»Human intelligence is literally the most powerful force in the Universe, limited at present only by the “wetware” of the brain. Artificial Intelligence, if designed to emulate human intelligence, could transform “Life, the Universe, and Everything,” from solving the existential problems on Earth today to exploring extraterrestrial worlds in the future.« — Jeffrey Watumull Chief Philosophy Officer and Director of AI at Oceanit, where they endeavor to create human-style AI and discover whether its principles obtain in extralinguistic domains (e.g., physics, biology). He is also cofounder of the Cambridge Institute for Exo-Language. Watumull studied at Cambridge, MIT, Oxford, and Harvard.

A.I. to E.T.

The Future of Science & Technology and The Fate of Our Civilization & The Cosmos

Transcript

1 00:00:00,080 –> 00:00:00,320

Lovely.

2 00:00:00,640 –> 00:00:12,960

So A.I. to E.T. and the ascent to the extraterrestrial begins terrestrial or should begin terrestrially with CL or CIEL.

3 00:00:15,160 –> 00:00:18,320

which is the Cambridge Institute for Exolanguage.

4 00:00:18,800 –> 00:00:22,320

Incidentally and serendipitously, CIEL is French for sky.

5 00:00:22,640 –> 00:00:33,600

But this is a pan-disciplinary center to be launched in the spring as a collaboration of academia and industry, primarily between Oceanit here in Hawaii and the University of Cambridge.

6 00:00:33,760 –> 00:00:43,040

And it is to explore the universality of intelligence, specifically its linguistic basis and possible extraterrestrial intelligences.

7 00:00:43,440 –> 00:00:57,120

And the theoretical foundations of CIEL is NOME, the Noetic Mathematical Engine, which is anthronoetic, that is human-level, human-style artificial intelligence, which I shall expound in due course.

8 00:00:57,360 –> 00:00:57,760

So

9 00:00:58,560 –> 00:01:10,960

With one of my PhD supervisors and co-founder, Professor Ian Roberts of the University of Cambridge, and with patronage from Virgin Galactic, SETI, and others, this is our future home in Cambridge at the Stephen Hawking Building.

10 00:01:11,200 –> 00:01:18,000

And the inset there is the Eagle Pub, which for some of you may know, this has some historical resonance.

11 00:01:18,000 –> 00:01:22,640

That’s where Watson and Crick celebrated the discovery of the structure of DNA.

12 00:01:23,600 –> 00:01:29,680

So to conjecture to the nature of life, intelligence, and language out there in the universe is a grand enterprise.

13 00:01:30,400 –> 00:01:48,160

I think to quote one scholar he described it as, given the very richness of the multidisciplinary and multicultural resources required by individual explanatory hypotheses, enables us to claim that this is the most complex multidisciplinary problem in all of contemporary science.

14 00:01:48,479 –> 00:01:53,120

And this is why CIEL will include not only linguists and mathematicians and philosophers such as myself, but

15 00:01:53,680 –> 00:02:05,080

other equally curious creatures such as astronomers, biologists, particularly astrobiologists, engineers of all breeds, AI researchers, anthropologists, classicists, literary scholars, artists, inter-allios.

16 00:02:05,080 –> 00:02:14,720

So for instance, befitting a books and music festival, three of our four or three of many future scientific advisors have books out this year.

17 00:02:14,720 –> 00:02:16,800

So one is Eric Kirschenbaum.

18 00:02:16,800 –> 00:02:17,920

He’s A Cambridge zoologist.

19 00:02:17,920 –> 00:02:20,160

He explains how the evolution of life

20 00:02:20,960 –> 00:02:24,160

on exoplanets could well resemble that on Earth.

21 00:02:24,160 –> 00:02:43,360

And Avi Loeb, who is the chair of astronomy at Harvard, has conjectured controversially that Oumuamua, the first interstellar object observed in our solar system, was no naturally occurring phenomenon, but in fact could well have been an artifact of an intelligent extraterrestrial civilization, perhaps in light sail.

22 00:02:44,360 –> 00:02:57,360

And finally, Chiara Marletto, who’s a physicist with whom I studied as an undergraduate at Oxford, she works on a new theory of physics, constructor theory, which endeavors to reformulate the laws of nature in computational terms and the physics of causation.

23 00:02:58,079 –> 00:03:14,800

This is all to explore strange new worlds, or rather to discover and formalize those principles of language that can be truly universal, from which wooden could construct a universal library of Babel with many profound implications for science, technology, philosophy.

24 00:03:15,440 –> 00:03:21,840

And on the immediate horizon, as Stuart was referencing, which is a project that we will be launching,

25 00:03:22,960 –> 00:03:25,520

to beam a message with SETI and METI.

26 00:03:25,520 –> 00:03:27,520

So SETI is the search for extraterrestrial intelligence.

27 00:03:27,560 –> 00:03:29,880

METI is messaging extraterrestrial intelligence.

28 00:03:29,880 –> 00:03:40,400

And the idea is to beam a message out into space in 2024, which is the jubilee of the storied Aracibo message designed by Frank Drake and Carl Sagan in the 1970s.

29 00:03:41,600 –> 00:03:42,000

And

30 00:03:42,320 –> 00:03:45,720

This will be the, as I said, the Jubilee, and we will send out an anniversary message.

31 00:03:45,720 –> 00:03:48,240

And this is about one of our projects, but it’s particularly provocative.

32 00:03:49,040 –> 00:03:53,760

Alas, we will need an alternate telescope because Aracibo died in 2020, but

33 00:03:54,320 –> 00:04:00,480

the FAST telescope in China could be adequate to the task, and SETI is negotiating that deal now.

34 00:04:00,640 –> 00:04:12,120

But one outrageous possibility, which I favor, is to encode the message in photons and imprint it on the cosmic microwave background, effectively etching it into space-time itself.

35 00:04:12,120 –> 00:04:15,520

I shall return to this outrageous conjecture at the end.

36 00:04:15,760 –> 00:04:16,320

So

37 00:04:17,120 –> 00:04:18,800

Why send the message?

38 00:04:18,800 –> 00:04:19,920

I think there are two reasons.

39 00:04:19,920 –> 00:04:24,200

The first is philosophical, of course, and the 2nd will be technological.

40 00:04:24,200 –> 00:04:33,520

But first, and my judgment foremost, is the philosophy represented and expressed in this.

41 00:04:37,440 –> 00:04:45,280

I met a traveler from an antique land who said, two vast and trunkless legs of stone stand in the desert.

42 00:04:46,080 –> 00:05:11,840

Near them, on the sand, half sunk, a shattered visage lies, whose frown and wrinkled lip and sneer of cold command tell that its sculptor well those passions read which yet survive, stamped on these lifeless things, the hand that unlocked them and the heart that fed, and on the pedestal these words appear.

43 00:05:13,280 –> 00:05:17,440

My name is Ozymandias, King of Kings.

44 00:05:18,080 –> 00:05:20,960

Look on my works, ye mighty, and despair.

45 00:05:25,280 –> 00:05:26,960

Nothing beside remains.

46 00:05:27,600 –> 00:05:36,159

Round the decay of that colossal wreck, boundless and bare, alone and level sands, stretch far away.

47 00:05:36,720 –> 00:05:48,640

So this sonnet written in 1811 by the poet Percy Shelley, husband to Mary Shelley, tells the tragedy of a civilization that’s doomed not simply to extinction, but to effacement.

48 00:05:48,800 –> 00:05:53,760

That is the knowledge of it, the beauty it created or failed to create is lost forever.

49 00:05:54,000 –> 00:05:57,840

And the profound philosophical question there is, why is this a tragedy?

50 00:05:58,360 –> 00:06:02,800

It is profound because like the absurdist philosopher Albert Camus,

51 00:06:03,440 –> 00:06:13,200

Quote, “I continue to believe that this world has no ultimate meaning, but I know that something in it has meaning and that is man because he is the only creature to insist on having one” end quote.

52 00:06:13,440 –> 00:06:17,280

So in other words, there’s no divine design for us in this universe.

53 00:06:17,280 –> 00:06:23,160

However, simply by saying that, saying these words, I’m committing myself that these words have meaning.

54 00:06:23,360 –> 00:06:27,600

So even though nothing has ultimate meaning, I’m committing myself to the meaning of my words.

55 00:06:27,600 –> 00:06:30,320

In other words, we create meaning via language.

56 00:06:30,800 –> 00:06:32,480

And Shelley’s Ozymandias

57 00:06:33,120 –> 00:06:38,080

understood this, hence his denouement, to etch his epigraph into eternity.

