English
Based on โThe rise and fall of civilizations | Eric Cline: Full Interviewโ from Big Think Watch the original video
The Day the World Ended (and Began Again): Lessons from the Bronze Age Collapse
More than three millennia ago, a sophisticated globalized world flourished around the Mediterranean. Powerful empires traded, negotiated, and intermarried, forming a complex web of interdependence. Then, in a shockingly short span of decades around 1200 BCE, this intricate system came crashing down, ushering in centuries of what historians call a โDark Age.โ
For Eric Cline, an archaeologist and ancient historian at George Washington University and author of โ1177 BC: The Year Civilization Collapsed,โ this period isnโt just a fascinating historical footnote. Itโs a stark mirror reflecting our own hyper-connected modern world, offering critical insights into the fragility of complex systems and the potential for a โperfect stormโ of catastrophes.
โThis was a time where people were basically globalized around the Mediterranean in a way that is not frequently seen,โ Cline explains. โAnd so what happened to them back then may have implications for us today. It turns out to be a much more important period to study than one might expect, even though itโs more than 3,000 years ago.โ
The โAncient G8โ: A World of Interdependence
The Late Bronze Age, spanning roughly 1700 to 1200 BCE, was an era of unprecedented internationalization. Cline likens it to an โancient G8,โ where major powers maintained constant contact, whether direct or indirect. This elite club included:
- Mycenaeans and Minoans in the Aegean (think the world of the Trojan War).
- Hittites in Anatolia (modern Turkey).
- Assyria and Babylonia in Mesopotamia (modern Iraq).
- Egyptians in the Nile Valley.
- Canaanites (in what is now Israel, Palestine, Lebanon, Syria).
- Cypriots on the island of Cyprus.
This wasnโt just casual acquaintance; it was a โsmall world networkโ where any two points were often just โtwo hopsโ away. Commercial, diplomatic, and even marital ties bound these civilizations together. Raw materials were the lifeblood of this network. Egypt supplied gold, Greece provided silver, and Cyprus was the primary source of copper. The most crucial, and problematic, commodity was tin, essential for making bronze (90% copper, 10% tin). Without tin, you couldnโt make tools, weapons, or art. Much of this tin traveled hundreds of miles, often from as far as Afghanistan.
Beyond raw materials, there was a vibrant trade in olive oil, wine, grain, and luxury goods. Cline recounts an anecdote of leather shoes sent from Crete to King Hammurabi in Babylon, only to be returned โ perhaps they were โtoo last millennium,โ he muses. At the highest echelons, kings engaged in elaborate gift-giving, exchanging items like solid gold daggers inlaid with lapis lazuli.
Diplomacy was cemented through dynastic marriages. Egyptian pharaohs like Amenhotep III and Akhenaten had harems filled with princesses from Mitanni and Babylonia, solidifying treaties. The fascinating detail? Egyptians never sent their own princesses abroad; they only received them.
This complex web of interdependence was both their greatest strength and their fatal flaw. It allowed these civilizations to reach unprecedented heights of prosperity and cultural exchange. But, as Cline warns, โThis globalized network is what rose them up to the highest levels, but it is then also what brought them crashing down at the end when all of that was cut.โ
What Does โCollapseโ Even Mean?
The term โcollapseโ itself is a subject of academic debate. Some scholars prefer โtransformation,โ arguing that societies didnโt vanish but merely changed. Cline, however, stands firm: โThe world that they had had in place in the 14th century and the 13th century BC, that world goes away. And it goes away within just a couple of decades in the early 12th century BC.โ
What truly collapsed was the network โ the commercial, diplomatic, and marital connections that linked them all. This wasnโt an overnight event for everyone, nor did every society disappear entirely. But for many, life as they knew it fundamentally altered. Cline compares it to the fall of the Roman Empire: โIt was catastrophic for its day.โ
This phenomenon is known as a โsystems collapse,โ a term coined by Colin Renfrew. It signifies the breakdown of a complex system where central economies, elite structures, and centralized governments vanish, often accompanied by massive population decline due to death and migration. The ensuing period is typically a โDark Age,โ marked by a regression to lower levels of socio-political and economic functioning.
The โPerfect Stormโ: Beyond a Single Cause
For decades, scholars sought a single, monocausal explanation for the Bronze Age collapse. The most popular culprit was the โSea Peoples,โ a mysterious confederation of invaders mentioned in Egyptian texts. These groups, with names like the Shardana, Shekelesh, and Peleset, were depicted as sweeping across the Mediterranean, attacking and destabilizing kingdoms. The Peleset, interestingly, are widely identified as the Philistines of biblical fame, who appear to have been Mycenaeans from Greece who fled to the Levant.
However, modern archaeology and a wealth of new data have fundamentally shifted this understanding. Cline and his colleagues now advocate for a โpolycausalโ explanation โ a โperfect stormโ of interconnected disasters. โYou need two of the explanations, three, four, something like that happening either all at once or in rapid succession,โ Cline asserts. โSo you donโt have time to recover from one catastrophe before the next one hits.โ
The Sea Peoples, Cline suggests, are largely โhistoryโs great scapegoats.โ Far from being the sole destroyers, they were more likely symptoms of a wider crisis, victims themselves, migrating in search of a better life. He draws parallels to the American Dust Bowl of the 1930s or modern-day refugees fleeing conflict, like those from the Syrian Civil War.
So, if not just the Sea Peoples, what else? The evidence points to a terrifying confluence of factors:
-
Mega-Drought and Famine: This is no longer a hypothesis but a scientific certainty. Data from stalagmites, dried-up lakes, and pollen analysis reveal a โmega-droughtโ lasting anywhere from 150 to 300 years, beginning around 1250 BCE and affecting a vast area from Italy to Iran. Such a prolonged drought would have devastated agriculture, leading directly to widespread famine. Texts from Ugarit (a vital port city in Syria) and Hittite archives explicitly plead for grain and report severe starvation. This environmental catastrophe likely spurred the migrations of groups like the Sea Peoples.
-
Invasions and Internal Rebellions: While the Sea Peoples werenโt the sole cause, there were invaders. New tablets from Ugarit describe โenemy shipsโ landing and overrunning port cities before advancing on Ugarit itself. The archaeological record confirms a devastating destruction layer at Ugarit, a meter deep, with bodies in the streets and arrowheads in walls. The city was abandoned for centuries. However, not all destruction was external. Some cities, like Hazor in Canaan or Mycenae in Greece, show signs of internal rebellion, where starving populations might have risen against their elites, burning palaces and temples but leaving common houses intact.
-
Earthquake Storms: The Eastern Mediterranean is a seismically active zone. Evidence suggests a series of powerful earthquakes, an โearthquake stormโ (or โearthquake sequenceโ in modern seismology), struck between 1225 and 1175 BCE. Off-kilter walls, slipped keystones, and bodies trapped under collapsed structures (like the teenage girl at Mycenae) point to widespread seismic activity. Even Troy, site of the legendary war, saw its โCity Sixโ destroyed by an earthquake, not human invaders. These natural disasters would have compounded the existing stresses, crippling infrastructure and further disrupting trade.
-
Disease: The โfour horsemen of the apocalypseโ often ride together. While harder to pinpoint archaeologically, historical accounts like Homerโs Iliad and the biblical Exodus mention plagues. More concrete evidence comes from the mummy of Egyptian pharaoh Ramses V (c. 1140 BCE), which shows pustules consistent with smallpox. A papyrus from Turin describes his death from plague and an unprecedented 16-month quarantine of the Valley of the Kings. Disease, especially in a time of drought, famine, and population displacement, would have been a final, devastating blow.
This โpoly crisisโ created a multiplier and domino effect. One catastrophe weakened societies, making them more vulnerable to the next. The disruption of a single vital trade route, like the tin supply, could cripple an entire network.
What Could Have Been Avoided? And What Came Next?
To understand the networkโs vulnerability, Cline collaborated with the US Army Corps of Engineers, running computer simulations of collapse scenarios. They found that the simultaneous fall of the Hittite Empire and the crucial port city of Ugarit would have been enough to shatter the entire globalized system. The Hittites did disappear, and Ugarit was utterly destroyed. Egypt, though severely weakened, โmuddled through,โ surviving the worst of the collapse. This suggests that while resilience is possible, the ability to adapt and react to unfolding crises is paramount.
The immediate aftermath was indeed a โDark Ageโ (c. 1200-800 BCE). Monumental building ceased, and writing systems like Linear B in Mycenaean Greece disappeared, leading to a loss of literacy. Historians often characterize this period as a regression, where people looked back at the Bronze Age as a โgolden age,โ reflected in epics like Homerโs Iliad and Odyssey.
Yet, necessity proved to be the mother of invention. The scarcity of tin forced a shift to a new metal: iron. Though known before, iron became widely adopted during and after the collapse, ushering in the Iron Age. The Phoenicians, survivors from central Canaan (modern Lebanon), not only thrived but also standardized the alphabet, which they then spread across the Mediterranean, laying the foundation for the Greek and Latin alphabets we use today. The Cypriots, too, demonstrated remarkable resilience, becoming pioneers in iron metallurgy.
History Rhymes
The story of the Late Bronze Age collapse is a powerful cautionary tale. It reveals how tightly woven global systems, while fostering prosperity, also create vulnerabilities. A cascade of seemingly disparate events โ climate change, resource scarcity, migration, conflict, and disease โ can overwhelm even the most powerful civilizations.
As Eric Cline concludes, โIโm not saying that weโre necessarily due for another collapse, but history does rhyme even if it doesnโt repeat.โ In an era of increasing global interdependence, climate change, and geopolitical instability, the echoes of 1177 BCE are perhaps louder than ever. Understanding how ancient civilizations navigated their โperfect stormโ might just be one of the most important lessons for our own survival.
Based on โAre Human Drivers Finally Obsolete? | Freakonomics Radioโ from Freakonomics Radio Network Watch the original video
The End of the Road for Human Drivers? Inside the Quest to Automate the Wheel
For centuries, the act of driving has been intrinsically human โ a task of skill, instinct, and sometimes, unfortunate error. From the horse-drawn carriage to the modern automobile, a person has always been at the helm. But what if that era is rapidly drawing to a close? What if the word โdriverโ soon conjures images not of a person, but of a machine, much like โdishwasherโ or โcomputerโ do today?
This profound shift is the subject of a deep dive by PJ Vote, host of the Search Engine podcast, who recently explored the two-part series on driverless cars for Freakonomics Radio. His journey into this topic began, surprisingly, not with technology, but with a bench-pressing injury.
โIโd gotten too into bench pressing,โ Vote recounted, โand I injured myself. I had a hernia and then I had to have a hernia repair.โ With limited mobility and in pain, a visit to a friend in San Francisco led him to take a Waymo robo-taxi. โIt was such an experience of the future that immediately becomes normal,โ he marvelled. โThe idea that I would press a button on my phone, a car would come out of nowhere driven by nobody. I would get in, watch the steering wheel turn itself.โ He likened the initial awe to a first airplane ride, which quickly becomes as mundane as an elevator by the third. This transformative experience sparked his curiosity: why wasnโt everyone talking about this impending revolution?
The Visionaries and the Skeptics
Voteโs investigation uncovers a cast of compelling characters driving this change. In the first part of his series, focusing on the car itself, he introduces Sebastian Thrun. A German-born roboticist and AI expert, Thrun lost a friend to a car accident as a teenager. His motivation for developing autonomous vehicles transcends mere convenience or profit; he genuinely seeks to reshape the modern world and save lives. His early ideas, once dismissed as crazy, have steadily gained plausibility over two decades.
The second part of the series, shifting focus to the driver, features the strongly opinionated politicians of Boston, grappling with the societal implications of this technology. Steven Dubner, host of Freakonomics Radio, noted a recurring theme throughout Voteโs reporting: โEvery time I spoke to someone, as they were talking, I thought everything they were saying makes senseโฆ And then I would go talk to the next person who saw things completely differently and it would just spin my head the other way.โ This struggle to reconcile competing, yet equally logical, interests lies at the heart of the driverless car debate.
Dubner himself confesses to being โanti-human driver for about 50 years.โ He readily admits his own driving flaws โ a temper, distractions, and a lack of skill. After riding in an autonomous test vehicle at Carnegie Mellon University, he was convinced: โGive me the autonomous vehicles. So plainly better than I am as a driver.โ
A Ghost of Jobs Past: The Knocker-Upper and the Lamp Lighter
To truly grasp the potential impact of driverless cars, Vote begins with a historical analogy. Imagine yourself nearly 200 years ago, waking before dawn to a hard rapping at your window. Thatโs the โKnocker-Upper,โ a job that existed for a century before alarm clocks, where a person walked the neighborhood tapping on windows to rouse workers. Outside, gas street lamps still burned, lit the night before by the โLamp Lighter,โ a cherished neighborhood fixture who made his rounds at dusk with a ladder and a light.
And then thereโs you, a professional driver โ a person on a coach, holding the reins of a horse, taking passengers where they needed to go. Today, the Knocker-Upper is your iPhone alarm, the Lamp Lighter is the electric street light. These jobs have vanished, replaced by technology. The driver, however, has persisted. The question now is whether this third job, and indeed the routine human task of driving for nearly everyone else, is also on the brink of transformation.
The Perils of Human Control
Alex Davies, author of Driven: The Race to Create the Autonomous Car, shares Voteโs fascination with human driving limitations. โI canโt always pay attention to everything,โ Davies admits, โthat I get tired.โ He recently resolved to be calmer on the road, especially with a baby on the way. This resolution underscores a stark reality: for most of us, driving is the riskiest behavior we routinely engage in. Davies himself was in an accident just months after speaking with Vote, totaling his car.
Safety is the core promise of the driverless car. Computers donโt get drunk, tired, or distracted. They never text or succumb to road rage. And these vehicles arenโt a distant dream; theyโre here. Robo-taxis like Waymo already operate in 10 American cities, providing millions of rides. China boasts an even wider rollout. In places like San Francisco or Austin, a driverless car is as common as an Uber, with passengers choosing between a human or a robot driver via their phone. This rapid evolution, happening now, promises to fundamentally reshape our cities and our lives.
Chapter 1: Dreams Without Drivers
The idea of a self-driving car is almost as old as the automobile itself. When humanity transitioned from horse-drawn carriages to mechanical vehicles, something crucial was lost: sentience. A horse wouldnโt simply run off a cliff if you dropped the reins. Early automobiles, powerful and non-sentient, faced passionate resistance. In the 1800s, people feared these โthunderingโ machines, partly due to job displacement. Horse breeders, farriers, feed suppliers, manure haulers, carriage manufacturers, and โTeamstersโ (original drivers of horse teams) all saw their livelihoods threatened.
Beyond jobs, cars were undeniably unsafe. Anti-car activists pushed for regulations like โred flag laws,โ requiring a person to walk ahead of an automobile waving a flag. One Pennsylvania proposal even suggested drivers disassemble their cars and hide the parts behind bushes if livestock were encountered. While extreme, these activists were directionally correct: cars initially wiped out many jobs and were incredibly dangerous. Cities like Detroit, which initially embraced cars without regulation, saw astonishing death rates, particularly among children, in the early 1900s.
It took decades for society to adapt: laws, licenses, driverโs education, road design improvements, highways, seat belts, and airbags all made driving safer. However, the smartphone has reversed some of that progress. Today, car-related deaths in America are as common as those from guns or opioids โ about one in a hundred people will likely know someone who dies in a car accident. This enduring problem fueled the desire to make cars more โsentient,โ like the horses they replaced. Early visions included radio-controlled cars or vehicles guided by magnets embedded in the road. The limiting factor was always the available technology.
Chapter 2: The DARPA Grand Challenges โ Forging a Future
The turn of the millennium brought a turning point. Deep within the Department of Defense, DARPA โ the agency behind GPS, the M16, and the early internet โ set its sights on autonomous military vehicles. In 2002, DARPAโs director, Tony Tether, proposed an unusual approach: a contest. The goal was to inspire tech innovation beyond simply building more websites. The prize for winning the โGrand Challengeโ was $1 million.
The rules were remarkably open: any vehicle, any design, as long as it didnโt attack other competitors (a question actually posed by a team with โBattlebotsโ history). Tetherโs initial idea of racing down the Las Vegas Strip was quickly dismissed as impractical, leading to the chosen venue: the desert outside Las Vegas. The true โdriverโ DARPA sought to replace was the American soldier, envisioning vehicles that could navigate roads potentially laden with explosives.
Two key engineers emerged from this formative period. Chris Urmson, a PhD student at Carnegie Mellon University, joined their โRed Teamโ to build โSandstorm,โ a bright red Humvee bristling with futuristic sensors. He grappled with the fundamental challenge of teaching a computer to control a vehicleโs steering, brakes, and throttle. Representing a different approach was Anthony Lewendowski, a charming, gangly entrepreneur from Berkeley. Lacking the resources of Carnegie Mellon, he opted for a standout design: โGhost Rider,โ the raceโs only self-driving motorcycle.
