14,000,605 AI Futures
"Any sufficiently advanced technology is indistinguishable from magic." - Arthur C. Clarke
In the expansive Marvel Cinematic Universe (MCU), the interplay of superhuman and technological marvels unveils a striking parallel to the evolving landscape of Artificial Intelligence (AI). No, I’m not talking about Jarvis; Dr. Strange's magical foresight, Spiderman's heightened senses, and Hawkeye's exceptional human capabilities resonate deeply with AI's potential trajectory: from the pinnacle of human achievement to the realm of magic.
Consider Dr. Strange in 'Infinity War,' masterfully navigating through 14,000,605 future scenarios to identify a single victorious outcome. This cinematic moment raises a profound question about our understanding of AI and intelligence: Why this specific number? It subtly highlights our tendency to impose finite limits on seemingly infinite possibilities, reflecting our biases in perceiving AI's capabilities.
Similarly, Spiderman's 'spidey senses,' enabling him to sense danger, and Hawkeye's expert precision in targeting symbolize the advanced stages of AI where rapid response and accuracy are paramount. These abilities mirror how humans, like athletes and chess grandmasters, have honed their cognitive skills to anticipate and react quicker than their adversaries.
Do You Want to Play a Game?
Before delving into the neuroscience underpinning intelligence, let's revisit the world of games, a historical testing ground for strategic prediction. Games, ranging from the intellectually demanding Chess and Go to the Viking's Hnefatafl, have always pushed the boundaries of our cognitive capabilities. My AI journey started with the early 80’s movie ‘WarGames.’ In a critical scene, we see a vivid example of how we think about games at the edge of cognitive ability, where a teenage Matthew Broderick engages in a game of Tic Tac Toe with an AI named Joshua. Initially designed for war simulations, the AI learns a crucial lesson in the futility of certain strategies: “… the only winning move is not to play.” Children learn this same lesson as they outgrow Tic-Tac-Toe. Chess and Go, on the other hand, play at our intellectual boundaries.
Chess grandmasters, often perceived as having near-omniscient foresight, are a testament to the human capacity for complex strategic thinking. Contrary to popular belief, their skill is not rooted in the ability to foresee countless move combinations but in their adeptness at pattern recognition and 'chunking' information. Studies by DeGroot and expanded upon by Herbert Simon demonstrate that grandmasters excel not due to superior memory but through rapid encoding of macro features of the chessboard.
“De Groot conducted his chess experiments on players from many different backgrounds, all the way from rank beginners to strong grandmasters…
…[he] exposed subjects very briefly, for 3-4 seconds, to positions taken from a game. He found that grandmasters and masters were able to recall the location of 93% of the pieces, while the experts remembered 72% and the class players merely 51%…
…In later (1973) studies conducted by Herbert A. Simon and W.G. Chase the experiments were conducted with real game positions and compared with random positions…
…in the random position players of all levels did approximately the same. Simon and Chase came to the conclusion that higher-ranked players use a form of chunking, or pattern-matching, that allows them to rapidly encode macro features of the positions.”
This cognitive technique of 'chunking' is prevalent in various skills. For instance, reading evolves from recognizing individual letters and phonemes to comprehending whole sentences and paragraphs. Like chess, it's about organizing information into manageable units, enabling higher-level processing and what seems like superhuman capabilities.
Why Evolution Builds Brains
Exploring the origins of intelligence, Max Bennett's "A Brief History of Intelligence" delves into the emergence of neurons in animals, a crucial evolutionary step. Bennett explains that fast and specific reflexes, vital for survival, necessitated the development of neurons and, subsequently, brains in animals. Of the three branches at the start of life, animals, as opposed to fungi and plants, developed neurons and, eventually, brains. This evolutionary arms race paved the way for the advanced predictive abilities we see in humans and, I believe, will pave the way for augmented intelligence.
"...it was the first time multicellular life ... sensed and moved with speed and specificity. The movement of plants and fungi takes hours to days; the movement of coral takes seconds.* ... the grasping of prey, opening of the mouth, pulling into the stomach, and closing of the mouth all require a well-timed and accurate orchestration of relaxing some muscles while contracting others. And this is why fungi don’t have neurons and animals do. Although both are large multicellular organisms that feed on other life, only the animal-survival strategy of killing level-two multicellular life requires fast and specific reflexes.”