58 00:06:38,360 –> 00:06:40,080

And he succeeded in a poetic sense.

59 00:06:40,080 –> 00:06:47,280

And our message beamed out to the cosmos in 2024 would be our sidereal sonnet, our sidereal symphony.

60 00:06:47,680 –> 00:06:54,080

And most importantly, this work of art, as it were, need never be observed by any intelligence out there to be meaningful.

61 00:06:54,080 –> 00:07:00,480

Ironically, it’s more meaningful, in fact, if it’s never read, because like art, it is meaningful in and of itself.

62 00:07:00,480 –> 00:07:16,000

And I think Camus concludes with words to the effect that the world has at least the truth of man, and our task is to provide justification against fate itself, and has no justification but man, hence he must be saved if we want to save the idea of life.

63 00:07:16,480 –> 00:07:17,440

And this

64 00:07:17,720 –> 00:07:23,040

this sidereal sonnet would save the idea of life, and I shall return to this at the end.

65 00:07:23,920 –> 00:07:27,160

The second reason for beaming this message in space is technological.

66 00:07:27,160 –> 00:07:28,000

Like the first reason

67 00:07:28,440 –> 00:07:31,200

It does not presuppose the existence of extraterrestrials.

68 00:07:31,440 –> 00:07:35,360

Indeed, CIEL is not even committed to the prediction that we will ever encounter such beings.

69 00:07:35,680 –> 00:07:43,159

It presupposes only the possibility of artificial intelligence here on Earth, a possibility that we’re seeking to transform into actuality at Oceanit

70 00:07:43,160 –> 00:07:45,000

with our AI named Gnome.

71 00:07:45,400 –> 00:07:47,680

The argument we had at CIEL and at Oceanit

72 00:07:47,920 –> 00:07:55,360

would proffer is the following, that for this message, we should create an AI, such as NOME, and beam it into the cosmos.

73 00:07:56,320 –> 00:08:07,520

This noetic mathematical engine is obviously a conceptual and phonetic homage to the linguist and philosopher Noam Chomsky, who was my former supervisor and one of the advisors to CIEL.

74 00:08:08,680 –> 00:08:14,760

And the architecture of NOME is designed to emulate the causal reasoning of the ideal human scientist.

75 00:08:14,760 –> 00:08:26,560

So we use language to conjecture theories and subject them to criticism and error correction to construct counterfactual statements and causal models that enter into explanations, which is the stuff of knowledge.

76 00:08:26,560 –> 00:08:28,800

And it is rational but chaotic and evolutionary.

77 00:08:29,360 –> 00:08:37,680

It was enacted by Galileo, whose logic is philosophical and mathematical and empirical, none of which was predominant and determinative, hence it was technically anarchic.

78 00:08:37,920 –> 00:08:38,600

Anything goes.

79 00:08:39,040 –> 00:08:46,520

of absurd thoughts, counter-intuitive theories of free-falling bodies, absurd experiments, and absurd thought experiments.

80 00:08:46,520 –> 00:08:52,880

So the leading power of Pisa experiment was surely a Godakin experiments, thought experiments, as were virtually all of the fabled

81 00:08:54,280 –> 00:08:55,680

experiments in the history of science.

82 00:08:56,240 –> 00:09:07,520

But as Chomsky said, as a latter-day Galileo, he said that pushing a precise but inadequate formulation to an absurd conclusion may be an important method of discovery.

83 00:09:08,040 –> 00:09:13,680

And that philosophy has grown here at Oceanit, but alas, my beard has not.

84 00:09:14,120 –> 00:09:14,480

one day.

85 00:09:14,800 –> 00:09:26,240

So Galileo created his Copernican theory in opposition to the contemporary empirical science and religious dogma, for the empirical data and religious teachings of his time contradicted his theory.

86 00:09:27,360 –> 00:09:35,440

Indeed, contrary to popular myth, Galileo’s theory was no better supported empirically than was that of the church, but critically, it was a better explanation.

87 00:09:35,760 –> 00:09:42,400

And the ex-foundation of his explanatory theory, the dissolution of its anomalies, and its eventual empirical corroboration,

88 00:09:43,080 –> 00:09:52,600

necessitated imaginative research, the invention of instruments, and ingenious theory-dependent construals and reconstruals of data, and the making of new evidence.

89 00:09:52,600 –> 00:09:59,920

So whatever method or non-method he needed, he used to create explanations of ever greater depth and with ever farther reach.

90 00:10:00,160 –> 00:10:04,880

Of course, contemporaneously, or contemporarily, his, you know, that

91 00:10:05,280 –> 00:10:11,520

and imagination, inventiveness, and ingenuity were regarded as heresy, conjuration, absurdity.

92 00:10:11,920 –> 00:10:27,760

And as the philosopher, the epistemologist of anarchy, Paul Feyerabend, so eloquently argued, he says that such unreasonable, nonsensical, methodological foreplay turns out to be an important precondition of clarity and empirical success.

93 00:10:28,240 –> 00:10:32,560

And such foreplay is not at all formulaic, but essentially aesthetic.

94 00:10:32,880 –> 00:10:43,640

And such epistemological anarchism, it dovetails quite swimmingly with Oceanit’s intellectual anarchy, which is essentially synonymous with critical rationalism.

95 00:10:44,320 –> 00:10:53,520

the philosophy of NOME, which is the theory of human-style knowledge, where we work to in which we’re working to encode at Oceanit in which we aspire to practice ourselves.

96 00:10:53,840 –> 00:10:59,600

In that epistemology, human knowledge creation, such as that engaged in by Galileo, is an evolutionary process as a set of

97 00:10:59,800 –> 00:11:29,200

conjecturing, criticizing these linguistically structured causal explanations in which the fittest or the best explanation survives conceptual analysis and empirical testing. So causal explanations are fundamental and indispensable to science as answers to why questions, transcending the merely statistical reduction summary of data. And the linguistic form of an explanatory conjecture assumes that of a counterfactual. Each sentence then encodes a modal verb like could or would or can or cannot,

98 00:11:29,560 –> 00:11:55,120

and these expressions of causal possibility or impossibility. So how do we as humans and potentially as AI generate such conjectures? With Chomsky, we understand that it is a grammar or a generative grammar. A grammar is a set of mathematically elegant rules generative of infinite complexity. And this is the core of our anthronoetic AI. And this that is the basis of our anthronoetic conjecture.

99 00:11:55,720 –> 00:12:24,720

which is that knowledge creation via this critical rationalism is substrate independent. That is, it need not be implemented in the wetware of brains. It could be in the hardware of a computer or in arbitrarily many other substrates. And any physical transformation is possible given the requisite knowledge. Knowledge is a kind of information, which is representational, causal, and explanatory. Information processing is computation. Hence, any physical transformation is possible given the requisite computation.

100 00:12:25,120 –> 00:12:54,880

Human intelligence is a universally powerful computation, and hence an AI endowed with human intelligence would be a universal intelligence. And this is controversial, but I enjoy this quote to, this is a David Deutsch, who’s a philosophical physicist. He says, “base metals can be transmuted into gold by stars and by intelligent beings who understand the processes that power stars, but by nothing else in the universe.” That is to say, the set of all conceivable transformations of physical

101 00:12:55,080 –> 00:13:20,320

objects, such as superluminal communication, some of them, some transformations can never happen because they’re forbidden by the laws of nature. Some, such as star formation, happen spontaneously, and some, such as converting air and water into trees or converting raw materials into a radio telescope, these are possible, but happen only when the requisite knowledge is present, for instance, that which is embodied in genes or in brains and minds.

102 00:13:20,560 –> 00:13:38,240

But those are the only possibilities. And it is that latter set, the set of transformations that are possible given the knowledge and minds that is the largest. Hence, everything that is not forbidden by the laws of nature is achievable given the requisite knowledge. That is, anthropometric minds are low and capable of creating. And we can think of it this way.