The 2004 Grand Challenge was, in Voteโs words, โan utter hysterical disaster.โ Ghost Rider toppled immediately because Lewendowski forgot to flip a stabilization switch. Every subsequent vehicle failed miserably. Sandstorm got stuck on a berm, its wheels spinning so hard they melted the tires, emitting plumes of black smoke before being shut down. Chris Urmson compared it to an Olympic marathon where the best runner only completes two miles.
However, the contest had achieved its underlying goal: it had โflushed all these inventors out,โ jumpstarting the scene that would develop autonomous technology. Observing the chaos that day was Sebastian Thrun, the legendary roboticist from Stanford. He saw a fundamental error: โall the teams treated this like a hardware problem.โ Thrun believed the true challenge was software, replacing the human driverโs mind, not just building bigger wheels. He was also driven by a humanitarian vision, imagining the potential to save millions of lives globally if traffic accidents could be eliminated.
Eighteen months later, in October 2005, DARPA doubled the prize to $2 million for the second Grand Challenge. Familiar faces returned: Urmson with two Carnegie Mellon vehicles, Lewendowski again with his still-failing motorcycle. But the new contender was Thrunโs Stanford team, with their comparatively โmeaslyโ blue Volkswagen SUV named โStanley.โ Thrun, a computer scientist, brought a focus on artificial intelligence, teaching Stanley to recognize roads and train itself by recording what its cameras saw. Stanley learned to differentiate between good driving surfaces (like grass) and bad ones (like mud) by measuring slipperiness and bumpiness. It detected patterns and generalized its learning thirty times a second, much like a human.
The second race was a resounding success. Multiple vehicles finished the challenging 132-mile course. The real question was who would do it fastest. Stanley, sandwiched between Carnegie Mellonโs behemoths, emerged victorious. Thrun described the sight of his blue car, a โdust cloudโ turning bluish, crossing the finish line as โunbelievably magical.โ It was a โmade-for-TV Kumbaya moment,โ a testament to a community of innovators, before the cutthroat competition for driverless cars truly began.
Chapter 3: Googleโs Secret Project โ From Desert to Public Roads
The DARPA Grand Challenge had not only showcased roboticists but also attracted an unexpected observer: Google co-founder Larry Page. Disguised in a hat and sunglasses, Page buttonholed Thrun, asking highly specific questions about his LiDAR system. Their connection wasnโt new; Thrun had previously fixed a small robot Page had built for telepresence meetings. Page, who had wanted to do his grad school thesis on autonomous vehicles, now saw tangible proof that the dream could be real.
Initially, Page hired Thrun and Lewendowski to build Google Street View, modifying Stanleyโs roof-mounted camera system. But soon, Page returned to his true dream: a driverless car. In 2009, Page approached Thrun with a mission: โSebastian, I think you should build a self-driving car that can drive anywhere in the world.โ Thrunโs immediate reaction was skepticism. โNo, taking the technology we built for this empty desert and putting in the middle of Market Street in San Francisco is going to kill somebody.โ
Page persisted, asking Thrun to provide a technical reason why it couldnโt be done. โThatโs the moment of incredible pain,โ Thrun recalled, โbecause I go home and I canโt think of a technical reason why not.โ This realization taught him a crucial lesson: โExperts are usually expert of the past, not the future, and if you ask an expert about innovation, something crazy new, theyโre the least likely person to say yes, it can be done.โ
Thus began Googleโs secret self-driving car project, โProject Chauffeur,โ in 2009. Led by Thrun, with Chris Urmson managing day-to-day operations, Anthony Lewendowski on hardware, and Dmitri Dolgov on planning, the small team of 11 engineers reported directly to Larry Page. Their nebulous goal was refined into two challenges: safely log 100,000 miles on public roads and complete the โLarry 1K.โ The Larry 1K involved driving 10 โtrickyโ 100-mile routes across California โ from the Bay Bridge to Lombard Street โ without a single human takeover.
To kickstart the project, the team licensed Stanfordโs DARPA Urban Challenge code. Lewendowski bought eight Toyota Priuses, retrofitting them with radar, cameras, and a spinning 360-degree LiDAR system. These cars, initially given names like โNight Riderโ before being numbered, were transformed into autonomous test vehicles. Don Bernett, a researcher on motion planning, joined the team, tasked with teaching the car nuanced behaviors like โnudgingโ โ the subtle shift a human driver makes to the left when a large truck passes on the right.
Early testing took place in secrecy, in the Shoreline Amphitheater parking lot near Googleโs offices and an empty airplane runway. Spring 2009 marked their first real road driving on the Central Expressway. Immediately, a critical flaw emerged: the car was โswerving wildly,โ like a โdrunken sailor.โ The small oscillations unnoticed on a runway became a significant problem on a public road.
The team adopted rigorous safety protocols. Two-person teams manned each car: a safety driver behind the wheel, ready to take over, and a partner watching a graphical interface, calling out discrepancies between sensor data and reality. This iterative process of logging errors, troubleshooting, and updating code was how the car learned. Bernett recalled how this intensely practical work made him obsess over human driving behavior. Why do humans drive the way they do? He found no easy answers, noting that โmachine learningโ now infers these โdeep truths.โ
He offered a fascinating example: the โlateral accelerationโ (the force that pushes you sideways) tolerated by passengers varies greatly with context. Two meters per second squared feels comfortable on a highway on-ramp, but โincredibly uncomfortableโ (like โMario Kartโ) during a U-turn in a cul-de-sac, where the limit is almost three times less. Human perception, not just physical forces, dictated comfort.
Despite these complexities, the unifying goal of the Larry 1K kept the team focused. By 2010, just a year in, they were on a roll, knocking out routes like Route One to Carmel and the Bay Area bridges. Each failure provided valuable data, leading to code improvements. They completed the Larry 1K in just over a year, nearly twice as fast as expected, celebrating each completed route with a $13.99 bottle of Corbell champagne.
By the fall of 2010, the โmiracleโ was complete. They had built a driverless car, with human supervision and extensive coding, that could safely navigate tricky California roads. They celebrated, throwing each other into Sebastian Thrunโs pool. But then, a new question emerged: โOkay, and now what?โ The path forward began to wobble. Competition would arrive, the team itself would face internal divisions, and some, believing the pace too slow, would take matters into their own hands in extreme ways. The era of driverless cars was just beginning, and the road ahead was anything but straight.
Based on โAndrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AIโ from No Priors: AI, Machine Learning, Tech, & Startups Watch the original video
The Loopy Era of AI: Andrej Karpathy on Living with Code Agents, AutoResearch, and Digital Psychosis
Imagine a world where you no longer write code, but instead โexpress your willโ to an army of AI assistants working tirelessly on your behalf. Where your home runs itself, orchestrated by a digital butler you text via WhatsApp. And where scientific research is conducted autonomously, with AI models recursively improving themselves, leaving human experts to merely contribute ideas to a queue.
This isnโt a distant sci-fi fantasy; itโs the present reality for Andrej Karpathy, a leading voice in AI and former head of AI at Tesla. In a recent conversation on the โNo Priorsโ podcast, Karpathy offers a captivating, at times bewildering, glimpse into what he terms the โAI psychosisโ of this โloopy era.โ His insights reveal a profound shift in how we interact with technology, work, and even conceive of intelligence itself.
The End of Manual Coding: A Personal Revolution
For Karpathy, the transformation has been swift and absolute. โCodeโs not even the right verb anymore,โ he declares, describing his new daily routine: โI have to express my will to my agents for 16 hours a day.โ This isnโt hyperbole. Since December, he estimates he hasnโt typed a single line of code himself. His workflow has flipped from 80% manual coding to a mere 20%, delegating the vast majority of tasks to AI agents.
This radical shift, Karpathy explains, stems from a โhuge unlock in what you can achieve as a person, as an individual.โ No longer bottlenecked by typing speed or manual execution, he now commands a fleet of digital helpers. He recounts his โperpetual state of AI psychosis,โ a mix of exhilaration and anxiety born from the rapid, unexplored capabilities now at his fingertips. The default workflow for software engineers, he argues, has fundamentally changed since late 2023, yet many outside the immediate AI frontier have yet to grasp the magnitude of this revolution.
Beyond Single Sessions: The Rise of โClaw-like Entitiesโ
The initial interaction with AI coding assistants like GitHub Copilot or OpenAIโs Codex often involves a single session, a brief exchange. But Karpathy and others are pushing far beyond this. The next frontier involves not just multiple agents, but persistent, โClaw-like entitiesโ that loop and operate autonomously.
He points to Peter Steinberg as a pioneer in this space, famous for a setup where numerous coding agents (like Codex) tile his monitor, each independently tackling different tasks. โItโs not just like hereโs a line of code, hereโs a new function,โ Karpathy explains. โItโs like hereโs a new functionality and delegate it to agent one. Hereโs a new functionality thatโs not going to interfere with the other one. Give it to agent two.โ This allows for โmuch larger macro actionsโ over a software repository, with agents simultaneously researching, planning, and writing code. The humanโs role morphs into a high-level manager, reviewing work and orchestrating these macro-level operations.
This new mode of interaction is โvery rewardingโ because it works, but itโs also โthe new thing to learn,โ contributing to Karpathyโs โpsychosis.โ The challenge becomes developing a โmuscle memoryโ for this meta-level command.
The Human as the New Bottleneck: A โSkill Issueโ
With AI agents handling the grunt work, the bottleneck has shifted dramatically. Itโs no longer about compute power or access to resources, but about the humanโs ability to effectively direct and leverage these tools. Karpathy vividly describes feeling โnervous when I have subscription left overโ on his AI platforms, implying he hasnโt maximized his โtoken throughputโโthe sheer volume of AI processing heโs commanding. This echoes a past anxiety of PhD students seeing their GPUs idle, but now itโs about tokens, not flops.
The prevailing sentiment when things donโt work is โskill issue.โ Itโs not that the AI canโt do it, but that the human hasnโt given good enough instructions, hasnโt configured the memory tools properly, or hasnโt found the right way to parallelize the agents. This โaddictiveโ nature of self-improvement, where โunlocksโ occur as one gets better at commanding AI, is a core driver of the โpsychosis.โ
Dobby the Elf Claw: Automating Life Itself
Karpathyโs personal life offers a compelling illustration of this agentic future. During a period of โClaw psychosisโ in January, he built โDobby the elf claw,โ an AI agent that manages his entire home.
Dobby began by autonomously scanning Karpathyโs local network, identifying smart home subsystems like Sonos speakers. With a few prompts, the agent reverse-engineered APIs, found endpoints, and soon music was playing in the study. Dobby then took control of lights, HVAC, shades, pool, spa, and even the security system. Now, Karpathy simply texts Dobby via WhatsApp: โDobby, sleepy time,โ and all lights go off. His security camera, paired with an AI model, sends him texts with images when a FedEx truck pulls up.
This personal automation highlights a deeper implication: the potential obsolescence of many bespoke apps. โThese apps that are in the app store for using these smart home devicesโฆ shouldnโt even exist,โ Karpathy posits. Instead, everything should be exposed as APIs, with agents acting as the intelligent glue. โThe customer is not the human anymore. Itโs like agents who are acting on behalf of humans.โ While this currently involves some โvi codingโ for technically proficient users, Karpathy predicts that within โa year or two or three,โ such home automation will be โtrivialโ and โfree,โ easily translating from non-technical human intent.
AutoResearch: AI Improving AI
The concept of removing the human bottleneck extends most dramatically to scientific discovery. Karpathyโs โauto researchโ project exemplifies this. His goal: โI donโt want to be like the researcher in the loopโฆ Iโm holding the system back.โ The objective is to arrange systems so they are โcompletely autonomous,โ allowing agents to run โfor longer periods of time without your involvement.โ
Karpathy, with two decades of experience training large language models (LLMs), put his own expertise to the test. He had painstakingly tuned a GPT-2 model by hand, believing it was โfairly well tuned.โ But after letting auto research run overnight, it returned with โtunings that I didnโt see,โ identifying subtle interactions between hyperparameters like weight decay and Adam betas that he had missed.
This single loop of auto research, Karpathy emphasizes, is just the beginning. Frontier AI labs are already exploring โrecursive self-improvement,โ where LLMs improve other LLMs. He envisions automated scientists generating ideas from papers and GitHub repos, funneling them into a queue for AI workers to test. โRemoving humans from all the processes and automating as much as possibleโ is the future of research, requiring a complete โrethinking of all the abstractions.โ
The Meta-Loop: Optimizing Research Itself
The โloopy eraโ extends even to the organization of research. Karpathy muses about โprogram MDsโโmarkdown files that describe how a research organization should work, detailing roles, processes, and even risk appetite. Just as auto research optimizes model parameters, one can imagine a โmeta-layerโ where AI optimizes these โprogram MDs.โ
โYou can definitely imagine that you have multiple research orgs,โ Karpathy explains, each with its own โcodeโ (its program MD). โAnd once you have code, then you can imagine tuning the code.โ This opens the door to AI designing better, more efficient research processes, leading to a perpetual loop of self-improvement at every level.
The Jagged Edge: AIโs Limitations and the Call for Speciation
Despite the astounding capabilities, Karpathy acknowledges the current limitations of AI. He describes the experience of interacting with these models as simultaneously talking to โan extremely brilliant PhD student whoโs been like a systems programmer for their entire life and a 10-year-old.โ This โjaggedness,โ where models excel in verifiable tasks but struggle with nuance, intent, or creativity, is a persistent frustration.
He illustrates this with the โatom jokeโ: โWhy donโt scientists trust atoms? Because they make up everything.โ This joke, he notes, was common years ago and is still what youโll get from state-of-the-art models today. While agents can โmove mountainsโ for hours on complex tasks, they offer the same โcrappy joke from 5 years agoโ because โitโs outside of the reinforcement learning. Itโs outside of whatโs being improved.โ This suggests a decoupling: being smarter at code generation doesnโt automatically translate to broader intelligence or creativity in all domains.
This โjaggednessโ leads Karpathy to question the current โmonoculture of modelsโ pursued by many labsโsingle, massive LLMs intended to be arbitrarily intelligent across all domains. He argues for โmore speciation in the intelligences,โ akin to the diversity of brains in the animal kingdom. Instead of an โoracle that knows everything,โ we might benefit from smaller, specialized models that retain a โcognitive coreโ but become highly efficient and competent in specific niches, like mathematics or particular programming languages. This โunbundlingโ could be driven by the need for efficiency, especially with compute constraints, though the โscience of manipulating the brainsโ (deep fine-tuning without losing capabilities) is still nascent.
A Future of Infinite Possibility and Perpetual Motion
Andrej Karpathyโs vision paints a picture of an AI-driven future that is both exhilarating and dizzying. We are entering a โloopy eraโ where AI agents are not just tools but collaborators, autonomous entities that will increasingly manage our digital and physical worlds. The human role is shifting from direct execution to high-level orchestration, leveraging AI to achieve unprecedented scales of productivity and discovery.
Yet, this transformation comes with its own anxieties: the โpsychosisโ of infinite possibilities, the constant feeling of a โskill issue,โ and the challenge of navigating AIโs brilliant but sometimes nonsensical jaggedness. As Karpathy concludes, the progression is obvious, but the path is still rough. The journey into the loopy era of AI promises to be one of perpetual motion, constant learning, and profound redefinition of what it means to be human in a world increasingly shaped by machine intelligence.
Based on โTerence Tao โ How the worldโs top mathematician uses AIโ from Dwarkesh Patel Watch the original video
Decoding Discovery: Terence Tao on AIโs Impact on the World of Math and Science
Terence Tao, often hailed as the โgreatest living mathematician,โ rarely needs an introduction. Yet, in a recent conversation with Dwarkesh Patel, Tao offered a fascinating perspective on how artificial intelligence is not just assisting, but fundamentally reshaping the landscape of scientific discovery. By drawing parallels between AI and historical scientific breakthroughs, particularly Johannes Keplerโs monumental work, Tao illuminated the profound shifts occurring in how we generate, verify, and ultimately understand scientific progress.
Kepler, the High-Temperature LLM
The discussion began with a journey back to the 17th century, to the story of Johannes Kepler and his quest to understand planetary motion. Building on the heliocentric model proposed by Copernicus (who himself was influenced by Aristarchus), Kepler initially clung to a beautiful, yet ultimately flawed, theory. Copernicus had suggested that planets orbited the Sun in perfect circles, a concept that largely fit observations gathered over centuries by Greek, Arab, and Indian astronomers.
Kepler, however, observed curious geometric relationships in Copernicusโs predicted orbit sizes. He posited a theory so elegant it seemed divinely inspired: the orbits of the six known planets could be nested within the five perfect Platonic solids (cube, tetrahedron, icosahedron, octahedron, dodecahedron). For instance, if the Earthโs orbit was enclosed in a cube, the outer sphere enclosing that cube almost perfectly matched Marsโs orbit. This theory, to Kepler, represented mathematical perfection aligning with Godโs design.
