels, in particular elements with fissile isotopes, such as Thorium, Uranium, and Plutonium, isotopes valuable for power in space, we cannot take them for granted in regions where we would use them for nuclear power on interstellar journeys. Journeys that might take Teraseconds, could last long enough to limit the choice of isotopes useful for propulsion power.

Even if their isotopes lasted long enough for a few trips, they would not do for general purpose propulsion.

The cycling and recycling of matter in this universe is obscure, confusing, and not yet deeply understood. Humanity is getting by on small leakages and scraps of matter and energy scattered around by a few generations of stars since the Big Bang. Still, that implies the consolation that if we look intelligently enough and determinedly enough we might be able to find more scraps before everything winds down, if ever it does.

I cannot cover all the possibilities, but exploration of stars as sources of neutrons, raw energy, and heavier nuclei, revealed interesting options for really far‑future prospects. Breeder reactors, possibly fusion reactors, and to some extent solar (stellar, really, of course) power could see us through for intermediate prospects and special purposes, but when we run low on thorium and heavier fission power, things grow more challenging.

Post‑Big‑Bang fissile isotopes are decaying; Homo futurens had better keep an eye open for new collisions that produce new fissile fuel elements.

A promising approach would be to begin by gathering iron and nickel, in essence the plentiful, non‑decaying, ash of nuclear fusion, in energetic processes such as supernovae. A production satellite in highly elliptical orbit around an M‑type dwarf star. It could use stellar power in close orbit, to drive accelerators to smash protons into masses of iron at enormous intensities. The results would be high concentrations of free neutrons at various energies, and they would be uniquely valuable at transmuting non‑fissile isotopes into usable fuels for breeder reactors.

Not all actinoid isotopes are suitable for fuels on interstellar voyages. Journeys between stars may take so long as to make it worthwhile to avoid wasting fuel halflives on nuclear decay. One way would be to include dormant fuels in the form of isotopes so long‑lived that their decay is negligible, even in millions of years.

It should not be assumed that just because an isotope such as U-238 is not of immediate use as a fuel, it lacks value. Such isotopes may be useful because they have long halflives, which is precisely what one demands of a good storage medium. The dormant fuel would be useless for fission in that form, but could be transformed into usable fuels on demand, either by exposing them to suitable neutron energies, or, more conveniently, to feed them directly into breeder reactors.

U‑238 for example could yield Pu-239, a particularly useful fission fuel. On Earth some otherwise valuable breeder fuels may be shunned because they produce inconvenient trans‑uranium isotopes, but in space, especially on long journeys, this need not be a problem, and could even make an asset of a perceived liability.

Note that the brute‑force process of fuel conversion by neutron irradiation of isotopes is not energetically favourable; their production generally requires more input energy than they yield in either fusion or fission, but M-type stars, “red dwarfs”, have huge lifetimes, and although that implies a slow energy output, it still is plentiful enough to drive our space‑based accelerators and processors for many Petaseconds. As long as the scale of the process is large enough to keep the output flowing, no one will be in a hurry; after all, the whole process will be carried out and managed by abiotic AI devices, and for them, the question of impatience does not arise. The production of the fuel isotopes amounts to the packaging of fuels, not their creation.

Having produced and suitably packaged a batch of desired nuclear fuel, the production plant could mount batches into identity‑coded cartridges, each holding say some tonnes of fuel. At the top of its elliptical orbit, it could launch each cartridge by rail‑gun into a documented collection orbit, and enter into dialogue with client or retail craft, that then would rendezvous with one or several orbiting cartridges, to collect them for export.

Ultimate solutions need not take any course precisely matching those discussed here, but the point is that for the foreseeable future our universe seems to provide sufficient negentropy worth harvesting.

Not only armies march on stomachs

An empty stomach is not a good political adviser.
Albert Einstein

For humans to be viable biological passengers on long‑range spacecraft and various classes of space outposts, only Homo futurens could be seriously considered, and only if certain classes of adjustment to their biology were implemented. Consider the following examples:

· Nutrition could change radically. In practically all organisms Darwinistic evolution has adapted nutritional needs to ecological opportunism; in effect all nutrition is specialised to match what is available, either as it is required in its finished form, or as material that must be processed into the finished form. What is vitally necessary, but freely available, will not necessarily be created from scratch, but assimilated as essential food.
Examples include vitamins, essential fatty acids, and essential amino acids. None of our vitamins is impossible to synthesise by means of existing enzymes; that is how they are created in the first place by the organisms that we eat.
We also could even in principle assimilate nitrogen from the air; certain microbes do it on a large scale.
Such capabilities, plus the ability to reprocess wastes, would change the ecologies of Homo futurens on Earth, in asteroidal dwellings or spacecraft.