103 00:13:39,040 –> 00:14:06,840

that to understand the laws of gravity, we should study black holes, where gravity is infinitely intense. Analogously, as some astrobiologists have argued, such as Sarah Walker, we should study life because that is where the laws of information are most intense. And with NoME, we go further and argue that we, to understand the laws of life, we should study intelligence and language, for that is where the laws of life are infinitely intense. And this is the

104 00:14:07,440 –> 00:14:34,800

radical and tendentious claim for today in 2020, AI does not mean a machine endowed with anthropometric, that is human-style intelligence, to say nothing of the conjecture that such intelligence is a fundamental force and the most causally effective force in the universe. It means machine learning, specifically the deep learning artificial neural networks that power your Google search, your Facebook, or your Meta as of today, and your Twitter feeds, Apple, Siri, Amazon, Alexa,

105 00:14:35,200 –> 00:15:02,880

your non-existent self-driving car and oodles of other apps and impressive and useful gadgets, which, as I said, are useful and interesting, but rather trivial in the cosmic scheme of things. Nevertheless, scientists, engineers, and Silicon Valley gurus and the public are projecting that machine learning will revolutionize the future. For instance, Kai-Fu Li had published a new book last month, prophesizing that in 2041, machine learning will attain this, or affect this revolution.

106 00:15:03,520 –> 00:15:28,960

that machine learning is now here in 2021 at a tipping point. The days of slow progress are over. In 2020, machine learning, such as deep learning, solved a 50-year-old riddle of biology called protein folding. The technology has surpassed humans and speech and object recognition, served up digital humans with uncanny realism in both appearance and speech, and earned passing marks on college entrance and medical licensing exams.

107 00:15:29,640 –> 00:15:55,200

It’s outperforming judges in fair and consistent sentencing and radiologists in diagnosing lung cancer, as well as powering drones that will change the future of delivery, agriculture, warfare. Finally, it’s enabling autonomous vehicles that drive more safely on the highways than humans. And as it continues to advance and new applications blossom, where does it all lead? And to this, I say, Every word of what you just said was wrong.

108 00:15:55,920 –> 00:16:20,720

It’s wrong because deep learning, indeed virtually all machine learning today is based on this delusion that intelligence, that an intelligent mind ought to abstract a general rule or pattern from particular data by some magical method of induction, seeking to erect these certain, you know, to erect certain knowledge on unquestionable foundations. And that is what Aristotle, centering here on the right, thought in ancient priests.

109 00:16:21,240 –> 00:16:40,560

what his empiricist descendants, such as Hume and Locke, thought in the 17th century Enlightenment, but what Aristotle’s teacher, Plato, on the left, understood, as did his Enlightenment descendants, the rationalists, such as Descartes and Leibniz, was that induction is logically impossible. In fact, there is no place for symbolic rules in artificial neural nets.

110 00:16:41,240 –> 00:16:53,600

And that what humans do in creating knowledge is to make bold, foundationless conjectures over an unbounded range. In some cases, consciously as scientists, but in most cases, subconsciously, that’s when children, quote unquote, learn.

111 00:16:53,840 –> 00:17:12,880

language automatically and efficiently and speedily from small and sparse data, a miraculous task impossible for any deep learning system today and I think ever in the future. And logically, this competence to generate conjectures itself cannot be pre-programmed, but it must be pre-programmed in the genes or encoded in an AI.

112 00:17:13,640 –> 00:17:43,560

Nevertheless, inductive logic or pseudo-logic, whose paragon is surely Bayesian statistics, continues to dominate not only AI, but all the sciences added into popular culture. For instance, in discussions of the pandemic, whenever you hear the trendy term priors, as in, in light of a new information, we need to update our priors, you’re being infected with a bit of Bayes. And this formula, whose eponymous inventor was an 18th century Presbyterian minister, it stipulates how one ought to calculate the amount of credence to assign particular

113 00:17:44,000 –> 00:17:55,360

That is, how strongly should we believe particular propositions or claims or hypotheses and so on, based on the evidence? That is, how should particular observations and the data and so forth change our beliefs?

114 00:17:55,800 –> 00:18:21,920

The mathematical formula says that the probability of a hypothesis A given some data B is determined by the probability of A and B divided by the probability of B. In other words, the probability of a hypothesis given the data is the probability of the hypothesis and the data, say the patient has COVID-19 and the test result is positive, divided by the probability of the data, hence, or here would be the total proportion of patients who tested positive.

115 00:18:22,520 –> 00:18:52,320

healthy and sick. And there are some additional technical details that describe how the prior probabilities enter into the calculation, how to update them a lot of new evidence, but you get the gist and so has the public. So to quote the New York Times in an article entitled How to Think Like an Epidemiologist, Bayesian reasoning comes awfully close to the working definition of rationality. In this century, Bayesian statistics has grown vastly more useful because of the kind of advanced computing power that did not exist 20 years ago. Bayesian

116 00:18:52,440 –> 00:19:18,960

statistics are rippling through everything from physics to cancer research, ecology to psychology. Bayesian reasoning is indeed the paragon of induction, the notion that we derive or extract or generalize our ideas from data seeking ever more probable beliefs and theories. And if this is indeed the definition of rationality, then it’s unsurprising to read in Scientific American that artificial intelligence researchers, including designers of Google self-driving cars,

117 00:19:19,320 –> 00:19:44,560

employ Bayesian software to help machines recognize patterns and make decisions. Bayesian programs sort spam from email, assess medical and homeland security risks, and decode DNA, among other things. One Nobel laureate in physics believes that Bayesian machines might be so intelligent that they make humans obsolete. So indeed, some physicists, such physicists are praying that Bayesianism can’t explain

118 00:19:45,080 –> 00:20:10,880

or I would say explain away the difficulties of quantum mechanics. Cubism, which is quantum Bayesianism, is increasingly popular for it so tightly sweeps away the ontological problem of explaining the objective reality of quantum mechanics under an epistemological rug. That is, the answer is not out there in the world, but in our minds. That is, the distortions of our subjective observations. So for instance, a quantum particle can be in a range of possible states,

119 00:20:11,320 –> 00:20:38,480

when an observer makes a measurement, she instantaneously collapses the wave function, that is the quantum state, into one of the possible states. Cubism argues that this collapse just reflects the updated knowledge of the observer. She didn’t know where the particle was before the measurement, but now she does. So whether there is such a phenomenon as the wave function collapse is an independent problem, but Bayesianism, like all inductive method, conceives the enterprise of science and rationality generally,

120 00:20:38,920 –> 00:21:07,360

as one of our beliefs about the world, not the world itself. And it is this philosophy of knowledge that dooms induction and the machine learning systems based on it to failure, including new quantum AI projects such as that at Google. Of course, I do not mean that such projects will be absolute failures. Our classical, that is non-quantum machine learning has already proven extremely useful in, as I said, oodles of gadgets. And I mean, but I do mean it will fail to attain human level, human style intelligence. And the proof of this

121 00:21:07,760 –> 00:21:27,840

is elementary. So first, let us state the assumptions of Bayesian science. So the assumptions are that the objective of science is or should be to increase our credence for true theories, and that the credences held by a rational thinker obey Bayesian probability.

122 00:21:29,000 –> 00:21:51,040

Now, the refutation is simple. If T is an explanatory theory, such as the sun is powered by nuclear fusion, then it’s negation, not T, that is, the sun is not powered by nuclear fusion, is not an explanation at all. Therefore, suppose implausibly for the sake of argument that one could quantify whatever this property that science endeavors to maximize.

123 00:21:51,640 –> 00:21:59,720

If T had an amount Q of that, then not T would have none at all, not 1 minus Q as the probability calculus.

124 00:21:59,920 –> 00:22:03,280

require if Q were a probability, QED.

125 00:22:03,280 –> 00:22:06,320

So Bayesianism is logically fallacious.

126 00:22:06,640 –> 00:22:08,400

And there are some corollaries of this.

127 00:22:08,400 –> 00:22:20,320

The conjunction of two theories, which are mutually inconsistent, such as quantum theory and general relativity, Einstein’s relativity theory of gravity, that the conjunction there we know is false.

128 00:22:20,560 –> 00:22:24,760

And therefore, it has zero probability of being true, and yet it represents understanding the world.

129 00:22:24,760 –> 00:22:27,360

And it’s an improvement over the ignorance of past centuries.

130 00:22:27,760 –> 00:22:41,840

And a second corollary is that if our best theories will be superseded eventually, as we do believe, and therefore we believe today they’re negations, it is still those false theories, not their true negations that constitute all of our deepest knowledge of the world.

131 00:22:42,240 –> 00:22:50,000

So the moral of this is that science seeks to maximize explanatory power, not increase our credence or our beliefs, our subjective states.

132 00:22:50,880 –> 00:22:56,560

So our desideratum should be to create powerful but improbable theories, theories that say more about the world.

133 00:22:57,280 –> 00:23:11,680

Now, with having, I think, we can discharge Bayesianism from our epistemology, we can begin to rationally conjecture solutions to problems in quantum mechanics, and such explanations will empower us to build revolutionary quantum technologies.