To confirm his grand vision, Kepler desperately needed accurate data. At the time, only one such dataset existed, meticulously compiled by the eccentric and wealthy Danish astronomer Tycho Brahe. Brahe had convinced the Danish government to fund an entire island observatory, where for decades he made naked-eye observations of planets like Mars and Jupiter, achieving a precision ten times greater than any before him.
Kepler began working with Brahe, but the older astronomer jealously guarded his data, releasing only fragments. Eventually, Kepler, in a move that might be considered scientific espionage, copied the data, leading to a contentious battle with Braheโs descendants. Armed with this unprecedented trove of information, Kepler set about verifying his beautiful Platonic solid theory.
To his profound disappointment, the data didnโt quite fit. His theory was off by about 10%. Years of relentless work, trying every conceivable โfudgeโ and adjustment, followed. It was a Herculean effort of data analysis that eventually led him to a shocking conclusion: planetary orbits were not perfect circles, but ellipses. This discovery formed Keplerโs first law. He then worked out his second law โ that planets sweep out equal areas in equal times โ and ten years later, after further arduous data collection, his third law, relating a planetโs orbital period to its distance from the Sun.
Kepler had no underlying theory to explain why these laws were true; they were purely empirical regularities derived from data. It would take Isaac Newton a century later to provide the theoretical framework (gravity and laws of motion) that unified and explained all three.
Dwarkesh Patel then drew a striking analogy: โKepler was a high-temperature LLM.โ He meant that Kepler, throughout his career, was essentially โtrying random relationships,โ some of which were wildly speculative (like his asides on โThe Harmonics of the Worldโ connecting planetary orbits to musical notes and astrology). Yet, when these wild ideas were rigorously tested against Braheโs โverifiable data bank,โ a few of them, like the elliptical orbits and the cube-square law, led to profound scientific progress. Tao concurred, emphasizing that this combination of prolific, often unconventional, idea generation coupled with stringent verification was key.
The Shifting Bottleneck of Scientific Progress
Traditionally, the history of science celebrates โeureka genius momentsโ โ the generation of novel ideas. A scientific problem typically involves many steps: identifying a fruitful problem, collecting data, strategizing data analysis, formulating a hypothesis, validating it, and communicating the findings. Idea generation has always been the โprestige part.โ
However, Tao observes that science has evolved. Classically, the two main paradigms were theory and experiment. The 20th century added numerical simulation, and the late 20th century ushered in the era of โbig data.โ Now, much scientific progress is driven by analyzing massive datasets first, then drawing patterns and deducing hypotheses โ almost reversing the classic scientific method of forming a hypothesis and then collecting data to test it. Kepler, in a sense, was an early data scientist, but even he started with preconceived theories before diving into Braheโs data.
This is where AI enters the picture. Tao argues that AI has driven the cost of idea generation โdown to almost zero,โ much like the internet drove down the cost of communication. This is an โamazing thing,โ but it doesnโt automatically create abundance or progress. Instead, it shifts the bottleneck.
โNow the bottleneck is different,โ Tao explains. โWeโre now in a situation where suddenly people can generate thousands of theories for a given scientific problem. Now we have to verify them, evaluate them.โ Human reviewers are already overwhelmed by AI-generated submissions to journals. The challenge is no longer generating ideas, but assessing which ones genuinely advance the field and which are โdead ends or red herrings.โ This is a problem we donโt yet know how to solve at scale.
The Challenge of Evaluating โTrue Progressโ
If we soon have โbillions of AI scientistsโ churning out theories, how do we discern real progress? Tao points out that human science has faced similar, albeit smaller-scale, challenges. He references the development of the โbitโ at Bell Labs in the 1940s โ a unifying concept with implications across many fields, not just the initial engineering problem it addressed. How do you identify such a transformative idea amidst millions of papers, especially if it initially lacks broad unifying power?
โA lot of itโs the test of time,โ Tao notes. Many great ideas, like deep learning itself or the transformer architecture (the foundation of modern LLMs), didnโt receive immediate widespread acceptance. Their fruitfulness became apparent only much later, as other scientists built upon them. The adoption of an idea can also depend on cultural and societal factors, like the base-ten numeral system becoming standard despite other possibilities.
Evaluating an ideaโs worth isnโt purely objective; itโs deeply intertwined with its context, past and future. Progress often isnโt linear. Copernicusโs heliocentric model, while conceptually simpler, was initially less accurate than the geocentric Ptolemaic system, which had been refined over a millennium with countless ad hoc fixes. It took Keplerโs elliptical orbits to make the heliocentric model superior.
โScience is always a work in progress,โ Tao states. โWhen you only get part of the solution, it looks worse than a theory which is incorrect but somehow has been completed to the point where it kind of answers all the questions.โ He cites Leibnizโs disagreement with Newtonโs action-at-a-distance gravity, and Newtonโs own puzzlement over inertial and gravitational mass equivalence โ mysteries only resolved by Einstein centuries later.
Sometimes, progress isnโt about adding new theories but โdeleting some assumptions that you have in your mind.โ The long-held Aristotelian notion that objects naturally rest made heliocentrism seem implausible (โHow come we werenโt all falling over?โ). Similarly, Darwinโs theory of evolution challenged the deeply ingrained idea of static species.
Tao sees us in a โcognitive version of the Copernican revolutionโ right now. We once believed human intelligence was the center of the universe, but now weโre confronting diverse forms of intelligence with different strengths and weaknesses. โOur assessment of which tasks require intelligence, which ones donโt, has to be reordered quite a bit.โ
Darwin vs. Newton: The Role of Communication and Data Loops
Dwarkesh Patel brought up a curious point from Edward Dolnickโs book The Clockwork Universe: Darwinโs Origin of Species (1859) came out two centuries after Newtonโs Principia Mathematica (1687), yet Darwinโs theory conceptually seems simpler. Thomas Huxley famously said upon reading Origin: โHow stupid not to have thought of that.โ No one said that about Principia. Why the delay?
Tao suggests it relates to the nature of evidence and communication. Newton could present equations that immediately predicted observable phenomena, like the Moonโs orbit. Darwinโs evidence for natural selection, while overwhelming, was โcumulative and retrospective.โ Lucretius had a similar idea in the first century BC, but it gained no traction because he couldnโt โrun some experiment and force people to pay attention.โ
Beyond data, communication is vital. Darwin was an โamazing science communicator,โ writing in plain English without equations, synthesizing disparate facts into a compelling vision. Newton, by contrast, wrote in Latin, invented entirely new mathematics to explain his work, and was notoriously secretive and competitive. His work only became widespread decades later when others simplified and explained it.
โThe art of exposition and making a case and creating a narrative is also a very important part of science,โ Tao emphasizes. โIf you have the data, it helps, but people need to be convinced, otherwise they will not push it further or take the initial investment to learn your theory and really explore it.โ This โsoft, squishy thingโ โ the social aspect of science, painting a narrative of gaps and future possibilities โ is โreally hard to reinforcement learn on.โ Perhaps, Tao muses, this persuasive, narrative-driven aspect will forever remain the human side of science.
AI for Math: The Erdลs Problem Frontier
Turning to the direct impact of AI on mathematics, Tao shared insights from the world of Erdลs problems โ a collection of 1100 open problems posed by the legendary mathematician Paul Erdลs. Recently, AI programs have solved about 50 of these. However, Tao notes, โit does seem like we have picked the low-hanging fruit.โ The initial flurry of โpure AI solutionsโ (where AI โone-shotsโ a problem) has slowed.
Tao uses a vivid analogy: imagine a mountain range with cliffs of varying heights (3 feet, 6 feet, 15 feet, mile-high), all shrouded in darkness. Weโre trying to climb as many as possible. AI tools are like โjumping machines that can jump two meters in the air, higher than any human.โ They sometimes jump in the wrong direction or crash, but sometimes they can reach the tops of the lowest walls that humans couldnโt. The recent success with Erdลs problems was this exciting period where AIs found and scaled these โlow ones.โ
A key limitation, however, is that current AI tools are โreally bad at creating partial progress or identifying intermediate stages that you should focus on first.โ They either succeed or fail. This contrasts with human mathematicians who โhill climb, make little markers, and try to identify partial things.โ
This leads to a โbearishโ and โbullishโ interpretation of AIโs current state. The bearish view is that AIs are only reaching certain โheights of wallโ โ not as high as humans. The bullish view, which Tao leans towards, is that once AIs achieve a certain โwaterlineโ of capability, they can fill every single problem available at that waterline. โWe canโt make a million copies of you and give each of them a million dollars of inference compute and have you do a hundred years of subjective time research on a million different problems at the same time. But once AIs reach Terence Tao-level, they could do that.โ
Complementarity and the Future of Math
Tao sees a future of โvery complementary science.โ AI excels at breadth, while human experts excel at depth. Our current scientific paradigms, focused on depth, need to be redesigned to leverage AIโs breadth. โWe should have a lot more effort in creating very broad classes of problems to work on rather than one or two really deep, important problems.โ AIs could โmap outโ entirely new fields, making โeasy observationsโ and identifying โcertain islands of difficulty,โ which human experts could then tackle.
This shift is already impacting Taoโs own productivity. While the โcoreโ of his work โ solving the most difficult part of a math problem โ still relies on pen and paper, AI has dramatically sped up โauxiliary tasks.โ Things like generating code, creating complex plots, conducting deeper literature searches, or even reformatting parentheses in a paper now take minutes instead of hours. This allows him to enrich his papers with more data, visuals, and context. โThe type of papers that I would write today, if I had to do them without AI assistance, would definitely take five times longer.โ However, he also notes that if he were to write a paper of the same functional depth as one from 2020, without the added features, the time savings arenโt as dramatic. AI makes papers โricher and broader, but not necessarily deeper.โ
Tao believes AI will revolutionize the โexperimental side of math.โ Unlike other sciences with their theoretical and experimental divisions, math has been almost entirely theoretical. Now, with AI, mathematicians can run large-scale experiments to test what methods work and what donโt, gather data on problem-solving effectiveness, and explore mathematics โat scaleโ โ an idea still in its infancy.
Artificial Cleverness vs. Artificial Intelligence
A crucial distinction Tao makes is between โartificial clevernessโ and โartificial intelligence.โ For Tao, true intelligence in a collaborative problem-solving context involves โadaptivity and continual improvement of the idea over time.โ Itโs about systematically mapping out what works and what doesnโt, evolving the strategy through discussion.
Current AIs, he says, โcan mimic this a little bit.โ They can โjump and fail, and jump and fail.โ But they canโt โjump a little bit, reach some handhold, stay there, pull other people up, and then try to jump from there.โ There isnโt a โcumulative process which is built up interactively.โ Instead, itโs largely โtrial and error and just repetition: brute force.โ When you run a new session with an AI, it has โforgotten what it just did.โ Its โunderstanding of math has not progressed.โ
This highlights the current frontier: while AI can apply existing techniques with impressive speed and accuracy, and even combine obscure methods to solve previously intractable problems (especially those lacking extensive literature), it struggles with true novelty and cumulative, evolving understanding. The โholes in the argument where none of the things are workingโ still require human ingenuity.
The progress, Tao concludes, is โsimultaneously amazing and disappointing.โ While we quickly acclimatize to AIโs stunning capabilities (like Google search two decades ago, or current college-level math-solving), the core challenge of fostering genuine, evolving intelligence remains. The future of science, as envisioned by Terence Tao, will be one where human depth and artificial breadth converge, leading to an โunrecognizableโ landscape of discovery โ a new era defined by an unprecedented partnership between human and machine.
Based on โInside Palantir: Building Software That Matters | Shyam Sankar on a16zโ from a16z Watch the original video
Americaโs Clarion Call: Reclaiming Innovation and Mobilizing for a New Era
In a world increasingly fraught with geopolitical tension and the specter of โhorrendous barbarism,โ a profound shift is underway in how America views its national security and technological future. Shyam Sankar, a titan from Palantir known for his behind-the-scenes influence, has stepped into the public spotlight with a stark message: the nation is at a critical juncture, facing a risk of โsuicide, not homicide,โ if it fails to mobilize its latent strengths. His insights, shared on a16z, offer a powerful diagnosis of Americaโs current challenges and a vigorous prescription for reclaiming its innovative edge.
From Shadow Architect to Public Provocateur
For years, Shyam Sankar was the โOG fixerโ at Palantir, a figure whispered about in Silicon Valley as the architect behind countless careers and a quiet force in defense tech. Founders like Trey Stevens of Anduril credit Sankar with single-handedly shaping their trajectories, introducing them to Palantir, nurturing their growth, and even โgiving them wings to fly awayโฆ to start something new.โ Yet, his influence remained largely out of the public eye until a few years ago.
What prompted this shift? Sankar describes it as โequal parts an act of desperation and act of optimism.โ After years witnessing the โfrog boilโ of stagnation within the Pentagon, a series of geopolitical eventsโfrom Russiaโs annexation of Crimea and Chinaโs militarization of the Spratly Islands to the failure of the JCPOA and the October 7th attacks in Israelโcatalyzed his conviction. โIt was kind of a radicalizing moment,โ he recounts. โWhat is going on here? We need to act.โ
Simultaneously, Sankar observed a re-emergence of entrepreneurial energy outside the building, with founders eager to build in the national interest. This confluence of urgent need and renewed spirit compelled him to articulate a โfundamental diagnosisโ: America had โaccidentally turned our back onโ the very things that led to past victories. His seminal โFirst Breakfastโ piece on defense reformation marked his public debut as a โstrident voice for what needs to happen in America.โ
The Erosion of Americaโs Defense Edge: A Historical Reckoning
Sankar argues that Americaโs current predicament stems from a series of post-Cold War missteps. Following the Soviet Unionโs collapse, the U.S. embraced a โpeace dividend,โ leading to a dramatic restructuring of its defense industrial base. The infamous โLast Supperโ dinner saw the number of prime defense contractors shrink from 51 to a mere five. The conventional wisdom suggested this consolidation simply reduced competition. Sankar offers a more nuanced, and troubling, explanation: โconsolidation bred conformity.โ
This conformity ushered in the โfinancialization of defense,โ where companies prioritized financial metrics like dividends and buybacks over innovation and growth. This environment, Sankar asserts, became hostile to โfoundersโ โ the daring, often heretical individuals who had driven Americaโs greatest military advancements.
He paints a vivid picture of these historical โhereticsโ:
- Heimman Rickover: The โfather of the nuclear navy,โ who battled against the Navyโs will and even Oppenheimerโs skepticism to build a revolutionary fleet.
- John Boyd: The notoriously difficult fighter pilot whose โheresyโ regarding energy-maneuverability theory and the OODA loop proved instrumental in the F-16โs design and the swift victory in Gulf War I.
- Andrew Higgins: The innovative Scots-Irishman whose โHiggins boatโ โ initially rejected by the Navy โ ultimately constituted 92% of all landing craft in World War II, enabling critical operations like D-Day.
These figures, Sankar emphasizes, were often โagainst the institution, the bureaucracy, the process.โ But in the post-Cold War era, such personalities were โexpungedโ and sought refuge in other parts of the American economy, particularly tech.
Compounding this issue was the unique nature of the Department of Defense as a โmonopsonyโ โ a single buyer. This allowed the DoD to impose rigid constraints on its suppliers, effectively creating โGalapagos tortoisesโ: exquisite, highly specialized companies perfectly adapted to the DoDโs unique ecosystem, but uncompetitive on the โmainlandโ of the broader economy. For years, there was โno front doorโ for outsiders to contribute, save for the intelligence communityโs In-Q-Tel.
Reclaiming the โAmerican Wayโ: A Vision for Mobilization
Sankarโs book, โMobilize,โ champions a return to a more integrated โAmerican industrial base.โ He points to World War II, where companies like Chrysler built both Minuteman missiles and minivans, and every consumer purchase subtly subsidized national security. This contrasts sharply with today, where 86% of major weapon system spending goes to defense specialists, up from just 6% in 1989.
The solution, Sankar argues, lies in:
- Inspiring Latent Heretics: The most crucial step is to empower and protect the founders and unconventional thinkers, both inside and outside government, who are willing to challenge the status quo. Leaders must โset the conditions to empower the heretics.โ
- Voluntary Civil-Military Fusion: While China enforces civil-military fusion, the U.S. should make voluntary collaboration irresistible. This means leveraging the vast R&D spending of the private sector and re-establishing pathways for skilled individuals to serve, much like World War IIโs direct commissioning of 100,000 experts.
- Founders Everywhere: Recognizing that founders arenโt just outside government. Figures like Colonel Drew Cukor, the โfather of Maven,โ exemplify this. Cukor, a Marine Colonel, driven by the operational failure that led to the Yazidi genocide, spearheaded the integration of AI into the Pentagon against immense bureaucratic resistance, facing absurd accusations and investigations. His story is a testament to the โincorruptibleโ commitment required of true heretics.
