Atlantis Dispatch 011:
in which ATLANTIS contemplates intelligent life…
“Triptych, Inspired by the Oresteia of Aeschylus.” Francis Bacon. 1981. © The Estate of Francis Bacon / DACS London 2020. All rights reserved.
October 15th, 2021
…begin transmission…
If you’ve been following the Atlantis Dispatches for a while now, you may have noticed a familiar pattern: when we send things out into space, we often do so in order to look back at ourselves. Are we totally self-absorbed? Or is there something about the view from space that makes us see more clearly
The late September launch of the Landsat 9 satellite is no exception to the pattern. NASA has been shooting out self-regarding satellites since 1972. The program is in concert with the U.S. Geological Survey and is designed, in part, to monitor Earth’s geology. It’s equipped with spectral sensors that help to detect minute changes in lakes, for instance, and allows us to survey the changing face of Earth’s forests, too. So, all this time, in the background, we’ve had this viewfinder running its reflective magic on our whole ecological world.
In light of this view, Atlantis experienced something of a swerve. We found ourselves wondering not only about the whole living ecosystem that forms Earth, but also about the strange (supposedly) intelligent creatures, who lately have been rearranging our planet in destructive ways. Looking down from the Landsat 9, we wondered about the intelligence at the core of this life. Are humans really intelligent? Could we identify intelligence if we came upon it? Perhaps the hunt for intelligent life is as difficult and fraught as the hunt for life itself!
Atlantis set sail to Silicon Valley, where the loudest clamors and claims about human intelligence rang out. From as far as the Bay we could hear the battle cries. Some said, “the greatest human brains, find ye-selves here!” Some said, “the human brain is such a #fail compared to an AI!” Some said, “most humans are morons!”
Now, Atlantis will admit, once we approached, we didn’t entirely want to wade into the quagmire of a humanoid smackdown—we might goop up the ship. In fact, when we thought about it, we wondered if we might do even worse than goop up the ship—maybe this great battle would set us afire like the Greeks who sailed to Sicily. Were we wading right into the eye of a war? And if so, was it our place to investigate or intervene? Is Atlantis a military vessel? We quivered and quavered and went to the hull to reflect upon Atlantis’s mission in this mad world. What is Atlantis doing here? What is the nature of Atlantis? What is Atlantis at all?
Well, as you may recall, our ship takes its name from that old city, the one that sunk for its hubris, and rose again in the form of Francis Bacon’s island, that utopia populated by scientists. Clearly, our Atlantis, this old and new Atlantis, is both Platonic and Baconian (no wonder we’re so often divided!). Well, what exactly does that mean? With respect to being Platonic, we’re kind of philosophical, maybe even a bit formal about things. We like the deep patterns of math and science, and we like the big questions even more.
As for being Baconian, well, reader, let us say this: we’re Baconian in the true sense of the word. The oft forgotten “Baconian,” has suffered from some low-fidelity cultural copying…but hey, at least it still persists.
Allow us to digress. Now, we’ll admit that old Frankie Bacon has a bit of a reputation, and well earned. He’s known as both the brilliant founder of modern science and an evil advocate for the conquest of nature. We think that he helped us immensely in the former respect (and maybe his reputation should be expanded in this regard); we think that in the latter respect he’s somewhat misunderstood (and our historical memory should be refined and aligned with what the man writes).
Bacon understood that thinking clearly about ourselves would require not only book learning, but also learning out in the world—technological learning, especially. He also knew that applying technology to the world could help us liberate ourselves—if, and only if, we are able to guard against the inevitable dangers that tech harbors.
In our next venture, we would have to be both curious and cautious, like Odysseus, sailing close past the Sirens. As you may recall, Odysseus tied himself to the mast so as not be drawn in to the island of discarded bones that the Sirens created, but kept his ears free in order to hear their strange music—a perfect synthesis of investigation and restraint.
Like our Atlantis, Bacon thought that humans would always use technology, and so they would need to learn to use it well. Bacon was not into using technology to destroy ourselves, though he knew it could. We are Promethean after all, so we’d better learn to handle fire.
How would we know how to do this? In part, by anticipation; in part, through experience. When Bacon wrote his New Organon, he gave us a brave new way to think about experiential learning: he introduced the idea of what we now call “experiment,” that structured kind of experience that is designed to yield insight. For Bacon, experimentation was not just a matter of tinkering, it was a way to become clearer about what our experience with new things actually reveals to us. We can’t know things apriori, especially when what we are trying to understand is something novel.
In short, Atlantis is Baconian in this sense: we are voracious; we venture forth into the world; we experiment properly; we guard against the dangers of our power; we’re a ship, not an ivory tower. Lo, that rhymes!