Bennett, Max. A Brief History of Intelligence (pp. 30-31). HarperCollins. Kindle Edition.
Understanding Human Intelligence
We have come a long way from the intelligence of corals. Jeff Hawkins, a pioneer in computing and neuroscience, builds upon this concept of intelligence in his work. He posits that the neocortex, the seat of our intelligence, operates based on a predictive algorithm. The model illustrates that we are in a constant state of prediction. Upon hearing the initial notes of a song, our brains not only anticipate the subsequent notes but also categorize the song by its title. This predictive mechanism extends beyond auditory cues; our visual system continuously forecasts what we see. For example, the blind spots in our vision, caused by the optic nerve at the back of our retinas, go unnoticed as our brains seamlessly 'fill in' or predict the missing visual information.
To elucidate this concept, consider these thought experiments:
While playing hide and seek with your children, you might spot a pair of small sneakers peeking out from under the living room curtain. Despite not seeing the entire figure, your brain reconstructs the image of your child hiding there.
Visualize approaching your home's front door to grasp the door handle. Should someone have playfully shifted the handle slightly to the left, your immediate recognition of this alteration again underscores our continual anticipation of what we expect to see, feel, hear, and more.
"...You can think about the world, move around in the world, and make predictions of the future because your cortex has built a model of the world. One of the most important concepts in this book is that the cortex’s hierarchical structure stores a model of the hierarchical structure of the real world. The real world’s nested structure is mirrored by the nested structure of your cortex.”
Hawkins, Jeff; Blakeslee, Sandra. On Intelligence (p. 125). Henry Holt and Co.. Kindle Edition.
Hawkins' theory aligns with the cognitive strategies of chess grandmasters, who use memory and pattern recognition to narrow down potential outcomes rather than calculating every possible move.
A Brief History of AI
AI's journey in chess mirrors this evolution of cognitive strategies. From IBM's Deep Blue, which employed brute-force calculation, to DeepMind's AlphaZero, which learned chess through self-play and developed novel strategies, AI has demonstrated an increasing alignment with human-like learning and thinking patterns, but in some respects brings an alien-like intelligence to chess:
"What’s also remarkable, though, Hassabis explained, is that it sometimes makes seemingly crazy sacrifices, like offering up a bishop and queen to exploit a positional advantage that led to victory. Such sacrifices of high-value pieces are normally rare. In another case the program moved its queen to the corner of the board, a very bizarre trick with a surprising positional value. “It’s like chess from another dimension,” Hassabis said.
Hassabis speculates that because Alpha Zero teaches itself, it benefits from not following the usual approach of assigning value to pieces and trying to minimize losses. “Maybe our conception of chess has been too limited,” he said. “It could be an important moment for chess. We can graft it into our own play.”
The Queen’s Gambit
Understanding intelligence as a sophisticated prediction system has profound implications for AI's role in augmenting human cognition. 'Queen's Gambit' offers a metaphorical glimpse into this future, where the protagonist visualizes chess moves, much like how AI might enhance our strategic thinking.
As we contemplate AI's future, from autopilot chess computers to collaborative formats like Centaur Chess, the question remains: How do we leverage AI to augment our cognitive capabilities? The answer lies in exploring multidimensional strategies, a topic I will delve into in next week's exploration of ten-dimensional chess and the evolving visions of Centaurian Chess.
The Blurred Line Between Technology and Magic
In conclusion, the journey from MCU's superheroes to chess grandmasters and AI innovators doesn't just entertain; it illuminates the profound connections between human creativity, technological advancements, and the future of AI. As we continue to explore these AI futures, reducing the search space from 14,000,605 possibilities, we edge closer to a world where the line between technology and magic becomes increasingly blurred.
Tech Tonic: The AI Thought Show with Kes & Scott
Continue exploring these captivating AI themes and join the conversation at our upcoming Tech Tonic event! For an immersive experience discussing AI's future, join us here: Tech Tonic: The AI Thought Show with Kes Sampanthar. Don't miss out on this engaging and thought-provoking session!