AI as the Slingshot for Re-industrialization
Central to Sankarโs vision is the transformative power of Artificial Intelligence. He rejects the โAI dumerismโ that portrays AI as an uncontrollable force. Instead, he asserts, โHumans are going to use AI to do X. Thereโs a choice here.โ AI, he believes, offers a historic opportunity to rectify the breakdown between wage growth and GDP growth that began in the 1970s.
โThere is an opportunity to give the American worker superpowers with AI,โ Sankar declares. This isnโt about symmetrical competition with China, but about a โDavidโs slingshotโ โ using technology to re-industrialize the country in entirely new ways. He highlights companies like Hadrian, which are achieving 50-100x productivity gains through tech-driven manufacturing.
This re-industrialization, Sankar argues, necessitates a return to the โcollocation of R&D and production.โ The โgreat lie of globalizationโ was the separation of innovation from production. As SpaceX demonstrates, innovation thrives when R&D engineers are on the factory floor, enabling rapid feedback loops. If you donโt make the thing, you canโt innovate on how you make the thing.
Inside the Army: A New Kind of Mobilization
Sankar isnโt just advocating; heโs acting. He recently joined the Army, along with other tech luminaries like Bob Muglia (former OpenAI chief research officer), Boz (Meta CTO), and Andrew Weil (former OpenAI chief product officer). This initiative, under General George and Secretary Driscoll, aims to bring senior tech expertise directly into the military.
His focus includes long-term force structure planning and treating software as a โmalleable weapon system.โ What has surprised him most from the inside? The โquality of talent in our green suitorsโ โ young, often self-taught individuals building โthe most compelling AI applicationsโ driven by โexistential stakes.โ These junior personnel, empowered by AI tools, can now build and deploy ideas in weeks, fostering a โbottoms-up innovation mission commandโ that plays to Americaโs unique military strengths.
The SAS Apocalypse and the Future of the Economy
Sankar also offers a provocative take on the โSAS apocalypse,โ the idea that AI will commoditize much of existing software. He introduces a critical distinction:
- Beta Software: Software that makes you similar to everyone else. This, he predicts, will โreally struggle,โ as AI allows custom โvibe-codingโ of solutions.
- Alpha Software: Software that allows you to express your differentiation and competitive advantage. Platforms focused on alpha will thrive, becoming โtoolkitsโ that enable unique strategies.
He argues that much of the โsoftware industrial complexโ has historically focused on โcan I sell it?โ rather than โdid it add value?โ The COVID-19 pandemic, where multi-billion dollar ERP systems collapsed while Zoom and Teams enabled remote work, served as a โSputnik momentโ for the industry.
Looking at the AI stack, Sankar believes value will accrue primarily at two layers: the โchips layerโ and the โAI infrastructure layerโ (what Palantir calls โontologyโ). Models, he suggests, are increasingly commoditized.
Ultimately, Sankarโs vision for AIโs economic impact is optimistic, provided humans exercise their agency. AI can reverse the trend of financial engineering over real engineering, empower workers, and drive re-industrialization. He champions a return to a โfounder personalityโ that prioritizes building and engineering, where the pathway to CEO runs through the CTO, as Elon Musk suggests. The goal, he concludes, isnโt to replace people but to make them better, to build โIron Man suitsโ for every worker, ensuring America remains dominant not just in technology, but in its very future.
Shyam Sankarโs voice is a powerful reminder that the future is not predetermined. Itโs a choice, shaped by leadership, courage, and a renewed commitment to the innovative spirit that once defined America. The call to mobilize is clear, and the opportunity to reclaim that vigor is now.
Based on โSilicon Valleyโs Big Bets on War Pay Off, and the Trump Family Business Looks to Transylvaniaโ from New York Times Podcasts Watch the original video
The Shifting Sands of Power and Profit: From Silicon Valleyโs War Bets to Trumpโs Transylvanian Gambit
In a world grappling with escalating conflicts, rapid technological shifts, and persistent economic anxieties, seemingly disparate headlines often converge to paint a vivid picture of our changing times. From tech giants pivoting to defense to a former presidentโs family business eyeing an unlikely real estate venture, and from the quiet desperation driving plasma donations to the unsettling rise of AI in literature, these stories reveal the complex currents beneath the surface of daily news.
Silicon Valleyโs New Front: The Business of War
The escalating tensions surrounding Iran have sent ripples across global markets and geopolitical strategies. While President Trump has publicly stated heโs โnot putting troops anywhereโ currently, he also hasnโt ruled out deploying forces โif needed.โ In a dramatic reversal of long-standing US policy, the administration is even considering unsanctioning Iranian oil โ a desperate move to bring down prices, effectively encouraging Iran to sell more oil even amidst conflict.
At the Pentagon, the financial stakes are soaring. The Defense Department is seeking an astonishing $200 billion to continue funding the war effort. This colossal sum, nearly a quarter of the entire annual defense budget, is already raising eyebrows among moderate Republicans. Yet, it also signals an unprecedented opportunity for a specific, once-reluctant sector: Silicon Valley.
For years, the tech industry, particularly companies like Google, Meta, and OpenAI, shied away from defense contracts. Google employees famously protested their companyโs work with the Department of Defense, invoking the โdo no evilโ motto as a moral compass against engaging in the business of war. But in a striking shift, those optics have dramatically changed.
Now, mammoth tech companies and nimble startups alike are actively pivoting into defense technology, striking deals worth hundreds of millions, even billions, of dollars. Theyโre building everything from advanced weapon systems to sophisticated software for the US government. Palantir, a data analytics company, has developed Project Maven, a system that aids the US government in selecting targets for air strikes. Google and OpenAI are leveraging their AI prowess to assist generals in the field. Even former Google CEO Eric Schmidt is leading projects to build counter-drone systems, now actively deployed by US assets in Iran to defend against Iranian drones.
With President Trump having allocated over a trillion dollars in defense spending, much of it is earmarked for the very kind of cutting-edge defense technology Silicon Valley is eager to produce. As global conflicts intensify, the once-unpopular โbusiness of warโ has become a lucrative frontier for the tech world, demonstrating a profound re-evaluation of ethical boundaries in pursuit of profit and strategic influence. This shift is further underscored by reports that Ukraine, having gained years of experience fighting off Russian drones (many of which were Iranian-made), has become a crucial source of expertise for other Middle Eastern nations seeking to defend against similar threats.
Trumpโs Transylvanian Ambition: Beyond the Glamour
Beyond the battlefields and boardrooms, another intriguing story unfolds in the unlikely setting of Transylvania, Romania. A new investigation by The New York Times has unearthed details of an unannounced real estate deal pursued by President Trumpโs family business. Unlike the glamorous tourist destinations like Bali and the Maldives where the Trump Organization typically operates, this project is slated for a semi-abandoned site alongside a military base and several enormous landfills.
Rebecca Ruiz, an investigative reporter for The Times who visited the site, described a landscape far from luxury: โYou could smell the stench. There were packs of wild dogs wandering. It felt like an unusual place to envision a luxury apartment complex and golf course.โ
The choice of location is perplexing to many, but it aligns with a pattern observed in Trumpโs second term: the Trump Organization actively seeks deals in places where the former president enjoys significant popularity. A Gallup poll last year revealed that over half of Romania held a favorable view of Trumpโs performance as president, making the country an outlier among EU nations. This unannounced Transylvanian venture, with its stark contrasts and strategic location, highlights the unique blend of politics and commerce that defines the Trump brandโs global real estate ambitions.
Americaโs Quiet Struggle: The Rise of the Plasma Economy
Closer to home, a different kind of economic reality is playing out in the unassuming middle-class suburbs of cities like Houston, Chicago, and Las Vegas. Businesses where individuals can donate plasma for cash are popping up with increasing frequency, shifting from their traditional clusters in low-income neighborhoods. Now, these centers are found nestled beside Orange Theory Fitness gyms and Charles Schwab offices.
Reporters from The Times visited several new Texas locations, finding long lines of people who never imagined theyโd be selling their plasma: a tech worker in his 30s, a sixth-grade special education teacher, a night-shift nurse. Most reported visiting the clinic twice a week, the maximum allowed under FDA regulations, to earn about $70 per donation.
This phenomenon underscores the quiet financial strain many middle-income Americans are experiencing, where an extra $70 can make a tangible difference for groceries or gas. One industry researcher described plasma centers as a โshadow safety net,โ offering a side income much like driving for Uber or Lyft.
The United States is a global anomaly in this regard, providing roughly 70% of the worldโs plasma because itโs one of the few countries that allows payment for donations โ a practice the World Health Organization discourages. While donating plasma is generally considered safe, there is remarkably little research on the long-term effects of frequent donations, adding a layer of concern to this multi-billion dollar industry built on the financial needs of its citizens.
The AI Author: Publishingโs New Frontier
Finally, the literary world is grappling with its own disruption: the rise of artificial intelligence. A buzzy new horror novel, โShy Girl,โ was set for release this spring, a revenge story about a woman held hostage. Its UK edition had been out for months, but the US publisher, Hachette, abruptly dropped the book over claims it was written with the help of AI.
This marks what appears to be the first commercial novel from a major publishing house to be pulled due to evidence of AI use. Early readers had voiced suspicions online, pointing to โnonsensical metaphors and repetitive bits.โ A publishing industry consultant, who ran the book through three different AI detection programs, confirmed that all three found text โlikely to be at least partly generated by AI,โ highlighting โcertain odd phrases like, โI pressed the phone to my lips. The screen cool and unyielding.โโ
The author has denied using AI herself, instead claiming that someone she hired to help edit the story was responsible. This incident has laid bare the publishing industryโs unpreparedness for this burgeoning technology. While most publishing contracts require authors to affirm their work is original, few companies have established measures or safeguards to verify it. As one consultant grimly noted to The Times, โAI bleeding into books is not merely inevitable. Weโre in the midst of it.โ
These diverse narratives, from the ethical dilemmas of tech and war to the surprising ventures of political figures, the quiet struggles of everyday Americans, and the disruptive force of AI, collectively illuminate the complex and rapidly evolving landscape of our modern world. They are the headlines that, when examined together, offer a deeper understanding of the forces shaping our present and future.
ํ๊ตญ์ด
โThe rise and fall of civilizations | Eric Cline: Full Interviewโ โ Big Think ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
๊ณ ๋ ๋ฌธ๋ช ์ โํผํํธ ์คํฐโ: 3์ฒ ๋ ์ ๋ถ๊ดด๊ฐ ์ค๋๋ ์ฐ๋ฆฌ์๊ฒ ๋์ง๋ ๊ฒฝ๊ณ
๊ธฐ์์ 1200๋ ๊ฒฝ, ์ง์คํด์ ๊ทผ๋ ์ง์ญ์ ์ญ์ฌ์ ์ ๋ก์๋ ๋๊ฒฉ๋ณ์ ๊ฒช์๋ค. ์๋ฐฑ ๋ ๊ฐ ๋ฒ์ฑํ๋ ๊ฐ๋ ฅํ ๋ฌธ๋ช ๋ค์ด ๋ง์น ๋๋ฏธ๋ ธ์ฒ๋ผ ์ฐ์์ ์ผ๋ก ๋ฌด๋์ ธ ๋ด๋ฆฐ ๊ฒ์ด๋ค. ์ด ์๊ธฐ, ๊ณผ์ฐ ๋ฌด์จ ์ผ์ด ์์๋ ๊ฑธ๊น? ๊ณ ๊ณ ํ์์ด์ ๊ณ ๋ ์ญ์ฌ๊ฐ์ธ ์๋ฆญ ํด๋ผ์ธ(Eric Cline) ๊ต์๋ ๊ทธ์ ์ ์ ใBC 1177: ๋ฌธ๋ช ์ด ๋ถ๊ดดํ ํด(1177 BC: The Year Civilization Collapsed)ใ์ ํ์์ ใBC 1177 ์ดํ(After 1177 BC)ใ๋ฅผ ํตํด ์ด ์์๊ป๋ผ ๊ฐ์ ๋ถ๊ดด์ ์์ธ์ ํํค์น๊ณ , ์ค๋๋ ์ฐ๋ฆฌ๊ฐ ์ง๋ฉดํ ๋ฌธ์ ์ ๋ํ ์ฌ์คํ ํต์ฐฐ์ ์ ์ํ๋ค.
ํด๋ผ์ธ ๊ต์๋ ํ๊ธฐ ์ฒญ๋๊ธฐ ์๋(Late Bronze Age, ์ฝ ๊ธฐ์์ 1700๋ ~ ๊ธฐ์์ 1200๋ )๋ฅผ ์ธ๋ฅ ์ญ์ฌ์ ๊ฐ์ฅ ์ค์ํ ์๊ธฐ ์ค ํ๋๋ก ๊ผฝ๋๋ค. โ์ด ์๊ธฐ๋ ์ฌ๋๋ค์ด ์ง์คํด ์ ์ญ์ ๊ฑธ์ณ ์ ๋ก ์๋ ๋ฐฉ์์ผ๋ก โ์ธ๊ณํโ๋์ด ์์๋คโ๋ ๊ทธ์ ์ค๋ช ์ฒ๋ผ, ๋น์์ ์ํธ ์ฐ๊ฒฐ์ฑ์ ํ๋์ ๋๋๋๋ก ๋ฎ์ ์๋ค. 3์ฒ ๋ ์ ์ ์ฌ๊ฑด์ด ์ค๋๋ ์ฐ๋ฆฌ์๊ฒ ์์ฌํ๋ ๋ฐ๊ฐ ํฌ๋ค๋ ์ฃผ์ฅ์ ๋ฐ๋ก ์ด๋ฌํ ์ ์ฌ์ฑ์์ ์ถ๋ฐํ๋ค.
์ฌ๋ผ์ง ํฉ๊ธ๊ธฐ: ํ๊ธฐ ์ฒญ๋๊ธฐ ์๋์ ๋ฒ์
ํด๋ผ์ธ ๊ต์๊ฐ ์ง์นญํ๋ โ๊ณ ๋ G8โ์ ํ๊ธฐ ์ฒญ๋๊ธฐ ์๋ ์ง์คํด์ ๊ทผ๋์ ์ง๋ฐฐํ๋ ์ฃผ์ ๋ฌธ๋ช ๋ค์ ์ผ์ปซ๋๋ค. ๊ทธ๋ฆฌ์ค์ ๋ฏธ์ผ๋ค์ธ(Mycenaeans)๊ณผ ๋ฏธ๋ ธ์์ธ(Minoans) (ํธ๋ก์ด ์ ์์ ๋ฐฐ๊ฒฝ), ์๋ํจ๋ฆฌ์(ํ ํ๋ฅดํค์)์ ํํ์ดํธ(Hittites), ๋ฉ์ํฌํ๋ฏธ์(ํ ์ด๋ผํฌ)์ ์์๋ฆฌ์(Assyria)์ ๋ฐ๋น๋ก๋์(Babylonia), ์ด์งํธ(Egypt), ํคํ๋ก์ค(Cyprus), ๊ทธ๋ฆฌ๊ณ ๊ฐ๋์(Canaan)์ ์ฌ๋๋ค์ด ๊ทธ๋ค์ด๋ค. ์ด๋ค์ ๋จ์ํ ์ด์์ ๋์ด, ๊ธด๋ฐํ ์์ ์ , ์ธ๊ต์ , ์ฌ์ง์ด ํผ์ธ ๊ด๊ณ๋ฅผ ํตํด ํ๋์ ๊ฑฐ๋ํ โ์๊ท๋ชจ ์ธ๊ณ ๋คํธ์ํฌ(Small World Network)โ๋ฅผ ํ์ฑํ๊ณ ์์๋ค.
โ์๋ก์๊ฒ ์ธ ๊ฑธ์ ์ด๋ด์ ์๋ ๊ด๊ณโ๋ผ๊ณ ํด๋ผ์ธ ๊ต์๊ฐ ์ค๋ช ํ๋ ์ด ๋คํธ์ํฌ๋, ์์ ๊ณผ ์ธ๊ต์ ํ๋ฐํ ๊ต๋ฅ๋ฅผ ํตํด ์ ์ง๋์๋ค. ์ด์งํธ๋ ๊ธ์, ๊ทธ๋ฆฌ์ค๋ ์์, ํคํ๋ก์ค๋ ๊ตฌ๋ฆฌ๋ฅผ ๊ณต๊ธํ์ผ๋ฉฐ, ์ฃผ์์ ๋ฉ๋ฆฌ ์ํ๊ฐ๋์คํ(๋ฐ๋คํ์จ ์ง์ญ)์์ ์์ฒ ํฌ๋ก๋ฏธํฐ๋ฅผ ์ด๋ํด ์๋ค. ์ด๋ค์ ๋จ์ํ ์ฌ์นํ์ ๋์ด, ์ฒญ๋๊ธฐ ์๋์ ํต์ฌ ์์์ธ ์ฒญ๋(๊ตฌ๋ฆฌ 90%, ์ฃผ์ 10%)์ ๋ง๋ค๊ธฐ ์ํด ์๋ก์ ์์์ ์์กดํ ์๋ฐ์ ์์๋ค. ์ฌ๋ฆฌ๋ธ์ , ์์ธ, ๊ณก๋ฌผ ๋ฑ ์๋ ์์๋ ํ๋ฐํ ๊ต์ญ๋์์ผ๋ฉฐ, ์ฌ์ง์ด ๊ฐ์ฃฝ ์ ๋ฐ์ด๋ ๊ธ ์๊ฐ ๋จ๊ฒ ๊ฐ์ ์ค์ ๋ฌผํ๋ค๋ ์๋ค ์ฌ์ด์ โ์ ๋ฌผ ๊ตํโ์ด๋ ์์ ํ๋์ ํตํด ์ค๊ณ ๊ฐ๋ค.
