AI Is Evolving in a Way That Could Be Hard to Stop
AI is capable of self-reproduction—should humans be worried?
Eric Weinstein is an American mathematician and economist. He earned his Ph.D in mathematical physics from Harvard University in 1992, is a research fellow at the Mathematical Institute of Oxford University, and is a managing director of Thiel Capital in San Francisco. He has published works and is an expert speaker on a range of topics including economics, immigration, elite labor, mitigating financial risk and incentivizing of creative risks in the hard sciences.
Eric Weinstein: There are a bunch of questions next to or adjacent to general artificial intelligence that have not gotten enough alarm because, in fact, there’s a crowding out of mindshare. I think that we don’t really appreciate how rare the concept of selection is in the machines and creations that we make. So in general, if I have two cars in the driveway I don’t worry that if the moon is in the right place in the sky and the mood is just right that there’ll be a third car at a later point, because in general I have to go to a factory to get a new car. I don’t have a reproductive system built into my sedan. Now almost all of the other physiological systems—what are there, perhaps 11?—have a mirror.
So my car has a brain, so it’s got a neurological system. It’s got a skeletal system in its steel, but it lacks a reproductive system.So you could ask the question: are humans capable of making any machines that are really self-replicative? And the fact of the matter is that it’s very tough to do at the atomic layer but there is a command in many computer languages called Spawn. And Spawn can effectively create daughter programs from a running program.
Now as soon as you have the ability to reproduce you have the possibility that systems of selective pressures can act because the abstraction of life will be just as easily handled whether it’s based in our nucleotides, in our A, C, Ts and Gs, or whether it’s based in our bits and our computer programs. So one of the great dangers is that what we will end up doing is creating artificial life, allowing systems of selective pressures to act on it and finding that we have been evolving computer programs that we may have no easy ability to terminate, even if they’re not fully intelligent.
Further if we look to natural selection and sexual selection in the biological world we find some very strange systems, plants or animals with no mature brain to speak of effectively outsmart species which do have a brain by hijacking the victim species’ brain to serve the non-thinking species.So, for example, I’m very partial to the mirror orchid which is an orchid whose bottom petal typically resembles the female of a pollinator species. And because the male in that pollinator species detects a sexual possibility the flower does not need to give up costly and energetic nectar in order to attract the pollinator. And so if the plant can fool the pollinator to attempt to mate with this pseudo-female in the form of its bottom petal, it can effectively reproduce without having to offer a treat or a gift to the pollinator but, in fact, parasitizes its energy. Now how is it able to do this? Because if a pollinator is fooled then that plant is rewarded. So the plant is actually using the brain of the pollinator species, let’s say a wasp or a bee, to improve the wax replica, if you will, which it uses to seduce the males.
That which is being fooled is the more neurologically advanced of the two species. And so what I’ve talked about, somewhat controversially, is what I call artificial out-telligence. Where instead of actually having an artificially intelligent species you can imagine a dumb computer program that uses the reward, through let’s say genetic algorithms and selection within a computer framework, to increasingly parasitize, using better and better lures, fully intelligent humans. And in the case of artificial intelligence I don’t think we’re there yet. But in the case of artificial out-telligence I can’t find anything that’s missing from the equation. So we have self-modifying code. You have Bitcoin so you could have a reward structure and blockchains. And there’s nothing that I see that keeps us from creating. Now that’s such a strange and quixotic possibility.
In this framework I don’t see an existential risk so that my friends who worry about machine intelligence being a terminal invention for the human species probably don’t need to be worried. But I think there’s a lot of exotica around artificial intelligence which hasn’t been explored and I think which is much closer to fruition. Perhaps that’s good. Maybe it’s a warning shot so thatwe’re going to find that just as we woke up to Bitcoin as digital gold we may wake up to a precursor to artificial general intelligence which alerts us to the fact that we should probably be devoting more energy into this absolutely crazy-sounding future problem which no humans have ever encountered.