· Wastes too, in particular, are of great potential interest; they are compounds that contain valuable nutrient elements in the form of various compounds that Homo ephemerens cannot assimilate, and that we currently excrete as unwanted in our foods or as metabolic end‑products, variously useless or harmful; substances such as cellulose, lignin, porphyrins, nitrogenous wastes, keratin, and collagen. Even many toxic substances such as some alkaloids and peptides and microbial cell walls and plasma membranes, could be processed as nutrients.
Darwinistically, the attempt to combine and maintain such capabilities would be suicidal, but in terms of teleological biotechnology we should be able to deal with most of them with a few hundred genes and a few thousand items of non‑coding nucleic acid configurations.
Bearing in mind where Homo futurens is likely to travel or settle in space, the reduction in waste matter and in the volume of nutrient rations would be of incalculable value. At present, compounds such as urea and uric acid are expensively excreted, and yet in principle, they could equally easily be re-processed into compounds such as glycine. Nor is there is any fundamental reason why it should be impossible to digest cellulose into cellobiose and glucose, than amylose into maltose and glucose.

· Smell, taste, and chemoception in general would need to change to accommodate nutrition in Homo futurens. Things that currently are offensive or unpleasantly bland, should be as appetisingly attractive as say, blue cheese or condiments are to Homo ephemerens. Many hazardous substances that are indetectable to Homo ephemerens (such as low concentrations of fluorine or radon?) could well be offensive to Homo futurens.

· Although the ability to recycle many foods, and digest others from scratch at no greater penalty than extra energy, would be priceless, and it would greatly reduce the burden of conscious passengers in space flight, or as pioneers in times of scarcity, it would not abolish all objections to unnecessarily including living passengers or crew on problematic missions, such as on space vessels. Not all nutrients would be be recyclable compactly, and there still would be problems of extra mass and accommodation, not to mention physical hazard, that would not be necessary for inorganic AIs, or for frozen, or embryonic, diapausing passengers in the form of eggs or seeds.

Hypermetamorphosis: Putting Away Childish Things

His eye happened to fall upon Alice: he turned round instantly,
and stood for some time looking at her with an air of the deepest disgust.
"What — is — this?" he said at last.
"This is a child!" Haigha replied eagerly, coming in front of Alice to introduce her. ... "
We only found it to-day. It's as large as life, and twice as natural!"
"I always thought they were fabulous monsters!" said the Unicorn. "Is it alive?"
"It can talk," said Haigha solemnly.

The Unicorn looked dreamily at Alice, and said "Talk, child."
Charles Dodgson. Through the Looking glass

Inevitably, Homo futurens, in our continued future and exploration, will face a universe of variety and challenge that we cannot anticipate — can hardly imagine. Our ecologies and circumstances, our entire ways of life, will expand indefinitely. Apart from whim and experimentation, this will require changes in our bodies and minds, and no less in our reproduction than anywhere else. Not only will sexual and emotional mores change beyond recognition, but the physical and functional adaptations must necessarily be drastic and varied. The very nature of interplanetary, let alone interstellar, flight and exploration, will demand repeated adaptation and expansion; some are likely to be dramatic.

Not that our current reproductive processes, physical, physiological, or psychological, are anything less than Heath Robinsonian. . .

Be all that as it may, new planets and space colonies, whether nightmarish or elysian, whether eremitic or gregarious, will each demand individual adaptations if we are to make the most of them, or even survive.

By the very nature of multicellular life, in which reproduction generally involves notionally mature entities producing new entities of reduced size of function adapted primarily to growth, metamorphosis is arguably universal. This commonly is so among the vertebrates, and In Homo ephemerens it is particularly messy and hazardous to all parties. As derived from details of past human evolution, we have reproduction mixed up with emotional attachment, with physical pleasure and pain, with incoherent details of physiology vulnerable to hazards ranging from physiology and infection to injury and educational disaster and delay.