์ด๋ฌํ ์ํธ ์์กด์ฑ์ ๋ฌธ๋ช ๋ค์ ์ ๋ก ์๋ ๋ฒ์์ผ๋ก ์ด๋์์ง๋ง, ๋์์ ์น๋ช ์ ์ธ ์ฝ์ ์ด ๋์๋ค. โ์ด๋ ๋๊ตฌ๋ ์๊ธ์์กฑํ ์ ์์๋คโ๋ ํด๋ผ์ธ ๊ต์์ ์ง์ ์ฒ๋ผ, ์ด ์ฐ๊ฒฐ๋ง์ด ๋์ด์ง๋ ์๊ฐ, ๋ชจ๋ ๊ฒ์ด ๋ฌด๋์ง ์๋ฐ์ ์์๋ค. ์ธ๊ต์ ๊ด๊ณ ์ญ์ ์ํธ ํ๋ ฅ์ ์ค์ํ ์ถ์ด์๋ค. ๋์๋ค์ ์กฐ์ฝ์ ๋งบ๊ณ ์ด๋ฅผ ๊ณต๊ณ ํ ํ๊ธฐ ์ํด ์๋ ๋ค์ ์๋ก์ ํ๋ ์ผ๋ก ๋ณด๋๋ค. ์ด์งํธ ํ๋ผ์ค ์๋ฉํธํ ํ 3์ธ์ ์์ผ๋ํค์ ํ๋ ์๋ ๋ฏธํ๋์ ๋ฐ๋น๋ก๋์ ๊ณต์ฃผ๋ค์ด ์์์ผ๋ฉฐ, ์ด๋ ๋น์ ์ธ๊ต์ ์ค์ํ ์๋จ์ด์๋ค.
๋ถ๊ดด์ ์๋ง: ๋จ์ํจ์ ๋์ด์ ๋ณตํฉ์ ์์ธ
์ค๋ซ๋์ ํ์๋ค์ ํ๊ธฐ ์ฒญ๋๊ธฐ ์๋์ ๋ถ๊ดด๋ฅผ โ๋ฐ๋ค ๋ฏผ์กฑ(Sea Peoples)โ์ ์นจ๋ต๊ณผ ๊ฐ์ ๋จ์ผ ์์ธ(monocausal)์ผ๋ก ์ค๋ช ํ๋ ค๋ ๊ฒฝํฅ์ด ์์๋ค. ๊ทธ๋ฌ๋ ํด๋ผ์ธ ๊ต์๋ ์ด๋ฌํ ์ค๋ช ์ด โ๋๋ฌด ๋จ์ํ๋คโ๊ณ ์ง์ ํ๋ฉฐ, ์ด์ ๋ ๋ณตํฉ์ ์์ธ(polycausal)์ ์ฃผ๋ชฉํด์ผ ํ๋ค๊ณ ๊ฐ์กฐํ๋ค. โ๋์ธ ๊ฐ์ง, ํน์ ๋ค ๊ฐ์ง ์์ธ์ด ๋์์ ๋๋ ๋น ๋ฅด๊ฒ ์ฐ์ด์ด ๋ฐ์ํ์ฌ, ํ ์ฌ์์์ ํ๋ณตํ ํ๋ ์์ด ๋ค์ ์ฌ์์ด ๋ฅ์ณ์๋คโ๋ ๊ฒ์ด๋ค. ๊ทธ๋ ์ด๋ฅผ โํผํํธ ์คํฐ(perfect storm)โ์ ๋น์ ํ๋ฉฐ, ๊ฐ ์์ธ์ด ์๋ก๋ฅผ ์ฆํญ์ํค๋ โ์น์ ํจ๊ณผ(multiplier effect)โ์ โ๋๋ฏธ๋ ธ ํจ๊ณผ(domino effect)โ๋ฅผ ์ผ์ผ์ผฐ๋ค๊ณ ์ค๋ช ํ๋ค.
์ด๋ฌํ ์ ๊ทผ ๋ฐฉ์์ ์ง๋ ์์ญ ๋ ๊ฐ ์๋ก์ด ๋ฐ์ดํฐ์ ์ฐ๊ตฌ ๊ฒฐ๊ณผ๊ฐ ์ถ์ ๋๋ฉด์ ๊ฐ๋ฅํด์ก๋ค. โ์ญ์ฌ๋ ๋ฐ๋ณต๋์ง ์์ง๋ง, ์ด์จ์ ์๋ค(history does rhyme even if it doesnโt repeat)โ๋ ํด๋ผ์ธ ๊ต์์ ๋ง์ฒ๋ผ, ๊ณ ๋ ์ฌํ์ ๋ถ๊ดด ๊ณผ์ ์ ์ค๋๋ ์ฐ๋ฆฌ๊ฐ ์ง๋ฉดํ ๊ธฐํ ๋ณํ, ์ด์ฃผ ๋ฌธ์ , ํฌ๋ฐ๋ฏน ๋ฑ์ ๋ณตํฉ ์๊ธฐ(poly-crisis)์ ๋ํ ์ค์ํ ๊ฒฝ๊ณ ์์ ๋์ง๋ค.
๋๊ฒฉ๋ณ์ ๋ถ๋ฅธ ์์ธ๋ค: ํผํํธ ์คํฐ์ ๊ตฌ์ฑ ์์
ํด๋ผ์ธ ๊ต์๋ ํ๊ธฐ ์ฒญ๋๊ธฐ ์๋ ๋ถ๊ดด๋ฅผ ์ด๋ํ ์ฃผ์ ์์ธ๋ค์ ๋ค์๊ณผ ๊ฐ์ด ์ ์ํ๋ค.
-
๊ฐ๋ญ๊ณผ ๊ธฐ๊ทผ(Drought and Famine): ๊ฐ๋ญ์ ๋จ์ํ ๊ธฐํ ํ์์ ๋์ด ์ฌํ ์ ์ฒด๋ฅผ ๋คํ๋๋ ์ฌ์์ด์๋ค. ๋ฆฌ์ค ์นดํํฐ(Rhys Carpenter) ๊ต์๊ฐ 1960๋ ๋์ ๋ฏธ์ผ๋ค ๋ฌธ๋ช ๋ถ๊ดด์ ์์ธ์ผ๋ก ๊ฐ๋ญ์ ์ ์ํ์ง๋ง, ๋น์์๋ ์ด๋ฅผ ๋ท๋ฐ์นจํ ์๋ฃ๊ฐ ๋ถ์กฑํ๋ค. ๊ทธ๋ฌ๋ ์ค๋๋ ์ฐ๋ฆฌ๋ ์ดํ๋ฆฌ์์์ ์ด๋์ ์ด๋ฅด๋ ๊ด๋ฒ์ํ ์ง์ญ์์ โ๋ฉ๊ฐ ๊ฐ๋ญ(mega-drought)โ์ ์ฆ๊ฑฐ๋ฅผ ํ๋ณดํ๊ณ ์๋ค. ๋๊ตด์ ์์ ์ฑ์ฅ ์ค๋จ, ๋ง๋ฅธ ํธ์์ ๊ฐ๋ฐ๋ฅ์ ํด์ ๋ฌผ ๋ถ์, ๊ทธ๋ฆฌ๊ณ ํ๋ฏธ๊ฒฝ์ผ๋ก ๊ด์ฐฐํ ๊ฝ๊ฐ๋ฃจ ๋ณํ ๋ฑ ๋ค์ํ ๊ณผํ์ ๋ฐ์ดํฐ๋ ๊ธฐ์์ 1250๋ ๊ฒฝ๋ถํฐ ์ต์ 150๋ , ๊ธธ๊ฒ๋ 300๋ ๊ฐ ์ง์๋ ๊ฐํนํ ๊ฑด์กฐ ๊ธฐํ๋ฅผ ์ฆ๋ช ํ๋ค.
์ด๋ฌํ ๊ฐ๋ญ์ ํ์ฐ์ ์ผ๋ก ๊ธฐ๊ทผ์ ๋ถ๋ฌ์๋ค. ์๋ฆฌ์ ๋ถ๋ถ ํด์์ ์ค์ํ ๊ตญ์ ํญ๊ตฌ์๋ ์ฐ๊ฐ๋ฆฌํธ(Ugarit)์์ ๋ฐ๊ตด๋ ์ ํ ํ์๋ โ์ฐ๋ฆฌ ๋์์ ๊ธฐ๊ทผ์ด ๋ค์๋ค. ๋์๋ฌ๋ผโ๋ ์ ๋ฐํ ๋ฉ์์ง๊ฐ ๋ด๊ฒจ ์๋ค. ํํ์ดํธ ์๊ตญ์์๋ ๊ธฐ์์ 1250๋ ๊ฒฝ๋ถํฐ โ๋ด ๋ ์ ๊ธฐ๊ทผ์ด ๋ค์๋ค. ์์ฌ์ ๋ฌธ์ ๋ค. ๊ณก๋ฌผ์ ๋ณด๋ด๋ฌ๋ผโ๋ ๋ด์ฉ์ ์์ ์ด ๋ฐ๊ฒฌ๋์์ผ๋ฉฐ, ์ด์งํธ๊ฐ ์ค์ ๋ก ๊ณก๋ฌผ๊ณผ ๋ง๋ฆฐ ์์ ์ ๋ณด๋ด ๊ตฌํธ ํ๋์ ํผ์ณค๋ค๋ ๊ธฐ๋ก๋ ์กด์ฌํ๋ค.
-
์ด์ฃผ์ ์นจ๋ต: โ๋ฐ๋ค ๋ฏผ์กฑโ์ ์ฌํด์(Migration and Invasion: Reinterpreting the โSea Peoplesโ): ์ ํต์ ์ผ๋ก โ๋ฐ๋ค ๋ฏผ์กฑโ์ ํ๊ธฐ ์ฒญ๋๊ธฐ ์๋ ๋ฌธ๋ช ๋ถ๊ดด์ ์ฃผ๋ฒ์ผ๋ก ์ง๋ชฉ๋์ด ์๋ค. ์ด์งํธ ํ๋ผ์ค ๋ฉ๋ฅด๋ตํ(Merneptah)์ ๋์ธ์ค 3์ธ(Ramses III)์ ๊ธฐ๋ก์ ๋ฐ๋ฅด๋ฉด, ์ด๋ค์ ๊ธฐ์์ 1207๋ ๊ณผ 1177๋ ์ ์ด์งํธ๋ฅผ ๊ณต๊ฒฉํ 9๊ฐ ์ฐํฉ ์ง๋จ์ด์๋ค. ๊ทธ๋ฌ๋ ํด๋ผ์ธ ๊ต์๋ ์ด๋ค์ ๋จ์ํ ์นจ๋ต์๊ฐ ์๋, โํผํด์์ด์ ์ต์์โ์ด๋ฉฐ โ๋ถ๊ดด์ ์์ธ์ด๋ผ๊ธฐ๋ณด๋ค๋ ์ฆ์โ์ผ๋ก ํด์ํ๋ค.
๋ฉ๊ฐ ๊ฐ๋ญ๊ณผ ๊ธฐ๊ทผ์ผ๋ก ์ธํด ์๋ถ ์ง์คํด(์์น ๋ฆฌ์, ์ฌ๋ฅด๋๋์, ์ดํ๋ฆฌ์ ๋ฑ)์์ ์ด ๊ณณ์ ์์ ์ฌ๋๋ค์ด ๋์ชฝ์ผ๋ก ๋๊ท๋ชจ ์ด์ฃผ๋ฅผ ์์ํ์ ๊ฒ์ด๋ผ๋ ์ค๋ช ์ด๋ค. ์ด๋ค ์ค โํ ๋ ์ (Peleset)โ์ผ๋ก ๋ถ๋ฆฐ ์ง๋จ์ ์ฑ๊ฒฝ์ ๋์ค๋ โ๋ธ๋ ์ ์ธ(Philistines)โ์ผ๋ก, ๊ทธ๋ฆฌ์ค ๋ฏธ์ผ๋ค ๋ฌธ๋ช ๊ถ์์ ์จ ์ฌ๋๋ค๋ก ์ถ์ ๋๋ค. ์ด๋ค์ ์๋ก์ด ์ถ์ ์ฐพ์ ์ด์ฃผํ์ง๋ง, ์ด๋ฏธ ๊ฐ๋ญ๊ณผ ๊ธฐ๊ทผ์ผ๋ก ๊ณ ํต๋ฐ๋ ๋๋ถ ์ง์คํด์ ๋์ฐฉํ์ฌ ๊ธฐ์กด ์ฃผ๋ฏผ๋ค๊ณผ ์ถฉ๋ํ๊ณ , ์ด๊ฒ์ด ์นจ๋ต์ผ๋ก ๋น์ณค์ ์ ์๋ค๋ ๊ฒ์ด๋ค. ์ฐ๊ฐ๋ฆฌํธ์ ์ ํ ํ์ ๊ธฐ๋ก๋ โ์ ์ ๋ฐฐ๋ค์ด ์๋ฅํ์ฌ ํญ๊ตฌ ๋์๋ฅผ ์ ๋ นํ๊ณ ์ฐ๊ฐ๋ฆฌํธ๋ก ์ง๊ฒฉํ๊ณ ์๋คโ๋ ๋ด์ฉ์ ์ด๋ฌํ ์ธ๋ถ ์นจ๋ต์ ์ฆ๊ฑฐ๋ฅผ ๋ณด์ฌ์ค๋ค.
๊ทธ๋ฌ๋ ๋ชจ๋ ํ๊ดด๊ฐ ์ธ๋ถ ์นจ๋ต์์ ์ํ ๊ฒ์ ์๋์๋ค. ๋ฏธ์ผ๋ค์ ๊ฐ๋์์ ํ์ด๋ฅด(Hazor) ๊ฐ์ ์ ์ ์ง์์๋ ๊ถ์ ๊ณผ ์ ์ ์ ๋ถํ์ง๋ง, ์ผ๋ฐ ์ฃผ๋ฏผ๋ค์ ์ฃผํ์ ์จ์ ํ ๋จ์ ์๋ ๊ฒฝ์ฐ๊ฐ ๋ฐ๊ฒฌ๋๋๋ฐ, ์ด๋ ์ธ๋ถ ์นจ๋ต๋ณด๋ค๋ ์๋ ๋ถ์กฑ๊ณผ ์ฌํ ๋ถํ๋ฑ์ผ๋ก ์ธํ โ๋ด๋ถ ๋ฐ๋(internal rebellion)โ์ ๊ฐ๋ฅ์ฑ์ ์์ฌํ๋ค.
-
์์ฐ์ฌํด: ์ง์ง์ ์ํ(Natural Disasters: The Threat of Earthquakes): ์ผ๋ถ ๋์์ ํ๊ดด๋ ์ธ๊ฐ์ ์์ด ์๋ ์์ฐ์ ํ, ์ฆ ์ง์ง์ ์ํ ๊ฒ์ด์๋ค. ๊ฑด๋ฌผ ๋ฒฝ์ด ๋น๋ค์ด์ง๊ฑฐ๋, ์์นํ ๋ฌธํ์ ์๊ธฐ๋์ด ๋น ์ ธ ์๋ ๋ฑ ์ง์ง์ ๋ช ํํ ํ์ ๋ค์ด ๋ฐ๊ฒฌ๋๋ค. ํนํ ๊ธฐ์์ 1225๋ ์์ 1175๋ ์ฌ์ด ์๊ฒํด์ ๋๋ถ ์ง์คํด ์ง์ญ์์๋ โ์ง์ง ํญํ(earthquake storm)โ ๋๋ ์ง์งํ์๋ค์ด โ์ง์ง ์ฐ์(earthquake sequence)โ์ด๋ผ๊ณ ๋ถ๋ฅด๋ ํ์์ด ๋ฐ์ํ๋ ๊ฒ์ผ๋ก ๋ณด์ธ๋ค. ์ด๋ ํ๋์ ๋จ์ธต์ ์์ ์๋ ฅ์ด ์์ ํ ํด์๋์ง ์์ ์์ญ ๋ ์ ๊ฑธ์ณ ์ฐ์์ ์ผ๋ก ์ง์ง์ด ๋ฐ์ํ๋ ํ์์ ์๋ฏธํ๋ค. ํธ๋ก์ด(Troy) ์ ์ ์ ํธ๋ก์ด 6์ธต์ด ์ธ๊ฐ์ด ์๋ ์ง์ง์ผ๋ก ํ๊ดด๋ ๊ฒ์ด ๋ํ์ ์ธ ์๋ค. ํ๋ ํ๋ฅดํค์์ ์ง์ง ๋ฐ์ ์ง๋๋ฅผ ํ๊ธฐ ์ฒญ๋๊ธฐ ์๋ ํ๊ดด๋ ๋์ ์ง๋์ ๊ฒน์ณ๋ณด๋ฉด ๊ฑฐ์ ์ผ๋์ผ๋ก ์ผ์นํ๋ ์๊ด๊ด๊ณ๋ฅผ ๋ณด์ธ๋ค.
