(CHACARITA, COSTA RICA) This strange conference is also part of an ongoing psychiatric experiment of sorts. The official meeting hall at the Hilton DoubleTree resort is not empty, but the few people here either sit silently or meander in irregular spirals. The real action is in the Hotel bar where attendees of the multiple gathered “societies” mingle, drink, make jokes, philosophize, pitch stories and crack open the bodies of Brazilian Wandering Spiders and sucking down the creatures’ ferocious venom, quite raw, before drinking very alcoholic antivenom chasers. All the ceremonial spider venom, decorative bananas, and banana crates makes for an unlikely mess. “It means you never can have enough napkins; it doesn’t matter: you always think you need more napkins to keep mopping up,” says one attendee as he wipes the proliferating stains from the dismembered venom glands, the drool and the blood.
The conference’s gathered attendees are part of an experiment to stimulate the growth of brain tumors, which will, it is hoped, improve how well each can imagine and model the future. The attendees who arrived in last week’s group say they can already feel the tumors growing in their skulls, or so they say. Meanwhile, a small cadre of Canadian science-fiction writers is clustered near the bar’s windows and the panoramic view of the coastline. Holding court among them is Peter Watts, whose novels Blindsight and Echopraxia partially inspired this unusual staged experiment. He is unimpressed. “This may be worth doing, for its own sake, but I doubt the outcomes will be as planned” he deadpans. “I’m not drinking that stuff.”
If things go as planned, a network of very small tumors will grow in the tissues that connect fore and anterior regions of the brain, resulting in either increased or decreased integration between them. Depending on your perspective either outcome will make for better or worse future-oriented cognition in those brains who host the tumors and the conference attendees whose bodies host those brains. Divided into informal groups of roughly two dozen members each, the hypotheses vary. Some here are betting that the particular venom/antevenom mixture and the resultant tumors will provide deeper integration of the prefrontal cortex, which specializes in anticipatory abstraction, and the anterior brain, where memories are stored, and so supposedly make for a deeper connection between past memory and future planning. Others, however, bet that the tumors will all but sever the connection, allowing the prefrontal cortex to induce futures in such a way that is “no longer hindered by the human individual autobiographical frame of reference” as put off by one particular excited attendee wearing a Hawaiian shirt, aviator sunglasses and sporting a mother-of-pearl cigarette holder in his teeth.
What is at stake, say the conference organizers, is nothing less than our ability to think about the future —individually and collectively— and in principle to make decisions based on this. During a particularly animated exchange on the conferences’s first day, the essential problem with our current ability to think the future was described by Armen Avanessian, a Viennese philosopher and curator, is tied to how our species’ sense of time evolved to suit the personal escapades of a single finite organism. “That’s the key problem and the key indication that the logic of the contemporary with its fixation on the present — you called it the human fixation on experience — that this presentism has difficulties or even completely fails in dealing with the logic of being constituted by the future."
Because the prefrontal cortex is the locus within the mammalian brain for the modeling of immediate actions in relation to past experiences and possible future consequences, it is the focus of debate and intrigue. Damage to to the cortex, such as in the famous case of railroadman, Phineas Gage, can result in profound changes in personality and an inability to act upon consequence goal-oriented models. In 1848, Gage’s prefrontal cortex was damaged by an iron pipe, but the rest of his brain was largely OK. He could understand and verbalize the longer-term consequences of his actions, but when faced with the performance of actual tasks, he could not help himself but to follow his most instantly-gratifying impulses.