The offspring is generated disastrously large inside the mother, for too long, and emerges from a disastrously inadequate channel, with all too frequently disastrous results for either mother or child. The morphology and physiology of the offspring are necessarily adapted to aspects of the successive stages of growth and education. And some of the most relevant stages are inside the mother and with imperfect and misleading education.

The first Gigasecond of the child’s metamorphosis is spent in growth and various aspects of education, both personal and social. Once those stages have passed, it is unusual for much improvement to ensue.

There are many options for improving on this reproductive strategy, which is adapted to the family-and-village environment of Homo ephemerens rather than the indefinite future and mental capacity of Homo futurens. The following are illustrative classes only. Some aspects might simply be impossible, depending on how much is possible in dealing with the CES.

· To begin with, in all classes that involve a “mother” (the actual gender need not be relevant to Homo futurens; in fact they might not have gender at all) the mother’s anatomy and physiology should be adapted to deliver the child through a safe passage, and not via a bony aperture, nor associated with any other physiological functions irrelevant to reproduction, including sexual or other social activity.

· There are many possible variations on the “mother” theme; some populations might prefer somatic reproduction of some sort, in which the parental body splits, or grows a new body externally like budding. It could grow an extra head looking backward, sharing parts of the brain for silent company. And the rear head could have extra arms for fine work.

· Apart from appropriate hormonal and nutritional controls, which should be under the “maternal” control of the parent, there should be conscious neural connections between mother and fetus that enable educational processes as the brain grows, emotional connection, and, by the time of parturition, if any, some degree of conversation. As will appear, the scope for this would be constrained, and would differ variously for some forms of somatic reproduction, but they would be functionally and socially important all the same. There is no reason to waste the time of gestation on pure ontogeny.

· The birth size of the young in cases when the young are internally grown and produced by some form of parturition, should be smaller than that of current human babies, which is adapted to preparing the child to grow in a world of uncertain hunter-gathering. Babies should be small enough for easy gestation and birth that is easy for parent and child. Something between the size of a chinchilla and a small cat might be good, say one or two kilogrammes. Within hours of birth the child should be agile enough to avoid being trodden on, and it should have a protective pelage. Babies born house‑trained would be a major bonus, particularly on spacecraft.

· The brain size of the fetus should be disproportionately large, but not so large as to interfere with precocious activity. The early brain would be far too small for a full‑sized human body, but unlike a brain large enough for Homo ephemerens, it would continue growing disproportionately until adulthood, by which time it would be larger than a mature Homo ephemerens brain. Much of its growth would be based on specific primordia that would not be required before adolescence at least. This is not as alien as it may sound; even Homo ephemerens brains grow selectively for roughly half a Gigasecond, and do not complete functional maturation for nearly a Gigasecond. Homo futurens brains could be expected to exceed that by a large margin, and in fact a lot of that growth would not be in the skull, but in the torso, in outgrowths from the spinal cord.

· It would be best if the placenta would include connections for conscious mental intercourse between the developing brain of the fetus and the mother’s central nervous system, perhaps her thalamus in particular. This could enable the foundational part of the child’s brain at birth to be sufficient for basic physiological functions, plus comprehension of basic emotions and social interactions. In advance, it could understand threats to safety and rewards for commitments, receive preadaptation to speech and reading, the nature of some real‑life physical facts and values and a number of conceptions of logic and maths.
Concern has been expressed that such an arrangement would permit the mother’s preconceptions and prejudices to be imprinted harmfully, but most of the child’s brain still would have to be constructed after birth and most of that construction would be to accommodate the intellect and the nucleus of congenital education passed on from mother as conversation rather than conversion. The input still would be valuable, but would be subjected to expansion of information and perspective as the child grows.

· Selected physical regions of the late fetal brain could be segregated as biological firmware — in effect organic ROM, in practice a region of specialised brain tissue tucked away in some corner of the brain — dedicated to fixed, organised, factual, data and principles, ranging from basic algorithms to scientific validation and functional ethics. A corresponding area of labile neural tissue — more like an organic scratch RAM — would be dedicated to conditional, editable, scratch memory. These congenital reference organs, comprising highly configured tissue, would comprise a small fraction of the adult brain volume, and probably would work on principles somewhat different from “natural” brain. A few cubic centimetres could easily have the capacity for the equivalent of whole shelves of reference documents.