134 00:23:12,000 –> 00:23:24,080

So quantum mechanics is the physics of the atomic and subatomic worlds, and understanding and using the power of this fundamental level of nature has transformed civilization scientifically, technologically, philosophically,

135 00:23:24,720 –> 00:23:47,960

We have only begun to explore and expound it, and with NoME for quantum mechanics, call it QUANToME we shall construct a novel and potentially revolutionary theory, subsuming quantum mechanics within NoME and thereby unify it surprisingly but beautifully and potentially profoundly with such seemingly disparate domains as linguistics and biomimicry, perchance even illuminating the mystery of consciousness.

136 00:23:47,960 –> 00:23:52,520

To begin, let us return to black holes in fact.

137 00:23:54,720 –> 00:24:03,600

I first cast my thought into black holes, as it were, studying under the mathematician Sir Roger Penrose at Oxford, where I first debated him in this legendary pub.

138 00:24:04,160 –> 00:24:13,680

He was awarded the 2020 Nobel Prize in Physics for the discovery that black hole formation is a robust prediction of the general theory of relativity, which is

139 00:24:14,080 –> 00:24:15,920

Einstein’s theory of gravity.

140 00:24:16,240 –> 00:24:33,680

So Penrose discovered mathematics of a sublime elegance to describe the shape that is the curved plane of space-time of black holes and explain why they must contain singularities, those points of infinite gravity and density where we know not how or if Einstein’s general relativity obtains.

141 00:24:34,080 –> 00:24:41,760

And from this mathematics was the, derived the idea of Penrose tiles, the shapes that cover or tile

142 00:24:42,360 –> 00:24:47,440

a surface or a plane such as space-time and certain specific and I think elegant patterns.

143 00:24:47,760 –> 00:24:51,120

These patterns are fractal all the way up and all the way down as it were.

144 00:24:51,680 –> 00:25:06,640

Thus, we can see that space-time is infinitely hierarchically structured complexity, such as one would find in Busan architecture, formally aesthetically analogous to the tilings of medieval Islamic art.

145 00:25:07,440 –> 00:25:18,000

And this is one of the most profound import of this discovery was that these fractal patterns occur everywhere can be composed and decomposed by beautiful sets of simple rules.

146 00:25:18,400 –> 00:25:28,840

And deducing such complexity from simplicity is the modus operandi of Oceanit philosophy and science, which Einstein himself so eloquently expressed as the

147 00:25:29,560 –> 00:25:36,400

The supreme task of the scientist is to arrive at the elementary laws from which the cosmos can be built up by pure deduction.

148 00:25:36,640 –> 00:25:46,240

So in Ushinetian terms, our task with quantum will be to arrive at elementary rules from which the cosmos, which is quantum mechanical, can be built up by pure deduction.

149 00:25:46,640 –> 00:25:51,360

So the intellectual bouquet of this pen rose walked it into my memory some months ago.

150 00:25:51,720 –> 00:26:02,480

as I was revamping the logical and mathematical rules whereby NoME generates linguistic structures, which are fractal in conforming to the Fibonacci sequence and the golden ratio, which I discussed last year at this festival.

151 00:26:02,880 –> 00:26:07,600

And the rules constitute a grammar, and the revamped grammar generates graphs.

152 00:26:07,920 –> 00:26:10,480

So let us proceed to the Platonic heaven.

153 00:26:11,120 –> 00:26:21,680

And we can define a hypergraph, which is equivalent to a syntax tree in natural language, simply as a set of abstract elements or hypernodes and their abstract relations or hyperedges.

154 00:26:22,080 –> 00:26:29,920

And to these graphs or syntax trees, we can apply grammars, that is, sets of rules like those that Penrose devised for tiling.

155 00:26:30,240 –> 00:26:34,640

For instance, here’s the rule that can be represented as a transformational grammar of graphs.

156 00:26:35,280 –> 00:26:47,880

And in this spirit, I was playing with one of my natural language grammars to determine whether it could generate Penrose-like tiles of Hofstadter’s butterfly, this fractal object that flutters betwixt general relativity and quantum mechanics.

157 00:26:47,880 –> 00:26:48,360

And it could.

158 00:26:48,360 –> 00:26:53,040

And hence, general relativity and quantum mechanics could be subsumed within QUANTNoME

159 00:26:53,840 –> 00:27:19,280

and in a unified mode of explanation of constructor theory, the gist of which is to explain physics not in terms of initial conditions and laws of motion, which is the standard mode, but rather in terms of calculations, first formalized by the great Alan Turing, which in physical terms means possible and impossible constructions or transformations, effectively calculating the universe with a linguistic Turing machine, which is itself a transformational Chomskian grammar.

160 00:27:19,720 –> 00:27:22,800

And this is the idea that was the birth of QUANTOAM.

161 00:27:23,760 –> 00:27:29,440

And here is one of QUANTOAM’s grammatical rules that can be written in algebraic form.

162 00:27:29,680 –> 00:27:50,640

Rewrite any A with BBB, rewrite any BB with an A, for instance, given A as an input, QUANTOAM generates this graph, which represents some physical system, a QUANTOAM system, say, where the blue represents strings, the state to the system, and the yellow represents the application to the grammar, the event to the system,

163 00:27:51,040 –> 00:27:56,400

And most importantly, and unique to our grammatical approach, we can represent causal relations, which are the orange lines.

164 00:27:57,000 –> 00:28:05,000

And the graph represents the complete system of all possible paths of history or evolution, and the complete system of causal relations within and between these paths.

165 00:28:05,000 –> 00:28:17,920

And this is where the QUANTOAM magic emerges, because the graphs that can be projected from this grammatical derivation can be Penrose tiled, and this is the shape that is generated corresponding to the different histories or evolutions.

166 00:28:18,320 –> 00:28:25,600

Now, the interpretation of QUANTOAM mechanics that naturally emerges from this model is that of the so-called many worlds interpretation.

167 00:28:25,920 –> 00:28:34,240

Intuitively, if a wee bit inaccurately, we can picture these possible grammatical computations as possible physical worlds or parallel universes, if you like.

168 00:28:34,720 –> 00:28:39,120

This is necessary to explain otherwise unexplainable quantum phenomena.

169 00:28:39,840 –> 00:28:55,040

The mystics out there who deny the many-worlds theory say that literally and vaingloriously, quote, do not endeavor to explain what appears impossible or the shut up and calculate method, shut up and calculate predictions, do not endeavor to explain the weirdness of quantum mechanics.

170 00:28:55,440 –> 00:28:57,920

However, we at Oceanit are critical rationalists.

171 00:28:57,920 –> 00:29:01,520

We’re not satisfied intellectually or aesthetically with mere predictions.

172 00:29:01,520 –> 00:29:06,480

On the contrary, in that Galilean mode, we seek explanations of the seen in terms of the unseen.

173 00:29:06,800 –> 00:29:10,000

and are logically committed to the deductive implications of our theories.

174 00:29:10,400 –> 00:29:18,400

And quantum theory logically implies the initially counterintuitive conclusion that all possible histories of everything and the mathematics of quantum substantiate this.

175 00:29:19,080 –> 00:29:27,680

And the set of parallel universes can converge in specific, they can converge in special conditions of so-called interference.

176 00:29:28,160 –> 00:29:33,440

And here is the fun of a grammar that generates a simple multiverse, which you can

177 00:29:33,760 –> 00:29:37,160

used with Penrose tiles and constructed and formed and understand its geometry.

178 00:29:37,160 –> 00:29:38,080

It’s very beautiful.

179 00:29:38,080 –> 00:29:43,280

So the many worlds theory emerges quite naturally from quantum’s graph grammars.

180 00:29:43,560 –> 00:29:48,560

And this has practical implications for quantum technology, of particular importance for quantum computation.

181 00:29:48,960 –> 00:29:53,640

Suffice to say that a general purpose quantum computer with its qubits, where a qubit is a

182 00:29:54,240 –> 00:30:03,600

zero in one universe and one in another, this could perform parallel calculations that are physically intractable for a classical computer where bits are either zero, one.

183 00:30:03,920 –> 00:30:13,200

The tricky bit, pardon the pun, is to choreograph the computation so that the parallel universes compute a single output in ours.

184 00:30:13,680 –> 00:30:20,800

And this requires a grammar, including natural language, as in fact has been discussed in some recent research.