For some attendees, freeing the prefrontal cortex once and for all will make for a human cognition that can imagine the future beyond the concerns of any one person, even in terms of his or her own death. Even the most diehard “disconnectors” (as they are called, somewhat derisively by the “connectors”) don’t think everyone should undergo this alteration. Such persons, it is thought, will have a crucial role to play in an increasingly granular global division of cognitive labor. That is, they foresee a division of cognitive labor at the social scale that corresponds to the division of specialization within the brain itself, with some brains having their prefrontal cortex “overclocked” and others having their anterior brain areas amplified. As each brain/person is able to induce and deduce experiences, patterns and models independently of others, the accumulated collective intelligence that emerges would be more complex and adaptive than when each organism/brain is responsible for performing all tasks. As one attendee with blood slowly dripping out of his ear put it to me with stern clarity, “specialization within a niche is basic function of a more complex ecology and now we are just using biological technologies, “the tumors,” (he emphasizes with air quotes) to catch up with the ecology we already have. We don’t see this as anticipatory but as playing catch-up.” After a weirdly long pause he concludes that “it’s a basic shift from mechanical to organic solitary, but now it’s about the core technology, “the brain,” (again the air quotes) and not just tools and equipment, such as fishing and making shoes.”
Brain-Hacking aside, the 20th century did see a shift would turn thinking about the future from a general self-preservation tactic or prophetic impulse to a specialized expert function. The growth of think-tanks, scenario planning, official governmental and corporate “100 year plans”marked a professionalization of “Futures” composition. As R. John Williams notes in his history of modern futurology, it also marks a shift from marking the future and what will happen (the world will end, the judgment will come, the economy will fall, etc) to a multiplicity of futures that could happen, dependent on big and little contingencies and their manifold interactions. Between 1870 and 1970, Prophecy (singular, immanent, foretold) would give way to Scenario (plural, emergent, contingent).
Other attendees at the conference, equally enthusiastic about gobbling Banana Spider venom, draw a very different lesson from this shift and and from how we might consider extruding social organization from neuroanatomy. For them, de-linking the prefrontal cortex from the rest of the brain (at individual or collective scale) is opposite of what is to be done. For the “connectors” the plan is to more deeply integrate cognitive specializations and, in time, even to merge and dedifferentiate them.
The “connectors” look at this this way. We understand that futural thought is not really encapsulated in any one part of the brain, but includes a wider network that controls the modulation, integration and even executive control over patterns of neural activity (integrating frontopolar cortex, prefrontal cortex, parietal cortex, premotor and dorsal anterior cingulate). It is the frontal-posterior integration across this network that allows for physical motor experience and memory to integrate with abstract and generic models of possible futures. It is thought that the Middle-Upper Paleolithic explosion in human intelligence may have had less to do with the growth of any one neuroanatomical component than with the evolution better integration between parts of the brain, such that one can place things in proper context and to switch between perspectives on the fly. In this way, intelligence, general or not, is this mental flexibility that allows for adaptiveness.
One prominent connector, wearing a well-appointed tool belt across his waist, sneers at a troop of disconnectors as they quietly debate: “that’s the Phinenes Gage Society over there.” On the other side of the bar, other attendees of undetermined allegiance are slumped over the table and being attended to by exasperated paramedics. Apparently this conference may run through six months of Costa Rica’s antivenin reserves in a couple weeks. Extra doses and special cocktails were brought in for the occasion but once he or she gets the taste for it, some find it hard to stop.
The connectors are also eager to extrapolate economic form from neuroanatomical form, but argue for more integration not less. “You can’t just go all symbolic without actual interaction with context.” Says one speaker who is not partaking in the venom/antivenom tumor experiment, but who is here in the capacity of economist interested in how deeper integration between practical experience and memory and projective abstraction could share “executive functions” of determining between conflicting plans for what to do next. Her paper, “On Economic Homologies Between Bogdanov's Tektonics and Genotypic-Phenotypic Optimization in Sussex School Evolutionary Robotics” drew parallels between the turn of the early 20th century Russian physician/philosopher’s integrative systems theory and recent techniques in how robot swarms model and solve problems. In evolutionary robotics, multiple variations in software code (standing in for genetic code) may be tried out between multiple robots (organisms) which will then draw from one another and optimize the code based on whatever bottom-up solution actually works best (realtime Lamarckian distribution). “It’s like deep learning but at a physical scale. We don’t always know why some solution or combination of code works, but we can build systems such that what does work can propagate up and out throughout the extended genome and phenome.” Her conclusion is that while the formal and anticipatory feats of abstraction by the pre-frontal cortex may still lead by the imagination of circumstances that have not yet been encountered and for which there is no “code” to employ, it cannot do so effectively if it does not benefit from bubbling emergent searching-finding-optimizing-replicating machine of bottom-up formfinding.