-
์ง๋ณ์ ๊ทธ๋ฆผ์: ๊ณ ๋ ์ธ๊ณ์ ํฌ๋ฐ๋ฏน(The Shadow of Disease: Pandemics in the Ancient World): ๊ฐ๋ญ, ๊ธฐ๊ทผ, ์ ์๊ณผ ํจ๊ป ์ง๋ณ์ ๋ฌธ๋ช ๋ถ๊ดด์ โ๋ค ๋ฒ์งธ ๊ธฐ์ฌโ์๋ค. ํธ๋ฉ๋ก์ค์ ใ์ผ๋ฆฌ์์ค(The Iliad)ใ์ ๋ฌ์ฌ๋ ๋ฏธ์ผ๋ค ๊ตฐ๋์ ์ญ๋ณ, ํ๋ธ๋ฆฌ ์ฑ๊ฒฝ ใ์ถ์ ๊ตฝ๊ธฐ(Exodus)ใ์ 10๊ฐ์ง ์ฌ์ ๋ฑ์ ๋น์ ์ ์ผ๋ณ์ ๋ํ ํฌ๋ฏธํ ๊ธฐ์ต์ผ ์ ์๋ค. ์ค์ ๋ก ๊ธฐ์์ 1350๋ ๊ฒฝ ํํ์ดํธ ์๊ตญ์์๋ ์ด์งํธ ์ ์ ํฌ๋ก๋ค์ด ์ฎ๊ฒจ์จ ์ญ๋ณ์ผ๋ก ์ํผ๋ฃฐ๋ฆฌ์ฐ๋ง 1์ธ(Sufi Lulu-Luma I)๋ฅผ ๋น๋กฏํ ์์กฑ๊ณผ ๋ฐฑ์ฑ๋ค์ด ํฌ๊ฒ ํฌ์๋์๋ค.
๋์ฑ ๊ฒฐ์ ์ ์ธ ์ฆ๊ฑฐ๋ ๊ธฐ์์ 1140๋ ๊ฒฝ ํ๋ผ์ค ๋์ธ์ค 5์ธ(Ramses V)์ ๋ฏธ๋ผ์์ ๋ฐ๊ฒฌ๋ ์ฒ์ฐ๋(smallpox) ํ์ ์ด๋ค. ํฌ๋ฆฐ ํํผ๋ฃจ์ค(Turin papyrus)๋ ๋์ธ์ค 5์ธ์ ๊ทธ์ ๊ฐ์กฑ์ด ์ญ๋ณ์ผ๋ก ์ฌ๋งํ์ผ๋ฉฐ, ์ด๋ค์ ๋งค์ฅํ๋ ๋ฐ 16๊ฐ์์ด ๊ฑธ๋ ธ๊ณ , ์ฌ์ง์ด ์๊ฐ์ ๊ณ๊ณก(Valley of the Kings)์ ๋ด์ํ๋ ๋ฑ โ์ธ๊ณ ์ต์ด์ ๊ฒฉ๋ฆฌ(quarantine)โ ์กฐ์น๊ฐ ์ทจํด์ก์์ ๊ธฐ๋กํ๊ณ ์๋ค. ์ด์ฒ๋ผ ์ง๋ณ์ ๋ถ๊ดด ์ง์ ๊ณผ ์งํ์ ์ฌํ๋ฅผ ๋์ฑ ์ทจ์ฝํ๊ฒ ๋ง๋ค์์ ๊ฒ์ด๋ค.
์์คํ ๋ถ๊ดด์ ๊ทธ ์ดํ: ์ํ๊ธฐ์ ์ฌํ์
์ด ๋ชจ๋ ๋ณตํฉ์ ์ธ ์์ธ๋ค์ด ํ๊บผ๋ฒ์ ๋ฎ์น๋ฉด์, ๊ณ ๋ ์ฌํ๋ โ์์คํ ๋ถ๊ดด(systems collapse)โ๋ฅผ ๊ฒฝํํ๋ค. ์ผ์๋ธ๋ฆฌ์ง ๋ํ์ ์ฝ๋ฆฐ ๋ ํ๋ฃจ(Colin Renfrew) ๊ต์๊ฐ ์ ์ํ ์ด ์ฉ์ด๋ ์ค์ ๊ฒฝ์ , ์์ ์๋ฆฌํธ, ์ค์ ์ ๋ถ์ ๋ถ๊ดด์ ํจ๊ป ์ธ๊ตฌ ๊ฐ์(์ฌ๋ง ๋ฐ ์ด์ฃผ)๊ฐ ๋ค๋ฐ๋ฅด๋ ๋ณตํฉ ์์คํ ์ ์ํด๋ฅผ ์ค๋ช ํ๋ค. ํ๊ธฐ ์ฒญ๋๊ธฐ ์๋์ ๋ถ๊ดด๋ ์ด๋ฌํ ์์คํ ๋ถ๊ดด์ ์ ํ์ ์ธ ์์๋ค.
์์คํ ๋ถ๊ดด๋ ์ฆ๊ฐ์ ์ธ โ์ํ๊ธฐ(Dark Age)โ๋ก ์ด์ด์ก๋ค. ๋ฏธ์ผ๋ค ๊ทธ๋ฆฌ์ค์์ ์ฌ์ฉ๋๋ ์ ํ B ๋ฌธ์(Linear B)๋ ์ฌ๋ผ์ก๊ณ , ๋๊ท๋ชจ ๊ฑด์ถ ๊ธฐ์ ๋ ์ํ๋ค. ์ฌ๋๋ค์ ๊ธ๊ฒฉํ ๋ฎ์ ์์ค์ ์ฌํ-์ ์น-๊ฒฝ์ ์ ๊ธฐ๋ฅ์ผ๋ก ํ๊ทํด์ผ ํ๋ค. ํ์ง๋ง ์ด ์ํ๊ธฐ๊ฐ ๋จ์ํ ํด๋ณด์ ์๊ธฐ๋ง์ ์๋์๋ค. โํ์๋ ๋ฐ๋ช ์ ์ด๋จธ๋๋ค(necessity is the mother of invention)โ๋ผ๋ ๋ง์ฒ๋ผ, ์ฃผ์ ๊ตฌํ๊ธฐ๊ฐ ์ด๋ ค์์ง์ ์ฌ๋๋ค์ ์๋ก์ด ๊ธ์์ธ ์ฒ (iron)์ ํ์ฉํ๊ธฐ ์์ํ๋ค. ํคํ๋ก์ค์ธ๋ค์ ์ฒ ์ ๊ฐ์ฅ ๋จผ์ ์ฌ์ฉํ์ผ๋ฉฐ, ์ด๋ ์ฒญ๋์ ๋์ฒดํ๋ ์ค์ํ ์์์ด ๋์๋ค.
๋ํ, ๊ฐ๋์ ์ค๋ถ(ํ ๋ ๋ฐ๋ ผ) ์ถ์ ์ ํ๋ํค์์ธ(Phoenicians)๋ค์ ๋ถ๊ดด ์ดํ ๊ฐ์ฅ ์ฑ๊ณต์ ์ผ๋ก ์ ์ํ๊ณ ๋ฒ์ฑํ ๋ฌธ๋ช ์ค ํ๋์๋ค. ์ด๋ค์ ๊ธฐ์กด์ ์๋ ์ํ๋ฒณ์ ํ์คํํ์ฌ ์ง์คํด ์ ์ญ์ผ๋ก ํผ๋จ๋ ธ๊ณ , ์ด๋ ๊ทธ๋ฆฌ์ค ์ํ๋ฒณ๊ณผ ๋ผํด ์ํ๋ฒณ์ ๊ธฐ์์ด ๋์ด ์ค๋๋ ๊น์ง ์ด์ด์ง๋ ๋ฌธ์ ์ฒด๊ณ์ ํ ๋๊ฐ ๋์๋ค. ์ฆ, ์ํ๊ธฐ๋ ํ๊ดด์ ํจ๊ป ํ์ ์ ์จ์์ ํ๊ณ ์์๋ ๊ฒ์ด๋ค.
ํ๋์ ๋์ง๋ ์ง๋ฌธ: ์ํธ ์ฐ๊ฒฐ์ฑ์ ์๋ ์ ๊ฒ
ํด๋ผ์ธ ๊ต์๋ ํ๊ธฐ ์ฒญ๋๊ธฐ ์๋์ ๋ถ๊ดด๊ฐ ์ค๋๋ ์ฐ๋ฆฌ์๊ฒ ์ฃผ๋ ๊ฐ์ฅ ํฐ ๊ตํ์ โ์ํธ ์ฐ๊ฒฐ์ฑ(interconnectedness)โ์ ์๋ฉด์ฑ์ด๋ผ๊ณ ๊ฐ์กฐํ๋ค. ๋น์์ โ์๊ท๋ชจ ์ธ๊ณ ๋คํธ์ํฌโ๋ ๋ฒ์์ ๊ฐ์ ธ์์ง๋ง, ๋์์ ํ ์ง์ญ์ ์ฌ์์ด ์ ์ธ๊ณ์ ์ธ ๋๋ฏธ๋ ธ ํจ๊ณผ๋ฅผ ์ผ์ผํค๋ ์ทจ์ฝ์ ์ด ๋์๋ค.
๋ฏธ ์ก๊ตฐ ๊ณต๋ณ๋(US Army Corps of Engineers)์์ ์ปดํจํฐ ์๋ฎฌ๋ ์ด์ ๊ฒฐ๊ณผ๋ ํฅ๋ฏธ๋ก์ด ๊ฒฐ๋ก ์ ๋์ถํ๋ค. ํํ์ดํธ์ ์ฐ๊ฐ๋ฆฌํธ๊ฐ ๋์์ ๋ถ๊ดดํ์ ๋ ์ ์ฒด ๋คํธ์ํฌ๊ฐ ์ํด๋ ๊ฐ๋ฅ์ฑ์ด ๊ฐ์ฅ ๋์๋ค๋ ๊ฒ์ด๋ค. ๋ฐ๋ฉด, ์์๋ฆฌ์๋ ๋ฐ๋น๋ก๋์ ๊ฐ์ ๋ค๋ฅธ ๋ฌธ๋ช ๋ค์ ๋ถ๊ดด๋ ์ ์ฒด ๋คํธ์ํฌ์ ์น๋ช ์ ์ด์ง ์์๋ค. ์ด์งํธ๋ ๋น๋ก ํฐ ์ด๋ ค์์ ๊ฒช์์ง๋ง, ๋ถ๊ดด์ ์ ์ํ๊ณ ์ด์๋จ์๋ค. ์ด๋ ๋คํธ์ํฌ ๋ถ๊ดด๊ฐ ์ฝ์ง ์์ผ๋ฉฐ, ์ฌ์์ ์ด๋ป๊ฒ ๋์ํ๋๋์ ๋ฐ๋ผ ํ๋ณต ํ๋ ฅ์ฑ(resilience)์ด ๋ฌ๋ผ์ง ์ ์์์ ์์ฌํ๋ค.
ํด๋ผ์ธ ๊ต์๋ ์ค๋๋ ์ฐ๋ฆฌ๊ฐ ๊ธฐํ ๋ณํ, ํฌ๋ฐ๋ฏน, ๊ฒฝ์ ์๊ธฐ, ์ง์ ํ์ ๊ฐ๋ฑ ๋ฑ ๋ณตํฉ์ ์ธ ์๊ธฐ์ ์ง๋ฉดํด ์์์ ์๊ธฐ์ํจ๋ค. ํ๊ธฐ ์ฒญ๋๊ธฐ ์๋์ ๋ฌธ๋ช ๋ค์ ์์ ๋ค์ด ๋ถ๊ดดํ๊ณ ์๋ค๋ ์ฌ์ค์กฐ์ฐจ ์ ๋๋ก ์ธ์งํ์ง ๋ชปํ์ ์ ์๋ค. ๊ณผ์ฐ ์ฐ๋ฆฌ๋ 3์ฒ ๋ ์ ์ ์ ์กฐ๋ค๋ณด๋ค ํ๋ช ํ๊ฒ ๋ค๊ฐ์ค๋ โํผํํธ ์คํฐโ์ ๋์ฒํ ์ ์์๊น? ์ญ์ฌ๋ ๋ฐ๋ณต๋์ง ์์ง๋ง, ๊ทธ ์ด์จ ์์์ ์ฐ๋ฆฌ๋ ๊ณผ๊ฑฐ์ ๊ตํ์ ํตํด ๋ฏธ๋๋ฅผ ์ค๋นํ ์งํ๋ฅผ ์ฐพ์์ผ ํ ๊ฒ์ด๋ค.
โAre Human Drivers Finally Obsolete? | Freakonomics Radioโ โ Freakonomics Radio Network ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
์ธ๊ฐ ์ด์ ์๋ ์ ๋ง ๊ตฌ์๋์ ์ ๋ฌผ์ด ๋ ๊น? ์์จ์ฃผํ์ฐจ์ ๊ฟ๊ณผ ํ์ค, ๊ทธ๋ฆฌ๊ณ ๊ตฌ๊ธ์ ๋น๋ฐ ํ๋ก์ ํธ
์ด์ ๋์์ ์์ ๋ผ๋ ๋ ์ด ์ฌ๊น? SF ์ํ์์๋ ๋ณด๋ ์์จ์ฃผํ์ฐจ๊ฐ ์ฐ๋ฆฌ์ ์ผ์์ ๊น์์ด ํ๊ณ ๋๋ ์๋๊ฐ ๋๋ํ๊ณ ์์ต๋๋ค. ํ์บ์คํธ โ์์น ์์ง(Search Engine)โ์ ์งํ์ PJ ๋ณดํธ(PJ Vote)๋ ์ต๊ทผ โํ๋ฆฌ์ฝ๋ ธ๋ฏน์ค ๋ผ๋์ค(Freakonomics Radio)โ์์ ํน๋ณ ํ์ ์ ํตํด ์์จ์ฃผํ์ฐจ์ ๊ณผ๊ฑฐ, ํ์ฌ, ๊ทธ๋ฆฌ๊ณ ๋ฏธ๋๋ฅผ ์ฌ์ธต์ ์ผ๋ก ํ๊ตฌํ๋ 2๋ถ์ ์๋ฆฌ์ฆ๋ฅผ ์ ๋ณด์์ต๋๋ค. ์ด ๊ธฐ์ฌ๋ ๊ทธ ์ฒซ ๋ฒ์งธ ์ด์ผ๊ธฐ์์ PJ๊ฐ ๋ค๋ ค์ค ๋๋ผ์ด ์ฌ์ ์ ๋ฐ๋ผ๊ฐ๋ฉฐ, ์ธ๊ฐ ์ด์ ์๊ฐ โ๊ตฌ์๋์ ์ ๋ฌผโ์ด ๋ ์ ์๋ค๋ ๋๋ดํ ์์ธก์ ๊ทผ๊ฑฐ์ ์์จ์ฃผํ ๊ธฐ์ ๋ฐ์ ์ ์จ๊ฒจ์ง ์ญ์ฌ๋ฅผ ํํค์ณ ๋ด ๋๋ค.
์ด์ ๋์์ ์์ ๋ผ๊ณ ์ถ์๋ ์ด์ : ๋ถ์์์ ์์๋ ์์จ์ฃผํ ๊ฒฝํ
PJ ๋ณดํธ๊ฐ ์์จ์ฃผํ์ฐจ์ ๊ด์ฌ์ ๊ฐ๊ฒ ๋ ๊ณ๊ธฐ๋ ๋ค์ ์์ธ์ ๊ฐ์ธ์ ์ธ ๊ฒฝํ์์ ์์๋์์ต๋๋ค. ๋ฒค์นํ๋ ์ค์ ๋๋ฌด ๋ชฐ๋ํ๋ค ๋ถ์์ ์ ์ด ์ํ๋ถ ํ์ฅ(hernia) ์์ ์ ๋ฐ๊ฒ ๋ ๊ทธ๋ ๊ทน์ฌํ ํต์ฆ๊ณผ ์ ํ๋ ๊ฑฐ๋์ผ๋ก ์ด๋ ค์์ ๊ฒช์์ต๋๋ค. ์ํ๋์์ค์ฝ์ ์๋ ์น๊ตฌ๋ฅผ ๋ฐฉ๋ฌธํ์ ๋, ๊ทธ๋ โ์จ์ด๋ชจ(Waymo)โ ๋ก๋ณดํ์๋ฅผ ์ฒ์ ์ด์ฉํ๊ฒ ๋ฉ๋๋ค.