185 00:30:21,120 –> 00:30:21,520

So

186 00:30:22,720 –> 00:30:28,400

In short, there’s a virtual wonderland of theories and technology that could be created with quantum technology, to which I’ll return.

187 00:30:28,720 –> 00:30:34,720

But how are we going to map the abstractions of QUANTNoME’s graph grammars to concrete QUANTNoME computation?

188 00:30:35,040 –> 00:30:40,240

Ideally, there would be some system that QUANTNoME could mimic to choreograph and its graphs.

189 00:30:40,480 –> 00:30:42,520

And I can generate such systems do exist.

190 00:30:48,680 –> 00:30:52,800

So of late, at Oceanit, we’ve been discussing biomimicry in the context of fungi.

191 00:30:53,040 –> 00:31:03,200

And I’ve become particularly besotted with these mycelial systems, these labyrinthia networks of subterranean fungi, which interconnect and literally enliven entire ecosystems of flora.

192 00:31:03,760 –> 00:31:15,920

To quote one scientist, by linking to the fungal networks, trees and plants can help out their neighbors by sharing nutrients and information or sabotage unwelcome plants by spreading toxic chemicals to the network.

193 00:31:16,960 –> 00:31:24,160

And indeed, these computational properties of this brain-like system would be the basis for the biomimicry of QUANTNoME.

194 00:31:24,800 –> 00:31:31,920

To quote another scientist, these fungal networks make communication between plants, including those of different species, faster and more effective.

195 00:31:32,320 –> 00:31:35,720

And these fungal systems are fantastically fractal.

196 00:31:36,480 –> 00:31:43,360

That is, these mycetal networks are gorgeous and graphical, and indeed, QUANTNoME can generate their graphical structure.

197 00:31:44,440 –> 00:31:45,120

It’s pictured here.

198 00:31:45,480 –> 00:31:58,560

And in fact, here’s where the mushrooms truly become magical in that the mycelial graphs are formally equivalent to the many worlds graphs of cloud mechanics in that continuing this derivation generates that shape which once Penrose tiled approximates the multiverse.

199 00:31:58,960 –> 00:32:04,080

So this is the conjecture that mimicking the choreography of information in the mycelial system

200 00:32:04,760 –> 00:32:15,040

which we know is fast and effective, could enable and enhance the performance of QUANTNoME systems at the software and/or hardware level, that is, quantum algorithms and/or quantum hardware.

201 00:32:15,360 –> 00:32:21,200

And as a toy model, a toy model could be a cellular automaton, the game of life.

202 00:32:21,440 –> 00:32:29,680

We could more efficiently and effectively compute the dynamics in my serial life with Penroe tiles, and we could mimic and transform these computations in QUANTNoME.

203 00:32:30,400 –> 00:32:33,880

to be efficiently and effectively equivalent to grammars.

204 00:32:33,880 –> 00:32:43,280

So, I shall work, in fact, it’s quite certain that only to this work that I discovered that Star Trek had independently converged on this idea.

205 00:32:45,360 –> 00:32:59,720

So, in the recent Star Trek series, there is this myosquial network, which is this discrete subspace that permeates the multiverse and which enables them to traverse an infinite number of roads leading everywhere, span the entire multiverse, that includes a known universe and all of the quantum realities.

206 00:32:59,840 –> 00:33:06,880

So sometimes science, Stephen Hawking said that today’s science fiction is tomorrow’s scientific fact.

207 00:33:07,440 –> 00:33:08,640

And it’s, I think it’s quite true.

208 00:33:08,640 –> 00:33:12,560

So I think that with

209 00:33:12,960 –> 00:33:31,440

One final conjecture would be to, the final pie in the sky, if you will, or a final mushroom in the sky, is in designing QUANTNoME, the computational system biomimicking fractal fungi by generating these graph theoretic structures via grammars, it may be the solution to the problem of understanding consciousness and recreating it in silicon.

210 00:33:31,720 –> 00:33:35,760

In other words, the computational implementation of QUANTNoME could be a conscious entity.

211 00:33:37,240 –> 00:33:40,080

And there are three reasons to conject that this is possible.

212 00:33:40,400 –> 00:33:47,200

First, as Penrose and I debated in that pub, the explanation for how consciousness emerges in the human brain

213 00:33:47,680 –> 00:33:52,320

could be quantum mechanical, as Penrose has argued, and I do not disagree.

214 00:33:52,320 –> 00:34:01,760

I differ only in what those quantum mechanical mechanisms, oh, I shan’t go into that here, but we do agree that these mechanisms are molecular, that is intra-neuronal or sub-cellular.

215 00:34:02,240 –> 00:34:07,120

And the second reason for this consciousness conjecture is not crazy.

216 00:34:07,120 –> 00:34:17,560

It’s the idea that these vast mycelial systems in nature may themselves already be conscious in some limited sense, as some scientists and philosophers have argued, for they can be

217 00:34:17,720 –> 00:34:27,360

as complex as brain and may use quantum mechanical processes, which we do know are involved in some biological processes such as photosynthesis and possibly avian navigation.

218 00:34:28,000 –> 00:34:33,840

And the third reason, third final reason, is that as I’ve argued, neurobiology hardware is not self-sufficient for consciousness.

219 00:34:34,160 –> 00:34:47,840

What is required is a linguistic Turing machine equivalent to a constructor, a system that can itself perform non-random trials and eliminate errors in this cybernetic dialectical feedback of what is effectively conjecture and criticism, and thereby construct a notion of self.

220 00:34:48,280 –> 00:34:52,400

Now, I’m not saying that these systems, say, fungi need to be able to talk to themselves.

221 00:34:52,720 –> 00:34:57,280

They do need to be able to generate graphs that enter into this dialectical cybernetic process.

222 00:34:57,280 –> 00:35:01,920

So I conjecture that the mathematical architecture of mycelial systems will enable them to do so.

223 00:35:02,520 –> 00:35:13,440

And so with biomimicry, QUANTNoME could be conscious too, in that its mathematical representation and computational implementation would mimic the computational, the quantum computational choreography of this consciousness.

224 00:35:14,600 –> 00:35:17,280

But at present, we also begin with some simpler problems.

225 00:35:17,280 –> 00:35:24,800

So one of the so-called revolutions of deep learning has been the solution to the folding problem of proteins.

226 00:35:25,200 –> 00:35:33,520

And as we all know all too well, the functionality of proteins, such as spike proteins, is determined by the three-dimensional structure into which they fold.

227 00:35:33,920 –> 00:35:36,880

It is an exceedingly complex problem to model and predict.

228 00:35:38,240 –> 00:35:43,360

But a general solution to it would indeed be revolutionary, deepening our understanding of the fundamental dynamics of life

229 00:35:43,800 –> 00:35:50,640

and enabling us to design new therapeutics and vaccines, supercharged synthetic biology, and many other applications.

230 00:35:50,640 –> 00:35:54,960

Now, this year it was reported that Google solved the problem, but that is not quite correct.

231 00:35:55,440 –> 00:36:06,240

It did solve an important problem, but not the general problem of computing the 3D folded structure of a protein from its amino acid sequence and the inverse of computing the sequence for a given structure.

232 00:36:06,240 –> 00:36:07,120

The reason it failed

233 00:36:08,160 –> 00:36:13,120

we submit is because a deep learning neural network is the wrong kind of AI for the task.

234 00:36:13,440 –> 00:36:16,640

The task is, in reality, linguistic, we would argue.

235 00:36:16,640 –> 00:36:31,360

That is, with our anthropic AI, with NoME, we can explain the complexity of protein folding in terms of this Chomsky hierarchy of formal grammars, where we can analyze the operations and structures as mathematically analogous to those that build the syntax of our sentences in proteins.

236 00:36:31,880 –> 00:36:40,480

And if we succeed in discovering and designing this grammar, we could generate those bespoke proteins required for novel medical interventions and synthetic biology.

237 00:36:41,480 –> 00:36:54,880

More profoundly, we could explain as we can linguistics and not really describe or predict as one can deep learning, what are the possible and impossible, grammatical and ungrammatical, both here on Earth and potentially anywhere in the universe, a multiverse.

238 00:36:54,880 –> 00:36:55,280

That is,

239 00:36:55,720 –> 00:36:57,640

We could discover fundamental laws of biology.

240 00:36:57,640 –> 00:37:04,400

For instance, we could distinguish life from non-life by the complexity of the information necessary to generate that structure.

241 00:37:04,720 –> 00:37:09,360

Life, like intelligence and language, generate structures that would not otherwise exist.