In this spirit, we asked ourselves, should we go forth into the great war over intelligence? A battle was certainly raging, and though we didn’t want to get caught in the crossfire, we figured the stakes were high enough to justify the strategic maneuver. After all, a misapprehension of our own cognitive powers—to understand, to predict, to imagine, to know—coupled with technological power, could positively destroy us. We see it every dispatch. We had to get involved.
Like Odysseus, therefore, we began by listening in, by running some thought experiments of our own, on the claims that we were hearing about the nature of intelligence.
When we dropped anchor in the Bay, we also dropped a sounding rod, that marvelous device that helps us plumb the depths of the sea. And down there, we started to see the fault lines of the battle. There were many fronts, reader. More than just two. From one corner, we heard those who said that intelligence was ultimately a matter of making the right binary choices, of following a reasonable path along a tree of Bayesian (Bay-esian?) decisions, so that the best we can do is to be less wrong. From another corner, a set of voices said that intelligence was best understood artificially, that only by looking at the technological ideal of cognitive power can we see what intelligence truly means. From a quieter corner still, there were those who were of the mind that intelligence had to do more with the grand imagination, with our ability to conjure worlds, to generate creative ideas and novelties that no one has thought before. Was one of these true? What were other voices saying? Were there better alternatives?
Atlantis felt overwhelmed at the subject, like we’d been stung by a torpedo fish. We were so overwhelmed, in fact, that the ship almost sank into a reverie of wonder.
But we shook off the fog, and said, “Atlantis, do not lose yourself. You are not without resources! LAND HO!” We sent word back to the Santa Fe Institute and found that we weren’t the only ones who had, on the beaches of that great computational valley, started to investigate the question.
First, we had, in our home port of SFI, the wonderous Davis Professor of Complexity, Melanie Mitchell, who has been asking about how much artificial intelligence relates to human intelligence since before old Atlantis’s winches were wound.
As Mitchell explains in her book, Artificial Intelligence: A Guide for Thinking Humans, we don’t necessarily gain much direct insight into the nature of intelligence from the entities we call AIs, since they don’t actually do anything like the kind of thinking that intelligent humans do. When we look at one of the most ubiquitous AI prototypes, self-driving cars manufactured in that thar Bay of yesterparagraph, we see that they seem woefully lacking in the basic cognitive ability to distinguish a snowman from a human on the street. Now we don’t know about you, but Atlantis isn’t entirely optimistic for the near future prospect of AIs illustrating the nature of intelligence. AIs may be able to beat chess and even Go, but if they can’t recognize a blockhead when they see it, we’re not sure that they can even begin to beat the game of life. No Terminator takeovers on the horizon. Clear waters ahead.
When Atlantis realized that AIs weren’t going to help us, we turned about again. Talk about self-reflection! Where could we spot an example of intelligence? Or where could we even come up with a viable working notion of intelligence at all? Can it be defined? Or do we find ourselves in an “intelligence” bog just as we often do in the “life” lagoon?
Well, reader, the ship began to sail on, on this great Odyssean adventure. We found ourselves Athens bound. On our way, we met another vessel, the Ship for Social Minds, that carried SFI external professor Simon DeDeo. In the gam he convinced us that what is really needed is for all of us to start to ask the question anew.
In a few weeks, DeDeo will be teaching a course on intelligence with the New Center for Research and Practice. As DeDeo remarks, “to study intelligence is to study the power and potential of thought itself.” And, for DeDeo, to see the range in powers of thinking, it’s helpful to look at intelligence in terms of utility, computation, knowledge, and reflection. (Three cheers for reflection! Splice the Mainbrace!)
DeDeo went on. It’s not just a theoretical problem, he said. It’s a political question, and the answers we arrive at, even provisional ones (which may be the only kind) shape deeply how we live.
So, what will life look like for an entity that understands intelligence most fully? Well, we don’t quite know yet. But DeDeo has also given us more food for thought. Recently, he published a little dialogue in Nautilus magazine to present one face of the challenge. And our readers might surmise that Platonic Atlantis loves a dialogue! It’s called “Ian and the Limits of Rationality,” and it delves into how we recognize patterns.
Ian is a student who challenges his teacher to justify testing him on the mathematical pattern that is formed by a partial list of prime numbers. His teacher tells him, Ian, “if you’re asking questions about completing a pattern, you’re looking under the hood at reason itself.” Ian is interested in knowing how we identify patterns when we can’t see them completely. In response, Ian’s teacher introduces him to the world of priors, the rules for the way we judge patterns before complete evidence of a pattern emerges. The two interlocutors, student and teacher, get down to the gritty questions and complex extra-logical beauty that may form the basis of reason.
Does Ian land on firm ground? Well, sail with us reader, for in the next few dispatches we’ll investigate intelligent life in its many forms. We’ll see what lies beyond the farthest shores—and we’ll dive deep, like, say, 20,000 leagues under the sea. What will we find there? Well, we don’t know what you think, but Atlantis has big plans to gam it up with a cephalopod.
Join us next time when Atlantis contemplates abstraction.