· It is important to understand that such ROM, no matter how loaded or installed, has no direct role in the child’s brain other than as a source of reference; the brain and education would develop normally in every other respect, but, we hope, rather better than otherwise. Downloads would not replace human intelligence or mental processes, any more than reading would; rather, they would serve as incorruptible reserves of information, rendering Homo futurens structurally resistant to bad-faith indoctrination and propaganda throughout life, and during youth in particular.

· It is hard to be sure what forms of propaganda and gullibility might be exploited, by such forms of antenatal instruction, ranging from religion and politics, to commerce and quackery, but it might be easier for dictatorships to concentrate on AI robots than humans. Even better, antenatally instilled scepticism in Homo futurens might render future humans particularly resistent to bad‑faith indoctrination and disinformation.

· The growth of the child, first from the pre‑larval stage in gestation to just after birth, then from pet size to human child size, and a few stages in between, could logically and literally be described as an elaborate form of the well‑known biological principle of hypermetamorphosis, in which there are distinct states of development in the life cycle, each specialised to functions appropriate to their respective stages of growth and activity.

· There are many possible non‑maternal versions of reproduction, some already in existence in nature here on earth. Some organisms, while in diapause, such as in the state of eggs, small larvae, or seeds, while frozen or dessicated, could be suited to transport for interstellar pioneering. While inactive, they would not require bulky, complex, artificial embryogenic or incubational equipment. At emergence, as the time of arrival approaches, they would need just awakening, feeding, and education. All of that could be managed by AGIs or AG‑CESs. That early education would be on the same lines as the pre‑birth loading of ROM content hypothesised for placental education by a biological mother, though perhaps more rigidly formulated. Note however, that information downloading into the brains of growing minds, is only a fractional aspect of education, and largely less important. It should however not be underrated, because in normal human schooling, a great deal of educational effort and capacity is wasted on volatile rote memorisation. Being able to store such information in stable forms, and without wasting time on misapplied educational effort, while instead paying profitable attention to proper education, is genuinely in valuable.

· Note too, that once a child has been raised from a diapaused state to normal social and reproductive activity, its future nature and history would not differ from that of any other member of its society or its offspring, any more than a twenty‑first‑century incubator baby would be limited in future to reproduction by incubator.

· There is no question of replacing Homo futurens with AGI or even AI‑CES; We cannot predict what the roles of biological members of society should be, but it is quite plausible that they might be more inclined to novel mental exploration and interdisciplinary connection between AI devices in general, and AI-Homo cooperation in particular. They also are likely to be more readily available for unexpectedly varied classes of work and exploration.

With the physiological vulnerabilities of the species engineered away, the final, unavoidable frontier of Homo futurens is the nature of subjective consciousness itself. Replacing the biological population entirely with AGI is no foreseeable objective; even if it eventually proved practicable it is not currently clear whether there would be any point to doing so; the challenge lies rather in how these entities interface.

This brings us to what still seems to be the ultimate unknown of a Petasecond civilization: the Cogito Ergo Sum (CES).

AI‑CES

MIND: A mysterious form of matter secreted by the brain.
Its chief activity consists in the endeavor to ascertain its own nature,
the futility of the attempt being due to the fact that
it has nothing but itself to know itself with.
Ambrose Bierce

The very fact that we deal with unknowns and have to face the reality that some unknowns are unknown, practically guarantees that we shall find ourselves in unpredictable intellectual collisions and traps of our own making

AI‑CES is a good example. It is a completely speculative prospect. The very concept of an AI-CES is a speculative obstacle. We currently possess hardly any empirical framework to define the CES, no mechanism to generate it, and no instrument to detect it. The only assertions of its meaningful existence that we can reasonably support are that, at least subjectively, there are states of the brain that represent its content, and that changes of those states affect its operation and content.

That might seem abstract to the point of immateriality, but far otherwise: it materially implies information content.

And information is material.

That is a very limited point, but very fundamental. For each of us the existence of our subjective consciousness is real. It remains strictly internal, bound by Descartes' centuries-old criterion: Cogito Ergo Sum: “I think therefore I exist”. We cannot empirically verify a CES within another human, let alone determine if an AGI's subjective awareness — if it exists at all — shares anything meaningfully in common with our own capacity for pain, joy, or boredom. We are effectively engineering entities of Petasecond longevity without understanding the nature or existence of their suffering, or any other mental or emotional state.