As the late afternoon drags on and the banners start to slip from overhead, another group here is interested in the venom/antivenom soup/tumor process and its specific ramifications for artificial intelligence. For them the machine intelligence lessons from evolutionary robotics are not allegorical; they are perhaps the central problematic. By and large, this group is less eager to talk to journalists, but sparks to life on the topic of Google DeepMind’s AlphaGo program that recently beat Lee Sedol, a world-class player from South Korea. They compete to articulate in their own words the significance of an unusual move it made in the second match, placing a black stone some distance from the locus of gameplay and sending the confused human contestant from the board needing to refresh. “He (Sedol) sept saying how beautiful the move was and that no human would have done that” they repeated over again, each in their words. “It can see patterns in things that we can’t.” Their interest is in how artificial intelligences, such as AlphaGo, may or may not conceive of the future. They are themselves divided as to whether its conceptions would more mechanistic and deterministic than normal human abstractions, or if its ideas on the future would be neither mechanistic nor deterministic at all, at least as we would recognize them.
Their arguments are intense, philosophical, and technical in different measures. “Evolutionary robotics is as non-teleological as evolution itself,” “so—what, deep learning can have no concept of the future at all?” “No more than a genome can understand the future.” “But future-oriented abstraction is already the result of evolutionary genetics! It already happened!” “But now we are starting from a very different initial state condition, one that is of our own making, and so the outcome is not just because of the brute force algos.” ”But how long is is initial, really? Maybe it’s all been initial and will continue to be initial for a very long time?” ”What does that have to do with anything? You ate one of those spiders didn’t you! If you throw up on me..!”
Those in the A.I. camp, already divided along the issue of mechanistic/deterministic thought, are divided once more as to how such an A.I might collaborate with any of the envonemomated humans at this gathering. Some suggest than only envenomated humans with disconnected prefrontal cortexes would be able to communicate at all with the A.I., others that only those with hyperintegrated frontal/anterior brains could (and indeed how/if the venom-tumor therapy works either way is as yet undetermined). “That's the real integration of a division of labor; not just between different parts of the human society but between those and different parts of non-human intelligence as well.” Others are certain that no communication is possible between the human and non-human forms other than through simple pre-determined rules-based frames (such as Go) but that the deliberate spawning of various mutant neuro-algorithmic cognitive forms is important in its own right. “In the face of a Sixth Great Extinction, post-carbon biodiversity will have to fill in for vertebrate biodiversity and probably more. It’s a terrible solution, but in a way, it has to be a holding pattern at least.”
Fire alarms send everyone out on to the beach, conference attendees of many allegiances and states of sobriety and toxicity stumble along in waves. The sun now set, the hotel’s floodlights are shone into the waves, casting long sharp shadows of the teetering and huddling futurologists. It’s not impossible that some genuinely fundamental breakthrough was approximated in the bar up there above us. It’s possible that one of these speculators chasing each other around the beach, yelling and laughing, is now incubating a brain tumor that will either integrate frontal and anterior brain functions or delink them decisively, which may or may not be a model for a social division of cognitive labor, which would or would not either enable or prevent them to from communicating with game playing A.I.’s. It’s possible that the tumors don’t really grow at all, or that they kill them in a boring normal way, or that they cause some other, perhaps more interesting, development to appear on the scene. In the meantime, the hotel staff beckons us back up hill through megaphones. Dozens of them silhouetted in the glare of the floodlights are all yelling at once, waving their arms and telling us to hurry! Something is happening.