242 00:37:09,360 –> 00:37:11,600

We call it my transmutation of gold example.

243 00:37:12,440 –> 00:37:20,440

Indeed, we can see here the profound consilience of life, intelligence and language that could be realized in our anthro-analytic artificial intelligence.

244 00:37:20,440 –> 00:37:30,480

And so returning to my original proposal with the Cambridge Institute for Exo- language, CIEL, the message we ought to beam into the cosmos is the code, I think, for our anthro-analytic AI.

245 00:37:31,200 –> 00:37:37,200

And this is because such a message would communicate the essence of us, which is not what we think, but it’s how we think.

246 00:37:37,520 –> 00:37:46,480

How we think, which is how we make discoveries, how we create new knowledge, is a necessary fact of our biology and the laws of nature that determine it, the fact that we are linguistic Turing machines.

247 00:37:46,880 –> 00:38:01,040

So we ought to send a theory or a causal model of our linguistic competence to create new objective knowledge, new problems, new arguments, new solutions, new theories, new works of art, via syntactic structures such as that animated here, and the form this theory assumes ought to be

248 00:38:01,160 –> 00:38:01,760

computable.

249 00:38:02,000 –> 00:38:05,240

It ought to generate the explanation explicitly and deterministically.

250 00:38:05,240 –> 00:38:08,560

It would be in a sense, therefore, self-explanatory.

251 00:38:08,960 –> 00:38:11,440

And that describes a computer program.

252 00:38:12,000 –> 00:38:26,000

For instance, here the program, here’s the program for one of NoME’s linguistic rules represented as a cellular automaton, a computer program encoding an explanatory theory of thinking that defines, you know, this defines when anthropy-like AI is, such as NoME

253 00:38:26,560 –> 00:38:28,720

and the artificial mind we’re creating here at Oceanit.

254 00:38:28,720 –> 00:38:37,760

But obviously, this message to the cosmos would be in 2024, will not yet be a mature AI.

255 00:38:38,320 –> 00:38:45,520

But even the rudiments of an embryological NoME, if you will, would suffice to communicate the ligaments of our intelligence or nature.

256 00:38:46,160 –> 00:38:56,800

An example of such a quidity of our linguistic competence is a mathematical formulation of the Platonic Aristotelian distinction between the possession and use of linguistic knowledge.

257 00:38:57,600 –> 00:39:04,400

Our toy model here is Euclidean geometry, which is mathematically analogous in a structure to natural language.

258 00:39:05,080 –> 00:39:11,360

So Euclid’s elements contain elaborate proofs, some such as the proof of the Pythagorean theorem.

259 00:39:11,920 –> 00:39:18,720

Now, here are some of our recent work where we demonstrated that the linguistic capacity of NoME to generate all the proofs in Euclid’s geometry.

260 00:39:19,520 –> 00:39:27,280

In Platonic and Aristotelian epistemology, the set of proofs, this entire set of proofs represents the knowledge we possess.

261 00:39:27,600 –> 00:39:30,480

Knowledge we can use is a subset.

262 00:39:30,720 –> 00:39:34,800

It’s the subset that we can construct stepwise that is efficiently provable.

263 00:39:35,680 –> 00:39:43,280

And because in our theory, thought necessarily assumes A linguistic form, the mathematical knowledge is ultimately a species of linguistic knowledge.

264 00:39:44,240 –> 00:39:50,560

So for instance, here’s the knowledge we possess of the proof of Euclid’s finest theorem, the uniqueness of the Platonic solids.

265 00:39:51,160 –> 00:39:56,240

And here’s the knowledge we can use, the set of possible proofs we can construct to derive that theorem.

266 00:39:56,720 –> 00:40:04,480

So literally the most informative message we could send to the extraterrestrials would be one that encodes the grammars our minds possess to use

267 00:40:05,040 –> 00:40:21,760

and that they use to generate such explanatory knowledge for the structure of proofs is generalizable to all forms of knowledge, which is why it is so important technologically here on Earth that progress in designing such a message is ipso facto progress in designing AI, which is the mission of our AI research at Oceanit and at CIEL.

268 00:40:22,400 –> 00:40:23,040

And incidentally,

269 00:40:25,040 –> 00:40:32,400

how many of you might have noticed that the constitutive shapes of the known logo are in fact the Platonic solids, not by accident.

270 00:40:32,800 –> 00:40:41,280

So what I’ve said of mathematical proofs and linguistic derivations obtains equally, for instance, of the assembly, as I said, of the assembly of biomolecules.

271 00:40:42,160 –> 00:40:52,320

But with that, we can transcend to convey notions of ourselves because anthropic AI, sending out into cosmos is to send a blueprint of ourselves because

272 00:40:53,000 –> 00:40:55,160

Not our physical selves, but our mental selves.

273 00:40:55,160 –> 00:40:58,440

For our minds, our software that run on the hardware of our brains.

274 00:40:58,440 –> 00:41:01,280

You are a mind, you are a brain, you are information.

275 00:41:01,680 –> 00:41:10,640

And because of anthronoetic AI, even the sketch of one, we can imagine sending in 2024, is simply the informational structure of a species' typical human mind.

276 00:41:10,960 –> 00:41:14,680

The message would seed humanity into the structure of the cosmos itself.

277 00:41:15,680 –> 00:41:22,960

And this area metaphysics is grounded in outrageous physics theory from none other than the old master, Sir Roger Penrose.

278 00:41:23,440 –> 00:41:29,200

And as of last week or so, he’s officially a member of our scientific advisory committee for CL.

279 00:41:29,680 –> 00:41:34,240

And his self-styled outrageous perspective, conformal cyclic cosmology,

280 00:41:34,720 –> 00:41:48,080

looks upon our universe as, quote, but one eon of an ending succession of expanding eons, each having an initial big bang, which is the conformal continuation of the remote exponential expansion of its previous eon.

281 00:41:48,480 –> 00:41:49,760

So in this theory,

282 00:41:50,480 –> 00:42:08,840

Like in Escher’s woodcut, infinity can be understood as a conformal representation of hyperbolic geometry, whose complexities I shall expound here, but which is in fact the kind of infinity most natural to our anthropoetic artificial intelligence, whose linguistic competence is based on the infinite use of finite minds.

283 00:42:08,840 –> 00:42:12,160

And here, the boundary of the circle represents infinity.

284 00:42:12,160 –> 00:42:14,560

It is conformal in the sense that angles

285 00:42:14,920 –> 00:42:20,240

here of the creatures are preserved whilst sizes can vary however you fancy.

286 00:42:20,560 –> 00:42:26,720

Now, Sir Roger sent me these slides of old-school transparencies that he still uses in lectures.

287 00:42:27,120 –> 00:42:41,840

And he proves with some lovely mathematics that the infinity in the singularity of the Big Bang, our distant past, is conformal to the infinity in our distant future as the universe continues to expand exponentially, from which it follows as

288 00:42:42,320 –> 00:42:53,120

Roger proves with some even lovelier mathematics that our big bang was the end of a preceding eon of the universe and that the end of our eon will be the big bang of the succeeding eon of the universe.

289 00:42:53,360 –> 00:43:02,080

Now, I shan’t go any deeper into the complexities of the form of geometry from which this theory derives, but suffice to say, the mathematics is sound and the physics it entails is not implausible.

290 00:43:02,600 –> 00:43:05,520

and in fact is corroborated empirically.

291 00:43:06,000 –> 00:43:09,120

And dark matter enters into this evidentiary basis, interestingly.

292 00:43:09,520 –> 00:43:25,600

And the relevance of this theory to CIEL, our message is that as Sir Roger discussed in his Nobel lecture last year, information from one eon could, if properly encoded, for instance, as photons or gravitational waves, pass into the next eon ad infinitum.

293 00:43:25,960 –> 00:43:31,040

And indeed, it is conceivable we could send a message with the technology of the breakthrough

294 00:43:31,800 –> 00:43:36,240

breakthrough star shot project, with whom we could collaborate at CIEL.

295 00:43:37,040 –> 00:43:51,680

It follows, therefore, that our message, this AI, the fingerprint of the human mind, as it were, could ring into eternity and seeding the future with the information constituent of us, a kind of information panspermia, if you will.

296 00:43:52,080 –> 00:43:59,760

Even Sir Roger and his collaborators have conjectured that it would not be so outrageous to search for evidence of extraterrestrial life from past eons,

297 00:44:00,160 –> 00:44:01,640

in the physics of this young.