But how are we to tell that any CES exists inside the AI (or, for that matter, inside the chimpanzee, or parrot, or cicada, or even another human)? How indeed, before we have even discovered a technology to characterise or work with a CES?

The only way for us to tell so far, is to ask it and take its word for its answer, or base guesses on its observed behaviour. And even when we have satisfied ourselves on that point, we cannot tell that its CES resembles our CES — we humans cannot even tell each other what it is like to be each other, let alone another species of animal, such as a bat. In fact, each of us cannot even tell how many CESs there might be in our own respective brains.

I have never seen any coherent, let alone cogent, characterisation of the nature or function of the CES; one common assertion is that it does not exist at all, and is our imagination. That I do deny, on the basis of the argument from the material nature of information.

And yet, however long I have discussed the topic, I still lack any idea of what any CES is for in biological terms, or what to do with it in any other terms.

CES seems to be the entity with which or in which we emote, care about, or wonder about, or explore ideas and feelings spontaneously. Without it we seem unable to feel pain or joy, but conversely, pain and joy are not well defined; we cannot say how each of us feels pain and joy, but we can tell that they differ between people much as, for example, our muscular control differs.

And yet, the idea of losing the whole universe to meaninglessness, even that crumb that our heritage amounts to, is as blasphemous an idea as is possible to anyone with a CES, who seriously thinks of how it implies the loss of everything that ever was beautiful or exciting.

But no one has yet spelt out what the CES is. We can tell at least that it implies information, because, at least where we find it inside us, there is a difference between whether it is there or not, and between which states — happy, hurt, angry, or sad — and the one thing that you cannot have information without, is difference, and difference is what you cannot have without information.

Arguably, the most fundamental functional difference between organisms like humans, and mechanisms like robots with their programs, is the fact that organisms practically universally come into being from a single cell that then divides to form a colony. That colony of cells must develop into into a working brain with a working CES.

The CES must grow, and it changes as it grows in contact with its environment and changes in the state of the brain. One does not even know how many CESs there are in our own brain; or how self‑aware such an unrecognised CES might be. It certainly seems from split-brain experiments, in which the corpus callosum is severed, that the two halves seem to work very much like one CES each, but that they can be shown to comprise two independent CESs.

Apparently similarly to dicephalous conjoined twins, but with little external sign or symptom.

Our robots and AIs on the other hand, simply get assembled, and the initial functional program gets loaded, though in practice such programs commonly are designed to learn more as they operate. Whether the biological brain differs fundamentally in its nature, from a neural network that learns artificially, I have no way of guessing. So far I simply assume that at some time in the not too distant future, we will understand the mechanism as well as we currently understand say, electromagnetic mechanisms.

How would an AGI with one or more CESs differ from a brain with one or more CESs? One thinks of differences such as that an AGI would be inactive without any instructions to proceed; that it would ask itself no spontaneous exploratory questions such as occur in humans (or other organisms); the likes of "that's funny. .. or strange. . .?" or "why?" or "would it not be nice if. . .?" or "how can I. . .?" or “In what way does that look familiar?” or “What would happen if. . .?” In general, arbitrarily playing with thoughts, and arbitrary curiosity, seem to be very human. Are they also very CES? Or does humanity depend on having multiple internal CESs in parallel?

Such mental activities strike me as being products of Darwinian selection for higher intelligence, but they also also seem to me to be in principle implementable in the thinking mechanisms of a mechanical AI. That seems to me more of a programming challenge, than a fundamental attribute.

But to what extent does having a CES imply pain or pleasure or emotions in general? We do not yet know. Even self‑aware humans have been known to lack a sense of pain, and some humans seem to lack various emotions altogether, or to have emotions distorted to extents that seem to be pathological. This clouds the ethical issue: supposing we were to find out how to instil a CES into a robot, how do we know whether we are creating a suffering being or not? How can we tell that an emotion is good or bad, functional or pernicious?

However it works, if we are to have AI‑CES prospectors, miners, marine and submarine monitors, inventers and improvisers, explorers, scientists, ambassadors, and leaders, they will need to be capable of finding for themselves, the right questions, as well as the right answers, in their routine or emergency operation. From one point of view aspects of those comparisons began as contrasts between humans and AIs, but from others, being able to characterise them suggests possibilities of machines with CES, or something similar.