Those among us who fear world domination at the metallic hands of super-intelligent AI have gotten a few steps ahead of themselves. We might actually be outsmarted first by fairly dumb AI, says Eric Weinstein. Humans rarely create products with a reproductive system—you never have to worry about waking up one morning to see that your car has spawned a new car on the driveway (and if it did: cha-ching!), but artificial intelligence has the capability to respond to selective pressures, to self-replicate and spawn daughter programs that we may not easily be able to terminate. Furthermore, there are examples in nature of organisms without brains parasitizing more complex and intelligent organisms, like the mirror orchid. Rather than spend its energy producing costly nectar as a lure, it merely fools the bee into mating with its lower petal through pattern imitation: this orchid hijacks the bee's brain to meet its own agenda. Weinstein believes all the elements necessary for AI programs to parasitize humans and have us serve its needs already exists, and although it may be a "crazy-sounding future problem which no humans have ever encountered," Weinstein thinks it would be wise to devote energy to these possibilities that are not as often in the limelight.
Once a week.
Subscribe to our weekly newsletter.
What makes some people more likely to shiver than others?
Some people just aren't bothered by the cold, no matter how low the temperature dips. And the reason for this may be in a person's genes.
Eating veggies is good for you. Now we can stop debating how much we should eat.
- A massive new study confirms that five servings of fruit and veggies a day can lower the risk of death.
- The maximum benefit is found at two servings of fruit and three of veggies—anything more offers no extra benefit according to the researchers.
- Not all fruits and veggies are equal. Leafy greens are better for you than starchy corn and potatoes.
An open letter predicts that a massive wall of rock is about to plunge into Barry Arm Fjord in Alaska.
- A remote area visited by tourists and cruises, and home to fishing villages, is about to be visited by a devastating tsunami.
- A wall of rock exposed by a receding glacier is about crash into the waters below.
- Glaciers hold such areas together — and when they're gone, bad stuff can be left behind.
The Barry Glacier gives its name to Alaska's Barry Arm Fjord, and a new open letter forecasts trouble ahead.
Thanks to global warming, the glacier has been retreating, so far removing two-thirds of its support for a steep mile-long slope, or scarp, containing perhaps 500 million cubic meters of material. (Think the Hoover Dam times several hundred.) The slope has been moving slowly since 1957, but scientists say it's become an avalanche waiting to happen, maybe within the next year, and likely within 20. When it does come crashing down into the fjord, it could set in motion a frightening tsunami overwhelming the fjord's normally peaceful waters .
The Barry Arm Fjord
Camping on the fjord's Black Sand Beach
Image source: Matt Zimmerman
The Barry Arm Fjord is a stretch of water between the Harriman Fjord and the Port Wills Fjord, located at the northwest corner of the well-known Prince William Sound. It's a beautiful area, home to a few hundred people supporting the local fishing industry, and it's also a popular destination for tourists — its Black Sand Beach is one of Alaska's most scenic — and cruise ships.
Not Alaska’s first watery rodeo, but likely the biggest
Image source: whrc.org
There have been at least two similar events in the state's recent history, though not on such a massive scale. On July 9, 1958, an earthquake nearby caused 40 million cubic yards of rock to suddenly slide 2,000 feet down into Lituya Bay, producing a tsunami whose peak waves reportedly reached 1,720 feet in height. By the time the wall of water reached the mouth of the bay, it was still 75 feet high. At Taan Fjord in 2015, a landslide caused a tsunami that crested at 600 feet. Both of these events thankfully occurred in sparsely populated areas, so few fatalities occurred.
The Barry Arm event will be larger than either of these by far.
"This is an enormous slope — the mass that could fail weighs over a billion tonnes," said geologist Dave Petley, speaking to Earther. "The internal structure of that rock mass, which will determine whether it collapses, is very complex. At the moment we don't know enough about it to be able to forecast its future behavior."
Outside of Alaska, on the west coast of Greenland, a landslide-produced tsunami towered 300 feet high, obliterating a fishing village in its path.
What the letter predicts for Barry Arm Fjord
Moving slowly at first...