298 00:44:01,640 –> 00:44:06,800

And to that end, he and his colleagues are joining us at CIEL.

299 00:44:07,200 –> 00:44:21,840

And it is even conceivable that we could append to the message encoding, the AI, a message encoding instructions for what’s called a von Neumann machine, a universal constructor to build the AI to recreate ourselves, but that’s perhaps too outrageous even for this presentation.

300 00:44:22,240 –> 00:44:23,680

So to conclude,

301 00:44:24,400 –> 00:44:32,720

The work of CIEL and Oceanit will explore the final frontier of AI and cosmology with near-term and far-out implications.

302 00:44:32,720 –> 00:44:48,400

The former includes applications from artificial general intelligence to computational approaches, understanding life, cancer, quantum mechanics, and it’s the, but it is the outrageous possibilities that are the most inspiring, truly the most Oceanitian, I submit, in recognizing no limits to our imaginations.

303 00:44:48,400 –> 00:44:48,720

Indeed,

304 00:44:49,280 –> 00:45:01,200

Should the most outrageous conjecture of information panstermia prove true, then our minds would be like Beethoven’s music, of which it was said that when everything passes and the world perishes, the Ninth Symphony will remain.

305 00:45:50,080 –> 00:45:50,480

Finito.

306 00:45:53,920 –> 00:45:54,440

Excellent.

307 00:45:54,440 –> 00:45:56,280

Thank you very much, Dr.

308 00:45:56,280 –> 00:45:58,160

Watumull, for the presentation.

309 00:45:58,160 –> 00:46:02,240

And we would like to open it up for any questions that people would have.

310 00:46:02,800 –> 00:46:04,080

We have a few minutes here.

311 00:46:04,080 –> 00:46:04,720

We can

312 00:46:05,280 –> 00:46:05,840

Bring it on.

313 00:46:08,240 –> 00:46:09,600

I enjoy the fencing.

314 00:46:10,240 –> 00:46:11,760

Entertain questions.

315 00:46:13,040 –> 00:46:15,200

In the meantime, I’ll ask one.

316 00:46:16,560 –> 00:46:33,040

So given that we want to, your discussion is about sending out a blueprint of ourselves, a blueprint of our minds, our language, and that sort of thing.

317 00:46:34,320 –> 00:47:03,120

Does this presuppose that there are entities out there that can view the universe in terms of blueprintable material, in terms of blueprintable things, that there is somehow some kind of order out there that the blueprint will make sense in?

318 00:47:04,160 –> 00:47:10,080

So the mission or the philosophy does not presuppose that any entities out there do exist.

319 00:47:10,360 –> 00:47:33,360

And when I, as a footnote, when I said that it would in fact be more important to send this message if there were no entities, the irony or the paradox there is that in fact, if we are the only creatures in this universe that create meaning, it’s therefore all the more imperative that we maximize the amount of meaning in the universe because we’re the only objects or constructors in the universe that are creating meaning.

320 00:47:33,360 –> 00:47:33,600

But

321 00:47:34,480 –> 00:47:40,160

Suppose, though, that there are extraterrestrial intelligences out there that could receive such a message.

322 00:47:40,320 –> 00:47:43,200

Yes, could they make sense of it?

323 00:47:43,680 –> 00:47:47,360

Is it not anthropocentric to presume that they would understand us?

324 00:47:47,840 –> 00:48:00,240

The conjecture here is that the properties of natural language in which this message would be encoded, the properties are fundamentally mathematics or mathematical, and mathematics is universal.

325 00:48:00,240 –> 00:48:01,600

The laws of mathematics are universal.

326 00:48:01,600 –> 00:48:03,120

That arithmetic does not differ

327 00:48:03,560 –> 00:48:07,040

on Alpha, it doesn’t differ in Alpha Centauri than it does here on Earth.

328 00:48:07,840 –> 00:48:17,760 2 plus 2 equals 4 everywhere in the universe has a law of nature, such that any intelligence anywhere in the universe will have converged on its evolution notions of mathematics.

329 00:48:17,760 –> 00:48:23,680

But even beyond mathematics, and so therefore you might expect what we would predict, conjecture that it’s

330 00:48:24,480 –> 00:48:31,200

language would, and its form of mentation and cognition, would resemble ours in its formal mathematical structure.

331 00:48:31,400 –> 00:48:37,360

And in fact, for those of you that are interested, you can Google a paper by myself and Chomsky called Rethinking Universality.

332 00:48:37,960 –> 00:48:43,920

where we discuss this very notion that intelligence will have converged on our form of natural language.

333 00:48:44,000 –> 00:48:53,040

But this message, as I was describing in the presentation, would not itself be semantic in that we would not be communicating or enumerating particular discoveries we have made.

334 00:48:53,040 –> 00:48:54,080

It would just be the

335 00:48:54,320 –> 00:48:58,240

the principles and the procedures by which we construct any particular expression.

336 00:48:58,240 –> 00:49:02,880

So it wouldn’t be any particular sentence, it’d be the method that we construct sentences.

337 00:49:03,160 –> 00:49:08,120

And thereby understand the method by which we are creative, not necessarily what we have created.

338 00:49:09,200 –> 00:49:10,320

We do have a question.

339 00:49:10,480 –> 00:49:12,080

We have a question from Bill Tam.

340 00:49:12,480 –> 00:49:14,160

Hello, Bill, thank you for asking.

341 00:49:14,960 –> 00:49:22,080

William Tam asks, how do photons survive the black hole singularity that destroys all information?

342 00:49:23,440 –> 00:49:26,640

So that’s a, that’s a, as Sherlock Holmes said, that’s a three-pipe problem.

343 00:49:27,040 –> 00:49:34,680

So the first pipe, the first problematic pipe would be the question whether information can be destroyed.

344 00:49:34,680 –> 00:49:40,160

So Stephen Hawking originally in the 1970s thought that information that slipped into a black hole was destroyed.

345 00:49:40,440 –> 00:49:50,000

But then subsequently he revised his opinion and lost a bet and decided that no, information is not destroyed because

346 00:49:50,640 –> 00:49:58,400

In many theories of quantum mechanics, information needs to be preserved or predicted to be preserved, and so thereby information does escape from a black hole that radiates out.

347 00:49:59,680 –> 00:50:05,840

Now, I, in fact, I’m agnostic as to whether or not information can be destroyed or even created.

348 00:50:06,720 –> 00:50:15,440

Penrose does assume that information can be destroyed, but if the photons at the end of our universe, at the end of the universe,

349 00:50:15,840 –> 00:50:18,880

It’ll be a very, as he as Penrith described, a very boring place.

350 00:50:19,200 –> 00:50:21,840

There’ll be nothing but photons and black holes.

351 00:50:22,320 –> 00:50:39,120

And, but at this crossover from our eon to the next eon, some photons which do not slip into black holes will, can, you can prove, traverse that boundary, as could gravitational waves from black holes that are colliding at the end of the universe.

352 00:50:39,120 –> 00:50:42,400

So one could today, in theory,

353 00:50:43,040 –> 00:50:45,320

if one examines the cosmic microwave background.

354 00:50:45,320 –> 00:50:53,200

And for those of you that don’t know, the cosmic microwave background is the evidence of the Big Bang that we can observe today.

355 00:50:53,200 –> 00:50:54,960

It’s the baby picture of our universe.

356 00:50:55,440 –> 00:51:05,640

And the conjecture from Penrose is that photons, gravitational waves, those would be imprinted on the cosmic microwave background.

357 00:51:05,640 –> 00:51:08,160

So you could observe those signals today.

358 00:51:08,760 –> 00:51:25,360

And so you can imagine, and some of the work at CIEL would be to imagine how precisely do you encode a message in photons, in particles of light, to make a pattern onto the microwave background of the next eon that could be interpreted by any civilizations that are there.

359 00:51:25,600 –> 00:51:33,800

Or as I said, even if there are no civilizations there to appreciate it, like art, I don’t think art needs to be observed to be meaningful.

360 00:51:34,400 –> 00:51:36,160

Art is meaningful in and of itself.

361 00:51:36,480 –> 00:51:40,240

and our minds are fundamentally aesthetic, artistic objects.

362 00:51:40,240 –> 00:51:44,000

So imprinting that into the universe could be possible.

363 00:51:44,000 –> 00:51:46,480

But yeah, so there are multiple problems.

364 00:51:46,720 –> 00:51:50,480

It’s more than a three-pipe problem, but those are excellent questions.