โ๋ฒํผ ํ๋๋ง ๋๋ฅด๋ฉด ์๋ฌด๋ ์ด์ ํ์ง ์๋ ์ฐจ๊ฐ ํ์ด๋์์. ์ฐจ์ ํ๋ฉด ํธ๋ค์ด ์ ์ ๋ก ๋์๊ฐ๋ ๊ฑธ ๋ณด๊ฒ ๋์ฃ . ๋ง์น ๋นํ๊ธฐ๋ฅผ ์ฒ์ ํ์ ๋์ ๊ฒฝ์ด๋ก์ ๊ฐ์์ด์. ์ธ ๋ฒ์งธ ํ ๋๋ ๊ทธ๋ฅ ์๋ฆฌ๋ฒ ์ดํฐ ํ๋ ๋๋์ด ๋ค์๊ณ ์.โ
์ด ๊ฒฝํ์ PJ์๊ฒ ๊น์ ์ธ์์ ๋จ๊ฒผ์ต๋๋ค. โ๋ง์ ๊ฒ์ด ๋ณํ ๊ฒโ์ด๋ผ๋ ์๊ฐ๊ณผ ํจ๊ป, ์ ์ฌ๋๋ค์ด ์ด ๋ณํ์ ๋ํด ๋ ๋ง์ด ์ด์ผ๊ธฐํ์ง ์๋์ง ์์ํจ์ ๋๊ผ์ต๋๋ค. ๊ทธ์ 2๋ถ์ ์๋ฆฌ์ฆ๋ ์ด ์ง๋ฌธ์ ๋ํ ๋ต์ ์ฐพ์๊ฐ๋ ์ฌ์ ์ ๋๋ค. ์ฒซ ๋ฒ์งธ ์ํผ์๋๋ โ์ฐจโ ์์ฒด์ ์ด์ ์ ๋ง์ถ๊ณ , ๋ ๋ฒ์งธ ์ํผ์๋๋ โ์ด์ ์โ์ ๋ํ ์ด์ผ๊ธฐ๋ฅผ ๋ค๋ฃน๋๋ค.
์ธ๊ฐ ์ด์ ์๋ ์ผ๋ง๋ โ์๋งโ์ผ๊น?
PJ ๋ณดํธ๋ ์ธ๊ฐ ์ด์ ์์ ํ๊ณ์ ๋ํด ์์งํ๊ฒ ์ด์ผ๊ธฐํฉ๋๋ค. โ์ธ๊ฐ์ด ์ด์ ํ๋ ๊ฑธ ๋ณธ ์ ์ด ์๋์? ์ฐ๋ฆฌ ์ค์ค๋ก๋ฅผ ํฌํจํด์์. ์ฐ๋ฆฌ๋ ๊ทธ๋ ๊ฒ ์ํ์ง ๋ชปํฉ๋๋ค.โ ์คํฐ๋ธ ๋๋ธ๋(Steven Dubner) ๋ํ ์ด์ ๋์ํ๋ฉฐ ์์ ์ ์ด์ ์ค๋ ฅ์ ํ์์ด ์์์ ๊ณ ๋ฐฑํฉ๋๋ค. โ์ ๋ ์ด์ ์ค๋ ฅ์ ๋ํ ํ์์ด ์์ด์. ์ ๋ ์ฑ์ง์ด ๊ธํ๊ณ , ์ฐ๋งํ์ฃ .โ
์ค์ ๋ก ์ฐ๋ฆฌ ๋๋ถ๋ถ์๊ฒ ์ด์ ์ ์ผ์์ ์ผ๋ก ํ๋ ํ๋ ์ค ๊ฐ์ฅ ์ํํ ์ถ์ ์ํฉ๋๋ค. ์์จ์ฃผํ์ฐจ์ ํต์ฌ์ ์ธ ํ๋ณด ๋ฌธ๊ตฌ๋ ๋ฐ๋ก โ์์ โ์ ๋๋ค. ์ปดํจํฐ๊ฐ ์ด์ ํ๋ ์ฐจ๋ ์์ฃผ์ด์ , ์กธ์์ด์ , ์ฃผ์ ์ฐ๋ง, ๋ฌธ์ ๋ฉ์์ง ์ ์ก, ๋ณด๋ณต ์ด์ ์ ํ์ง ์์ต๋๋ค. ๊ทธ๋ฆฌ๊ณ ์ด๋ฌํ ์์จ์ฃผํ์ฐจ๋ ๋ ์ด์ ๋ฏธ๋์ ์ด์ผ๊ธฐ๊ฐ ์๋๋๋ค. ์จ์ด๋ชจ ๊ฐ์ ๋ก๋ณดํ์๋ ์ด๋ฏธ ๋ฏธ๊ตญ 10๊ฐ ๋์์์ ์๋ฐฑ๋ง ๊ฑด์ ์ดํ์ ์ ๊ณตํ๊ณ ์์ผ๋ฉฐ, ์ค๊ตญ์์๋ ๋ ๊ด๋ฒ์ํ๊ฒ ํ์ฐ๋๊ณ ์์ต๋๋ค. ์ํ๋์์ค์ฝ๋ ์ค์คํด ๊ฐ์ ๋์์์๋ ์์จ์ฃผํ์ฐจ๊ฐ ์ฐ๋ฒ(Uber)๋งํผ์ด๋ ํํ ํ๊ฒฝ์ด ๋๊ณ ์์ต๋๋ค.
์ฌ๋ผ์ง ์ง์ , ๊ทธ๋ฆฌ๊ณ ๋ณํ์ ์๊ณ
์์จ์ฃผํ์ฐจ์ ๋ฑ์ฅ์ ์ธ๋ฅ ์ญ์ฌ์์ ๋ฐ๋ณต๋์ด ์จ ๊ฑฐ๋ํ ๋ณํ์ ํ๋ฆ๊ณผ ๋ง๋ฟ์ ์์ต๋๋ค. PJ๋ 200๋ ์ ์ผ๋ก ๊ฑฐ์ฌ๋ฌ ์ฌ๋ผ๊ฐ ์ธ ๊ฐ์ง ์ง์ ์ ์์ํ๊ฒ ํฉ๋๋ค.
- ๋ ธ์ปค์ ํผ(Knocker-upper): ์๋ ์๊ณ๊ฐ ๋ฐ๋ช ๋๊ธฐ ์ , ๊ธด ๋ง๋๊ธฐ๋ก ์ฌ๋๋ค์ ์ฐฝ๋ฌธ์ ๋๋๋ ค ๊นจ์์ฃผ๋ ์ง์ . ์ด์ ๋ ์ค๋งํธํฐ ์๋์ด ๊ทธ ์๋ฆฌ๋ฅผ ๋์ ํฉ๋๋ค.
- ๋จํ ๋ผ์ดํฐ(Lamp Lighter): ๋ฐค์ ๊ฐ์ค๋ฑ์ ์ผ๊ณ ์๋ฒฝ์ ๋๋ ์ง์ . ์ด์ ๋ ์ ๊ธฐ ๊ฐ๋ก๋ฑ์ด ์๋์ผ๋ก ๋ฐํ์ง๋๋ค.
- ๋ง๋ถ(Coachman): ๋ง์ ํ๊ณ ์น๊ฐ์ ์ํ๋ ๊ณณ์ผ๋ก ๋ฐ๋ ค๋ค์ฃผ๋ ์ง์ . ์ด์ ๋ ์๋์ฐจ ์ด์ ์๊ฐ ๊ทธ ์ญํ ์ ํฉ๋๋ค.
์ด ์ค ์ฒซ ๋ ์ง์ ์ ์ญ์ฌ ์์ผ๋ก ์ฌ๋ผ์ง ์ง ์ค๋์ ๋๋ค. ํ์ง๋ง โ์ด์ ์โ๋ ์ฌ์ ํ ์กด์ฌํฉ๋๋ค. ๊ทธ๋ฌ๋ PJ๋ ์ด ์ง์ ๋ํ ๋ณํ์ ๊ธฐ๋ก์ ์ ์๋ค๊ณ ๋งํฉ๋๋ค. โ์ด์ ์๋ผ๋ ๋จ์ด๊ฐ ์ง๊ธ์ ์ธ๊ฐ์ ๋ ์ฌ๋ฆฌ๊ฒ ํ์ง๋ง, ๊ณง ์๊ธฐ์ธ์ฒ๊ธฐ(dishwasher), ํ๋ฆฐํฐ(printer), ์ปดํจํฐ(computer)์ฒ๋ผ ๊ธฐ๊ณ๋ฅผ ์ง์นญํ๋ ๋จ์ด๋ก ๋ฐ๋ ์ ์์ต๋๋ค.โ
๊ฟ์ ์ค๋๋์์ง๋ง, ํ์ค์ ๋ฌ๋๋ค: ์๋์ฐจ์ ๋ฑ์ฅ๊ณผ ์ด๊ธฐ ์ ํญ
ํฅ๋ฏธ๋กญ๊ฒ๋, ์ธ๊ฐ ์ด์ ์๋ฅผ ๊ธฐ๊ณ๋ก ๋์ฒดํ๋ ค๋ ๊ฟ์ ๋จํ ๋ผ์ดํฐ์ ์๋๋งํผ์ด๋ ์ค๋๋์์ต๋๋ค. ๋ง์ฐจ์์ ์๋์ฐจ๋ก ๋์ด์ค๋ฉด์ ์ธ๋ฅ๋ ํ ๊ฐ์ง ์ค์ํ ๊ฒ์ ์์์ต๋๋ค. ๋ฐ๋ก โ๊ฐ๊ฐ(sentience)โ์ ๋๋ค. ๋ง์ ๊ณ ์๋ฅผ ๋์๋ ์ ๋ฒฝ์ผ๋ก ๋ฌ๋ ค๊ฐ์ง ์์ต๋๋ค. ํ์ง๋ง ์๋์ฐจ๋ ๊ทธ๋ ์ง ์์ต๋๋ค.
1800๋ ๋ ํ๋ฐ, ์ฆ๊ธฐ๊ธฐ๊ด์ฐจ์ ๊ฐ์๋ฆฐ ์๋์ฐจ๊ฐ ๋๋ก๋ฅผ ์ง์ฃผํ๊ธฐ ์์ํ์ ๋, ์ฌ๋๋ค์ ์ด ์๋ก์ด ๊ธฐ์ ์ ๋ํด ๊ฒฉ๋ ฌํ ๋ ผ์์ ๋ฒ์์ต๋๋ค. ๋๋ ค์์ ํฌ๊ฒ ๋ ๊ฐ์ง์์ต๋๋ค.
- ์ผ์๋ฆฌ ์ํ: ๋ง๋ถ, ๋ง๊ตฌ๊ฐ ์ฃผ์ธ, ๋ง๊ตฌ ์ ์กฐ์ฌ, ๋ง๋ถ ์ฒ๋ฆฌ์ ์ ๋ฑ ๋ง๊ณผ ๊ด๋ จ๋ ์๋ง์ ๋ ธ๋์๋ค์ ์ผ์๋ฆฌ๊ฐ ์ฌ๋ผ์ง ๊ฒ์ด๋ผ๋ ์ฐ๋ ค๊ฐ ์ปธ์ต๋๋ค. ์ค๋๋ โํ์คํฐ์ฆ(Teamsters)โ๋ผ๋ ๋จ์ด๊ฐ ํธ๋ญ ์ด์ ์ฌ ๋ ธ์กฐ๋ฅผ ์ฐ์์ํค์ง๋ง, ์๋๋ ๋ง์ ๋ชจ๋ ์ฌ๋๋ค์ ์ง์นญํ์ต๋๋ค.
- ์์ ๋ฌธ์ : ๊ท์ ์์ด ๋์ ๋ ์๋์ฐจ๋ ์์ฒญ๋ ์ฌ๋ง๋ฅ ์ ๊ธฐ๋กํ์ต๋๋ค. ์ด์ ๋ฉดํ, ์ ํธ๋ฑ, ๋ฐฉํฅ ์ง์๋ฑ์ด ์๋ 1900๋ ๋ ์ด ๋ํธ๋ก์ดํธ์์๋ ์๋ง์ ์ฌ๋ง์๊ฐ ๋ฐ์ํ๊ณ , ๊ทธ ์ค ์๋น์๋ ์์ด๋ค์ด์์ต๋๋ค.
์ผ๋ถ ๋ฐ(ๅ)์๋์ฐจ ์ด๋๊ฐ๋ค์ โ์ ๊ธฐ ์กฐ๋ก(Red Flag Laws)โ์ ๊ฐ์ ๊ท์ ๋ฅผ ํตํด ๊ธฐ์ ์ ๋์ ์ ๋ง๊ฑฐ๋ ๋ฆ์ถ๋ ค ํ์ต๋๋ค. ์๋ฅผ ๋ค์ด, ํ์ค๋ฒ ์ด๋์์์๋ ๊ฐ์ถ์ ๋ง๋๋ ๋ฌด๋ง์ฐจ(horseless carriage) ์ด์ ์๋ ์ฐจ๋ฅผ ๋ถํดํ์ฌ ๋ค๋ถ ๋ค์ ์จ๊ฒจ์ผ ํ๋ค๋ ๋ฒ์์ด ์ ์๋๊ธฐ๋ ํ์ต๋๋ค(์ฃผ์ง์ฌ์ ๊ฑฐ๋ถ๊ถ ํ์ฌ๋ก ๋ฌด์ฐ). ์ด๋ฌํ ๊ท์ ๋ ํฉ๋นํ๊ฒ ๋ค๋ฆด์ง ๋ชจ๋ฅด์ง๋ง, ๊ทธ๋ค์ ์์ธก์ ์ด๋ ์ ๋ ์ณ์์ต๋๋ค. ์๋์ฐจ๋ ์ด๊ธฐ์๋ ๋ง์ ์ผ์๋ฆฌ๋ฅผ ์์ด๊ณ , ๋งค์ฐ ์ํํ์ต๋๋ค.
์ฌํ๊ฐ ์๋์ฐจ์ ์ ์ํ๋ ๋ฐ๋ ์์ญ ๋ ์ด ๊ฑธ๋ ธ์ต๋๋ค. ๋ฒ๊ท, ์ด์ ๋ฉดํ, ์ด์ ๊ต์ก, ๋๋ก ์ค๊ณ, ๊ณ ์๋๋ก, ์์ ๋ฒจํธ, ์์ด๋ฐฑ ๋ฑ์ด ๋ฐ๋ช ๋๋ฉด์ ์ด์ ์ ๋ ์ํํด์ก์ต๋๋ค. ํ์ง๋ง ์ค๋งํธํฐ์ ๋ฑ์ฅ์ ์ด๋ฌํ ๋ฐ์ ์ ์ผ๋ถ ๋๋๋ ค ๋๊ธฐ๋ ํ์ต๋๋ค. ์ค๋๋ ๋ฏธ๊ตญ์์ ์๋์ฐจ ์ฌ๊ณ ์ฌ๋ง์ ์๋ ์ด๊ธฐ๋ ์คํผ์ค์ด๋(opioids) ์ค๋ ์ผ๋ก ์ธํ ์ฌ๋ง์ ์์ ๋น์ทํ๋ฉฐ, ์ฝ 100๋ช ์ค 1๋ช ๊ผด๋ก ๋ฐ์ํฉ๋๋ค.
์ฌ๋ง์์ ํผ์ด๋ ์์จ์ฃผํ์ ๊ฟ: DARPA ๊ทธ๋๋ ์ฑ๋ฆฐ์ง
์๋์ฐจ์ ์์ ๋ฌธ์ ๋ฅผ ๊ทผ๋ณธ์ ์ผ๋ก ํด๊ฒฐํ๊ณ , ์๋์ฐจ๋ฅผ ๋ค์ โ๊ฐ๊ฐ์ โ์ผ๋ก ๋ง๋ค๋ ค๋ ๊ฟ์ ์ค๋ซ๋์ ์ด์ด์ ธ ์์ต๋๋ค. ์ด๊ธฐ์๋ ๋ฌด์ ์กฐ์ข ์ฐจ๋์ด๋ ๋๋ก ๋ฐ์ ์์์ ์ค์นํ๋ ๋ฑ ๋ค์ํ ์์ด๋์ด๊ฐ ์๋๋์์ง๋ง, ๊ธฐ์ ์ ํ๊ณ์ ๋ถ๋ชํ์ต๋๋ค.