Image source: whrc.org
"The effects would be especially severe near where the landslide enters the water at the head of Barry Arm. Additionally, areas of shallow water, or low-lying land near the shore, would be in danger even further from the source. A minor failure may not produce significant impacts beyond the inner parts of the fiord, while a complete failure could be destructive throughout Barry Arm, Harriman Fiord, and parts of Port Wells. Our initial results show complex impacts further from the landslide than Barry Arm, with over 30 foot waves in some distant bays, including Whittier."
The discovery of the impeding landslide began with an observation by the sister of geologist Hig Higman of Ground Truth, an organization in Seldovia, Alaska. Artist Valisa Higman was vacationing in the area and sent her brother some photos of worrying fractures she noticed in the slope, taken while she was on a boat cruising the fjord.
Higman confirmed his sister's hunch via available satellite imagery and, digging deeper, found that between 2009 and 2015 the slope had moved 600 feet downhill, leaving a prominent scar.
Ohio State's Chunli Dai unearthed a connection between the movement and the receding of the Barry Glacier. Comparison of the Barry Arm slope with other similar areas, combined with computer modeling of the possible resulting tsunamis, led to the publication of the group's letter.
While the full group of signatories from 14 organizations and institutions has only been working on the situation for a month, the implications were immediately clear. The signers include experts from Ohio State University, the University of Southern California, and the Anchorage and Fairbanks campuses of the University of Alaska.
Once informed of the open letter's contents, the Alaska's Department of Natural Resources immediately released a warning that "an increasingly likely landslide could generate a wave with devastating effects on fishermen and recreationalists."
How do you prepare for something like this?
Image source: whrc.org
The obvious question is what can be done to prepare for the landslide and tsunami? For one thing, there's more to understand about the upcoming event, and the researchers lay out their plan in the letter:
"To inform and refine hazard mitigation efforts, we would like to pursue several lines of investigation: Detect changes in the slope that might forewarn of a landslide, better understand what could trigger a landslide, and refine tsunami model projections. By mapping the landslide and nearby terrain, both above and below sea level, we can more accurately determine the basic physical dimensions of the landslide. This can be paired with GPS and seismic measurements made over time to see how the slope responds to changes in the glacier and to events like rainstorms and earthquakes. Field and satellite data can support near-real time hazard monitoring, while computer models of landslide and tsunami scenarios can help identify specific places that are most at risk."
In the letter, the authors reached out to those living in and visiting the area, asking, "What specific questions are most important to you?" and "What could be done to reduce the danger to people who want to visit or work in Barry Arm?" They also invited locals to let them know about any changes, including even small rock-falls and landslides.
The famous cognition test was reworked for cuttlefish. They did better than expected.
- Scientists recently ran the Stanford marshmallow experiment on cuttlefish and found they were pretty good at it.
- The test subjects could wait up to two minutes for a better tasting treat.
- The study suggests cuttlefish are smarter than you think but isn't the final word on how bright they are.