365 00:51:51,920 –> 00:51:53,760

So just to make sure

366 00:51:55,600 –> 00:51:58,720

I understood something and for the benefit of everyone.

367 00:51:59,120 –> 00:52:18,960

When we talk about the mathematical structure and the universality of mathematics and mathematical language, and we apply it to this, I don’t know if you could call it the mushroom network or I think you call it the mycelial network.

368 00:52:20,240 –> 00:52:21,280

This then

369 00:52:21,920 –> 00:52:28,880

as you showed when you were diagramming the way we think as we create sentences, they all look very similar.

370 00:52:28,880 –> 00:52:36,320

Can you explain that a little bit and go back into that thing, connecting everything?

371 00:52:36,320 –> 00:52:43,200

And this, I believe you used the word conceal approach or conceal activity.

372 00:52:43,600 –> 00:52:43,880

Indeed.

373 00:52:44,160 –> 00:52:50,480

So the fundamental operations that generate syntactic structures, the structures of human sentences,

374 00:52:50,800 –> 00:52:54,680

They are recursive in that they can refer to themselves.

375 00:52:54,680 –> 00:53:01,280

So in a sentence of English, you can embed clauses within clauses.

376 00:53:01,520 –> 00:53:04,320

So the cat on the mat ate the dollar.

377 00:53:04,320 –> 00:53:07,920

And so you can embed them infinitely and they continually refer to themselves.

378 00:53:08,120 –> 00:53:10,640

You can generate very complex fractal structures.

379 00:53:10,920 –> 00:53:14,080

And fractals are the Mandelbrot set or the Fibonacci sequence.

380 00:53:14,360 –> 00:53:20,640

where there is self-similarity on multiple levels of complexity, all the way from micro scales to macro scales.

381 00:53:21,120 –> 00:53:34,520

And those same operations that build fractal structures in natural language, and fractals are interesting not merely because they’re aesthetically beautiful, but because they’re computationally optimal in the sense of maximizing the amount of information with minimal resources.

382 00:53:34,520 –> 00:53:36,960

And so they’re very Leibnizian.

383 00:53:36,960 –> 00:53:40,080

So Leibniz in 17th century thought that nature,

384 00:53:42,400 –> 00:53:47,280

maximizes beauty with the simplest possible means, infinite use of finite means.

385 00:53:47,600 –> 00:53:56,080

And so natural language conforms perfectly with that Leibnizian metaphysic, as does mycelial networks.

386 00:53:56,880 –> 00:54:08,160

But as I said, that universality of fractal geometry and recursive processes that refer to themselves is not only universal to natural systems, but to systems that we construct artificially, because as I said, it’s very efficient.

387 00:54:08,640 –> 00:54:09,320

And I think it’s the

388 00:54:10,000 –> 00:54:23,520

the secret or the key to consciousness too, because it’s self-reference, constructing a notion of the self that’s unique to humans based on this influence of competence, whereby you can use language to effectively talk to yourself, thereby creating a self.

389 00:54:23,520 –> 00:54:33,040

That’s why you could imagine this fractal recursive process operating across humans, mysterial networks, and beyond, obviously artificial intelligence.

390 00:54:34,080 –> 00:54:36,000

If I may interrupt, we have another question.

391 00:54:36,000 –> 00:54:38,160

Thank you, Ariana White.

392 00:54:39,600 –> 00:54:51,600

Ariana asks, how does the crossover into a new eon occur if the hypothetical fate of our universe is a state of maximum entropy where work cannot be done?

393 00:54:52,240 –> 00:54:59,520

Yes, so that is the, that was the, as Penrose termed it, the mammoth in the room.

394 00:55:01,040 –> 00:55:10,440

the problem of the second law of thermodynamics, which is the law of nature that states that entropy, which is disorder, increases over time.

395 00:55:10,440 –> 00:55:20,080

So the conjecture of the Big Bang is that the universe started out in a state, a very uniform state, and ever since has been increasing in disorder over time.

396 00:55:20,080 –> 00:55:23,360

So things degrade and dissipate over time.

397 00:55:23,360 –> 00:55:28,560

And eventually, yes, the universe would be in a state of maximum disorder, maximum entropy.

398 00:55:29,200 –> 00:55:31,920

But there is, it’s very complicated.

399 00:55:32,080 –> 00:55:43,120

And so the one interesting exception, local exception to the law of entropy, it is local in spatial and temporal terms, is life.

400 00:55:43,600 –> 00:55:58,640

Life is a process that, as Erwin Schrodinger proved or observed in the 1940s, life is a unique process in terms of rebelling against entropy for some time, in the sense that life, that’s objects, we and other life in the universe, is organizing

401 00:55:59,440 –> 00:56:02,000

information to pattern matter.

402 00:56:02,200 –> 00:56:04,440

And that’s, in fact, I think a good definition of what life does.

403 00:56:04,440 –> 00:56:13,840

And this is from the astrobiology of Sarah Walker, which is that what life is, life is information, abstract causation, patterning matter.

404 00:56:13,840 –> 00:56:22,640

So we’re organizing and temporarily rebelling against the second law, which is accelerating the increase of entropy and disorder over time.

405 00:56:23,040 –> 00:56:26,400

So life, which I think

406 00:56:26,800 –> 00:56:42,320

There’s an interesting unity, consilience, once again, of life, intelligence, and language, which is that by analogy, so if one is interested in the laws of, say, gravitation, one studies black holes, because that is where gravity is most intense.

407 00:56:42,960 –> 00:56:53,600

If one is interested in the laws of life, or if one is interested in the laws of information, as an abstract causation that’s patterning matter, then one should study light, because that’s where information is most intense.

408 00:56:54,080 –> 00:56:57,040

And where life is most intense is, I think, intelligence.

409 00:56:57,360 –> 00:57:00,240

And where intelligence is most intense is language.

410 00:57:00,920 –> 00:57:04,400

And so with language, it’s the ultimate reversal of entropy.

411 00:57:04,640 –> 00:57:08,560

So we’re creating these bubbles of order in a sea of chaos.

412 00:57:08,560 –> 00:57:20,480

But yes, over time, so one reason or motivation, yes, to propagate messages into the future would be to propagate bubbles of order and non-randomness and meaning.

413 00:57:21,040 –> 00:57:22,520

to the end of the universe and into the next.

414 00:57:22,560 –> 00:57:27,720

But yes, the precise physics of how, because at the end of the universe, it will be maximum entropy.

415 00:57:27,720 –> 00:57:29,880

But so this gets quite technical.

416 00:57:29,920 –> 00:57:41,520

So for those of you that we should discuss this online over a bottle of wine, but very briefly, at the end of the universe, all the entropy, all the entropy, all the disorder, vast majority of it will be in black holes.

417 00:57:42,160 –> 00:57:47,960

Even today, in our state of the universe at present, the overwhelming amount of entropy is in black holes.

418 00:57:47,960 –> 00:57:51,680

And so at the end of the universe, there’ll basically only be black holes and photons.

419 00:57:51,920 –> 00:58:07,360

So that begins to explain how at the crossover, the initial state of the next ion, its big bang, will be in a state of low entropy because black holes eventually evaporate by Hawking radiation.

420 00:58:07,400 –> 00:58:08,600

They pop out of existence.

421 00:58:08,600 –> 00:58:10,520

And so you’ll have a very uniform state.

422 00:58:11,120 –> 00:58:16,560

with photons, which photons move at the speed of light, so they have no mass.

423 00:58:16,800 –> 00:58:24,920

And so you begin the universe as you did at our big bang, no mass, massless particles and life, the universe and everything.

424 00:58:24,920 –> 00:58:28,560

But we should send me an e-mail and we’ll discuss this a little bit.

425 00:58:28,560 –> 00:58:30,000

I enjoy Zoomposia.

426 00:58:30,240 –> 00:58:33,320

So, you know, Plato had his symposia where they have wine, intellectual debate.

427 00:58:33,760 –> 00:58:38,400

Alas, with the pandemic, we’re restricted perhaps from a symposium, but a Zoomposium

428 00:58:38,880 –> 00:58:41,280

I would delight in some dialogue.

References

• Jeffrey Watumull, A Turing Program for Linguistic Theory, lingbuzz/001550, Jun., 2012.

• Jeffrey Watumull, Language Is a “Quite Useless” Tool: A Rejoinder to Fedorenko, Piantadosi, and Gibson’s “Language Is Primarily a Tool for Communication Rather Than Thought”, lingbuzz/008535, Oct., 2024.