Imagine establishing an AI in a spacecraft on a multi‑Terasecond voyage. We are convinced that a CES‑less AI will not suffer boredom, any more than a clock would. In fact we even could design an AGI that switches itself off, perhaps to save energy whenever it has no functional reason to remain active. It still could respond to a trigger event to wake it up as required, much as a human might set an alarm clock and sleep undisturbed until it is time for action. Things need not be any worse for an AGI‑CES if it can switch itself off, being woken automatically, perhaps after a few Teraseconds, either when alarmed or as scheduled.

What we can be sure of is that if an AGI is destroyed after having recorded information, that could amount to arbitrarily great loss of discoveries. Could that loss be even greater in dealing with AGI‑CES? Perhaps, or perhaps not.

Again, if a conscious AGI‑CES is left with nothing to do, or is subjected to harm, it conceivably might suffer boredom or pain, but that need not follow logically. We certainly should accept that if it actually does suffer, we must accept the responsibility, just as we would in our treatment of a human or animal; furthermore, if we cannot show that it does not suffer, we must assume that precautions against unnecessary suffering are adequate.

To avoid such abuses in dealing with artefacts equipped with subjective sensitivities for whatever purpose, does require suitable precautions, but they do not seem challenging: an AGI‑CES with no mental demands on its time, and no concept of physical weariness, might emulate Leonhard Euler, who, when he was blind and insomniac in his old age, could lie for hours calculating log tables till daylight; the next night he would continue where he had left off. I do not assert that this would satisfy an idle AGI‑CES, but such devices would have petabytes of data in brains or in storage, accessible for background data mining, working on data from libraries, or information gathered by the instrumentation.

Many entail unacceptably high risk — self‑sacrificial roles — possibly unavoidable expenditure, but unacceptable for humans when artefacts could be sacrificed instead.

We cannot however dictate all future principles in advance; our current obsession with the sanctity of human life cannot be absolute; for example, what are we to do when given the choice of saving one life or saving several? Or saving one friend rather than saving many strangers?

Even apart from such dilemmas, the nature of Homo futurens changes with time, and the changes can affect our basis for ethical decisions. Suppose one has deliberately reproduced somatically, with shared subjective consciousness till just before the choice of who shall die in averting a deadly situation, say in deciding who is to get the last survival craft for escaping from the stricken spacecraft when there is an accident. What if every clone is to get a seat in a survival craft, leaving one subjectively distinct passenger to die for lack of any hope of rescue?

What if the distinct personality has special knowledge of great importance, whereas the clones have duplicated knowledge available? What if, on a hugely valuable spacecraft, there is a Homo futurens who has lived for a Petasecond and wishes not to prolong life, once the knowledge has been delivered; but suppose a disaster is impending and the passenger can only be saved at the cost of the spacecraft, including a the entire crew of AGIs?

Unknown unknowns. . .

The Future, Not the End

Science is as yet in its infancy, and we can foretell little of the future
save that the thing that has not been is the thing that shall be;
that no beliefs, no values, no institutions are safe.
The future will be no primrose path. It will have its own problems.
Some will be the secular problems of the past,
giant flowers of evil blossoming at last to their own destruction. Others will be wholly new.
Whether in the end man will survive his ascensions of power we cannot tell.
But the problem is no new one. It is the old paradox of freedom
re-enacted with mankind for actor.
John Burdon Sanderson Haldane

It is important to remember that Homo ephemerens, as we are now, is a dead end, and doomed to ignominious suffering and extinction, taking much of the living world with him. This essay urges that instead, we aim for constant improvement and — yes — fun, making ourselves the first species in the history of Earth to, as Fitzgerald put it: “make game of that which makes as much of thee”. . .

We have far to go first: The age of the amphibians lasted for tens of millions of years; the age of reptiles over 150 million. As Homo sapiens, we have lasted less than 1 million. If we do not kill ourselves with war, we could be taken out by an astronomic disaster any year now.

Not to make the most of our technological advances, and starting really soon, would be criminal self deception. You and I will not see the rewards, but we can work towards building a future for a worthwhile humanity.

That too, is worth while.

Full Duplex

Wednesday, March 11, 2026

Future, Futility, or Finity

Homo