Proof that some people are less patient than invertebrates<iframe width="730" height="430" src="https://www.youtube.com/embed/H1yhGClUJ0U" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe><p> The common cuttlefish is a small cephalopod notable for producing sepia ink and relative intelligence for an invertebrate. Studies have shown them to be capable of remembering important details from previous foraging experiences, and to adjust their foraging strategies in response to changing circumstances. </p><p>In a new study, published in <a href="https://royalsocietypublishing.org/doi/10.1098/rspb.2020.3161" target="_blank" rel="noopener noreferrer">The Proceedings of the Royal Society B</a>, researchers demonstrated that the critters have mental capacities previously thought limited to vertebrates.</p><p>After determining that cuttlefish are willing to eat raw king prawns but prefer a live grass shrimp, the researchers trained them to associate certain symbols on see-through containers with a different level of accessibility. One symbol meant the cuttlefish could get into the box and eat the food inside right away, another meant there would be a delay before it opened, and the last indicated the container could not be opened.</p><p>The cephalopods were then trained to understand that upon entering one container, the food in the other would be removed. This training also introduced them to the idea of varying delay times for the boxes with the second <a href="https://www.sciencealert.com/cuttlefish-can-pass-a-cognitive-test-designed-for-children" target="_blank" rel="noopener noreferrer">symbol</a>. </p><p>Two of the cuttlefish recruited for the study "dropped out," at this point, but the remaining six—named Mica, Pinto, Demi, Franklin, Jebidiah, and Rogelio—all caught on to how things worked pretty quickly.</p><p>It was then that the actual experiment could begin. The cuttlefish were presented with two containers: one that could be opened immediately with a raw king prawn, and one that held a live grass shrimp that would only open after a delay. The subjects could always see both containers and had the ability to go to the immediate access option if they grew tired of waiting for the other. The poor control group was faced with a box that never opened and one they could get into right away.</p><p>In the end, the cuttlefish demonstrated that they would wait anywhere between 50 and 130 seconds for the better treat. This is the same length of time that some primates and birds have shown themselves to be able to wait for.</p><p>Further tests of the subject's cognitive abilities—they were tested to see how long it took them to associate a symbol with a prize and then on how long it took them to catch on when the symbols were switched—showed a relationship between how long a cuttlefish was willing to wait and how quickly it learned the associations. </p>
All of this is interesting, but what use could it possibly have?<img type="lazy-image" data-runner-src="https://assets.rebelmouse.io/eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpbWFnZSI6Imh0dHBzOi8vYXNzZXRzLnJibC5tcy8yNTcxNzY2MS9vcmlnaW4uanBnIiwiZXhwaXJlc19hdCI6MTY2MTM0MzYyMH0.lKFLPfutlflkzr_NM6WmnosKM1rU6UEIHWlyzWhYQNM/img.jpg?width=1245&coordinates=0%2C10%2C0%2C88&height=700" id="77c04" class="rm-shortcode" data-rm-shortcode-id="7eb9d5b2d890496756a69fb45ceac87c" data-rm-shortcode-name="rebelmouse-image" data-width="1245" data-height="700" />
A diagram showing the experimental set up. On the left is the control condition, on the right is the experimental condition.
Credit: Alexandra K. Schnell et al., 2021<p> As you can probably guess, the ability to delay gratification as part of a plan is not the most common thing in the animal kingdom. While humans, apes, some birds, and dogs can do it, less intelligent animals can't. </p><p>While it is reasonably simple to devise a hypothesis for why social humans, tool-making chimps, or hunting birds are able to delay gratification, the cuttlefish is neither social, a toolmaker, or is it hunting anything particularly <a href="https://gizmodo.com/cuttlefish-are-able-to-wait-for-a-reward-1846392756" target="_blank" rel="noopener noreferrer">intelligent</a>. Why they evolved this capacity is up for debate. </p><p>Lead author Alexandra Schnell of the University of Cambridge discussed their speculations on the evolutionary advantage cuttlefish might get out of this skill with <a href="https://www.eurekalert.org/pub_releases/2021-03/mbl-qc022621.php" target="_blank" rel="noopener noreferrer">Eurekalert:</a> </p><p style="margin-left: 20px;"> "Cuttlefish spend most of their time camouflaging, sitting and waiting, punctuated by brief periods of foraging. They break camouflage when they forage, so they are exposed to every predator in the ocean that wants to eat them. We speculate that delayed gratification may have evolved as a byproduct of this, so the cuttlefish can optimize foraging by waiting to choose better quality food."</p><p>Given the unique evolutionary tree of the cuttlefish, its cognitive abilities are an example of convergent evolution, in which two unrelated animals, in this case primates and cuttlefish, evolve the same trait to solve similar problems. These findings could help shed light on the evolution of the cuttlefish and its relatives. </p><p> It should be noted that this study isn't definitive; at the moment, we can't make a useful comparison between the overall intelligence of the cuttlefish and the other animals that can or cannot pass some variation of the marshmallow test.</p><p>Despite this, the results are quite exciting and will likely influence future research into animal intelligence. If the common cuttlefish can pass the marshmallow test, what else can?</p>