2002๋ , ๋ฏธ๊ตญ ๊ตญ๋ฐฉ๋ถ ์ฐํ์ ๋น๋ฐ ์ฐ๊ตฌ๊ธฐ๊ด์ธ DARPA(Defense Advanced Research Projects Agency, ๊ณ ๋ฑ๋ฐฉ์์ฐ๊ตฌ๊ณํ๊ตญ)๊ฐ ์ด ๊ฟ์ ๋ถ์ ์งํ์ต๋๋ค. GPS, M16 ์์ด, ์ด๊ธฐ ์ธํฐ๋ท, ํ๋ ๋ฐํฐ ๋๋ก (Predator drone) ๋ฑ์ ๊ฐ๋ฐํ DARPA๋ ์์จ ๊ตฐ์ฉ ์ฐจ๋ ๊ฐ๋ฐ์ ๊ด์ฌ์ ๊ฐ์ก์ต๋๋ค. ๋น์ DARPA ๊ตญ์ฅ์ด์๋ ํ ๋ ํ ๋(Tony Tether)๋ โ์ฝํ ์คํธ๋ฅผ ์ด์ด ์ด ๋ชจ๋ ๊ธฐ์ ๋ค์ ์ ๋๋ก ๋ ์์จ์ฃผํ์ฐจ๋ก ๋ง๋ค ์ ์๋์ง ๋ณด์โ๊ณ ์ ์ํ์ต๋๋ค.
๊ทธ์ ์๋ ์์ด๋์ด๋ ๋ผ์ค๋ฒ ์ด๊ฑฐ์ค ์คํธ๋ฆฝ์ ๋ฌ๋ฆฌ๋ ๊ฒ์ด์์ง๋ง, ๋์ ์ ์ฒด๋ฅผ ๋ง๋น์ํฌ ์๋ ์์๊ธฐ์ ๊ฒฐ๊ตญ ๋ผ์ค๋ฒ ์ด๊ฑฐ์ค ์ธ๊ณฝ์ ์ฌ๋ง์์ ๊ฒฝ์ฃผ๋ฅผ ํผ์น๊ธฐ๋ก ํฉ๋๋ค. DARPA๋ โ๋ฏธ๊ตญ ๋ณ์ฌโ๋ผ๋ ์ด์ ์๋ฅผ ๋์ฒดํ๊ณ ์ ํ์ต๋๋ค. ์ง๋ขฐ๊ฐ ๊น๋ ค ์์์ง ๋ชจ๋ฅด๋ ๋๋ก๋ฅผ ์ค์ค๋ก ์ด์ ํ ์ ์๋ ์ฐจ๋์ด ํ์ํ๋ ๊ฒ์ ๋๋ค. ์๊ธ์ 100๋ง ๋ฌ๋ฌ. ๊ท์น์ ๋งค์ฐ ์์ ๋ก์ ์ต๋๋ค. ์ฐจ๋ ๊ฐ ํต์ ๋ง ๊ธ์ง๋์์ ๋ฟ, ๋ฐํด ์๋ ํํ๋ ์ ํ์ด ์์์ต๋๋ค. ์ฌ์ง์ด โ๋ค๋ฅธ ์ฐจ๋์ ๊ณต๊ฒฉํ ์ ์์ต๋๊น?โ๋ผ๋ ์ง๋ฌธ์ โ์ ๋๋คโ๋ ๋ต์ด ๋์์์ ์ ๋๋ก ์ฐธ๊ฐ์๋ค์ ์์๋ ฅ์ ๋์ณค์ต๋๋ค. ๋ง์ ์ฐธ๊ฐํ์ โ๋ฐฐํ๋ด(Battlebots)โ ๋ํ ์ถ์ ๋ค์ด์์ต๋๋ค.
2004๋ ๊ทธ๋๋ ์ฑ๋ฆฐ์ง: ์ฒ์ฐธํ ์คํจ
2004๋ 3์์ ์ฒซ ๋ฒ์งธ ๊ทธ๋๋ ์ฑ๋ฆฐ์ง๋ โ์์ ํ ํ์คํ ๋ฆฌ์ ์ฌ์(utter hysterical disaster)โ์ด์์ต๋๋ค.
- ์ค์๋ ๋ฅด์๋์คํค(Anthony Lewendowski)์ โ๊ณ ์คํธ ๋ผ์ด๋(Ghost Rider)โ: ์ ์ผํ ์์จ์ฃผํ ์คํ ๋ฐ์ด์์ง๋ง, ์์ ํ ์์คํ ์ค์์น๋ฅผ ์ผ๋ ๊ฒ์ ์์ด๋ฒ๋ ค ์ถ๋ฐ๊ณผ ๋์์ ์ ๋ณต๋์์ต๋๋ค.
- ์นด๋ค๊ธฐ๋ฉ๋ก (Carnegie Mellon) ํ์ โ์๋์คํฐ(Sandstorm)โ: ๋ฐ์ ๋นจ๊ฐ์ ํ๋น(Humvee)์ ๋ฏธ๋ ์งํฅ์ ์ธ ์ผ์๋ค์ด ์ฅ์ฐฉ๋ ์ฐจ๋์ด์์ง๋ง, ์ถ๋ฐ ์งํ ํ๋๋ฏธ์ ๊ฑธ๋ ค ๋ฐํด๊ฐ ํ๋๋ค๊ฐ ํ์ด์ด๊ฐ ๋ น์ ๊ฒ์ ์ฐ๊ธฐ๋ฅผ ๋ฟ์ผ๋ฉฐ ๋ฉ์ถฐ ์ฐ์ต๋๋ค.
๋๋ถ๋ถ์ ์ฐจ๋์ด ์ฒ์ฐธํ๊ฒ ์คํจํ์ต๋๋ค. ์ฐธ๊ฐ์๋ค์ ๋ก๋ด์ด 26๋ง์ผ ๋ง๋ผํค์์ ๊ฒจ์ฐ 2๋ง์ผ์ ๋ฌ๋ฆฐ ๊ฒ์ ๋น์ ํ๋ฉฐ ์ค๋ง๊ฐ์ ๊ฐ์ถ์ง ๋ชปํ์ต๋๋ค. ๊ทธ๋ฌ๋ ์ด ๋ํ๋ ์์จ์ฃผํ ๊ธฐ์ ๊ฐ๋ฐ์ ์ํ โ์ฅโ์ ๋ง๋ จํ๋ค๋ ์ ์์ ํฐ ์๋ฏธ๊ฐ ์์์ต๋๋ค. ๋น์ ์ด ๊ด๊ฒฝ์ ์ง์ผ๋ณด๋ ํ ์ธ๋ฌผ์ด ์์๋๋ฐ, ๋ฐ๋ก ์คํ ํฌ๋ ๋ํ๊ต์ ์ ์ค์ ์ธ ๋ก๋ด ๊ณตํ์ **์ธ๋ฐ์ค์ฐฌ ์ค๋ฐ(Sebastian Thrun)**์ด์์ต๋๋ค.
์ค๋ฐ์ ํต์ฐฐ: ํ๋์จ์ด ์๋ ์ํํธ์จ์ด ๋ฌธ์
์ธ๋ฐ์ค์ฐฌ ์ค๋ฐ์ ์ฒซ ๋ฒ์งธ ์ฑ๋ฆฐ์ง๋ฅผ ๊ด์ ํ๋ฉฐ ์ฐธ๊ฐํ๋ค์ด ๊ทผ๋ณธ์ ์ธ ์ค๋ฅ๋ฅผ ๋ฒํ๋ค๊ณ ์๊ฐํ์ต๋๋ค.
โ๋ชจ๋ ํ์ด ์ด๊ฒ์ ํ๋์จ์ด ๋ฌธ์ ๋ก ๋ค๋ฃจ๋ ๊ฒ์ ๋ดค์ต๋๋ค. โ๋ ํฐ ๋ฐํด์ ๋ ํฐ ์ฐจ์ฒด๋ฅผ ๋ง๋ค์ด์ผ ํดโ๋ผ๊ณ ์๊ฐํ์ฃ . ํ์ง๋ง ์ ๋ โ์ ๊น๋ง, ์ด ์ฑ๋ฆฐ์ง๋ ์ฌ๋ง์ ์ฃผํํ ์ ์๋ ์์จ์ฃผํ์ฐจ๋ฅผ ๋ง๋๋ ๊ฑด๋ฐ, ์ฌ๋์ด ํ๋ฉด ์ ๋ฌ๋ฆฌ๋ ๋ ํฐ์นด๋ฅผ ๋น๋ฆด ์ ์์์? ๋ฌธ์ ๋ ์ฌ๋์ ์ด์ ์์์ ๋นผ๊ณ ์ปดํจํฐ๋ก ๋์ฒดํ๋ ๊ฑฐ์ผ. ์ด๊ฑด ๋ ํฐ ํ์ด์ด์ ๋ฌธ์ ๊ฐ ์๋๋ผ ์ํํธ์จ์ด ๋ฌธ์ ์ผโ๋ผ๊ณ ์๊ฐํ์ต๋๋ค.โ
๋ก๋ด ๊ณตํ๊ณผ ์ธ๊ณต์ง๋ฅ(AI)์ด๋ผ๋ ์ด์ค ๋ฐฐ๊ฒฝ์ ๊ฐ์ง ์ค๋ฐ์ ๋ก๋ด ์ด์ ์์ โ๋๋โ์ ์ง์คํ์ต๋๋ค. ๊ทธ๋ ๊ตฐ์ฉ ์ฐจ๋์ ๋์ด ์ ์ธ๊ณ ๊ตํต์ฌ๊ณ ์ฌ๋ง์ 100๋ง ๋ช ์ ๊ตฌ์ํ ์ ์๋ ๊ธฐ์ ์ ๊ฟ๊ฟจ์ต๋๋ค.
2005๋ ๊ทธ๋๋ ์ฑ๋ฆฐ์ง: ๊ธฐ๊ณ ํ์ต์ ์น๋ฆฌ
18๊ฐ์ ํ, DARPA๋ ์๊ธ์ 200๋ง ๋ฌ๋ฌ๋ก ๋ ๋ฐฐ ์ฌ๋ ธ๊ณ , ๋ ๋ฒ์งธ ๊ทธ๋๋ ์ฑ๋ฆฐ์ง๊ฐ ์ด๋ ธ์ต๋๋ค. ์นด๋ค๊ธฐ๋ฉ๋ก ํ์ ํฌ๋ฆฌ์ค ์์จ(Chris Urmson)์ ๋ ๋์ ์ฐจ๋์ผ๋ก ๋ค์ ์ฐธ๊ฐํ๊ณ , ์ค์๋ ๋ฅด์๋์คํค๋ ์ฌ์ ํ ์๋ํ์ง ์๋ ์คํ ๋ฐ์ด๋ก ์์ ์์ ํ๋ฝํ์ต๋๋ค. ๊ทธ๋ฆฌ๊ณ ์ค๋ฐ์ ์คํ ํฌ๋ ํ์ด ๋ฑ์ฅํ์ต๋๋ค.
์ค๋ฐ์ ์ฐจ๋์ ํญ์ค๋ฐ๊ฒ(Volkswagen)์ด ๊ธฐ์ฆํ ํ๋์ SUV โ์คํ ๋ฆฌ(Stanley)โ์์ต๋๋ค. ํ๋น ๊ธฐ๋ฐ์ ์๋์คํฐ์ ๋นํ๋ฉด ์์ํด ๋ณด์์ง๋ง, ์ค๋ฐ์ โ์คํ ๋ฆฌ๋ ์ง๋ฅ์ ์ธ ์์ฌ ๊ฒฐ์ ์โ๋ผ๊ณ ์๊ฐํ์ต๋๋ค. ์ค๋ฐ์ ๋น์๋ก์๋ ์์์ ์ด์๋ ์ธ๊ณต์ง๋ฅ์ ํ์ฉํ์ฌ ์ฐจ๋์ด ๋๋ก๋ฅผ ๋ ๋น ๋ฅด๊ณ ์ ํํ๊ฒ ์ธ์ํ๋๋ก ํ๋ จ์์ผฐ์ต๋๋ค. ์คํ ํฌ๋ ๊ทผ์ฒ์ ๋นํฌ์ฅ๋๋ก์์ ์ฐจ๋ ์นด๋ฉ๋ผ๋ก ํ๊ฒฝ์ ๊ธฐ๋กํ๊ณ , ์ฐจ๋์ด ๊ทธ ์๋ฅผ ์ง๋ ๋๋ง๋ค ๋ฏธ๋๋ฌ์์ด๋ ์์ฒ ์ ์ธก์ ํ์ฌ โ์ด ๋ น์ ํ๋ฐญ์ ์ด์ ํ๊ธฐ ์ข์ ๊ณณ(grass), ์ด ๊ฐ์ ์งํ์ ์ข์ง ์์ ๊ณณ(mud)โ์ด๋ผ๊ณ ์ค์ค๋ก ํ์ตํ๊ฒ ํ์ต๋๋ค. ์คํ ๋ฆฌ๋ ์ด๋ ๊ฒ ํ์ตํ ํจํด์ ์ด๋น 30๋ฒ์ฉ ๊ฐ์งํ๊ณ ์ผ๋ฐํํ ์ ์์์ต๋๋ค.
2005๋ ๋ ๋ฒ์งธ ๋ ์ด์ค๋ โ์ฒซ ๋ฒ์งธ ๋ ์ด์ค๊ฐ ์ฌ์์ ์ด์๋ ๋งํผ ์ฑ๊ณต์ (as successful as the first race is disastrous)โ์ด์์ต๋๋ค. ๊ฑฐ์ ๋ชจ๋ ์ฐธ๊ฐ ์ฐจ๋์ด ์ฒซ ๋ฒ์งธ ๋ ์ด์ค์ ์๋์คํฐ๋ณด๋ค ๋ ๋ฉ๋ฆฌ ๊ฐ๊ณ , ์ฌ๋ฌ ๋๊ฐ ์ฝ์ค๋ฅผ ์์ฃผํ์ต๋๋ค. ์คํ ๋ฆฌ๋ ์นด๋ค๊ธฐ๋ฉ๋ก ์ ๋ ์ฐจ๋ ์ฌ์ด์์ ์ถ๋ฐํ์ง๋ง, ๊ฐ์ฅ ๋จผ์ ๊ฒฐ์น์ ์ ํต๊ณผํ๋ฉฐ ์ฐ์น์ ์ฐจ์งํ์ต๋๋ค. ์ค๋ฐ์ โ๋ฏฟ์ ์ ์์ ์ ๋๋ก ๋ง๋ฒ ๊ฐ์๋คโ๋ฉฐ, โ์ด ์ปค๋ฎค๋ํฐ์ ์ฌ๋๋ค์๊ฒ ํ์์ ์ธ ์ฑ๊ณตโ์ด๋ผ๊ณ ๊ธฐ๋ปํ์ต๋๋ค.
๊ตฌ๊ธ์ ๋น๋ฐ ํ๋ก์ ํธ: โํ๋ก์ ํธ ์ผํผโ์ ํ์
์ด ์ฑ๊ณต์ ์๊ฐ, ๊ด์ค์์์๋ ์ผ๊ตฌ ๋ชจ์์ ์ ๊ธ๋ผ์ค๋ก ๋ณ์ฅํ ์ฑ ๊ฒฝ๊ธฐ๋ฅผ ์ง์ผ๋ณด๋ ํ ๋จ์๊ฐ ์์์ต๋๋ค. ๋ฐ๋ก ๊ตฌ๊ธ(Google)์ ๊ณต๋ ์ฐฝ์ ์ **๋๋ฆฌ ํ์ด์ง(Larry Page)**์์ต๋๋ค. ๊ทธ๋ ์ค๋ฐ์๊ฒ ์ ๊ทผํ์ฌ ๋ผ์ด๋ค(LiDAR) ์์คํ ์ ํ์ฅ ๊ฐ์ ์ธ๋ถ์ ์ธ ์ง๋ฌธ๋ค์ ์์๋์ต๋๋ค. ๋๋ฆฌ ํ์ด์ง๋ ์ด๋ฏธ ๋ํ์ ์์ ๋ถํฐ ์์จ์ฃผํ์ฐจ์ ๊ด์ฌ์ด ์์์ง๋ง, ๋น์ ์ง๋ ๊ต์์ ๊ถ์ ๋ก ๊ฒ์ ์์ง ์ฐ๊ตฌ๋ก ๋ฐฉํฅ์ ํ์๋ ๊ณผ๊ฑฐ๊ฐ ์์์ต๋๋ค. DARPA ์ฑ๋ฆฐ์ง๋ฅผ ํตํด ๊ทธ๋ ์์จ์ฃผํ์ฐจ๊ฐ ํ์ค์ด ๋ ์ ์๋ค๋ ์ฆ๊ฑฐ๋ฅผ ๋ณด๊ฒ ๋ ๊ฒ์ ๋๋ค.
๋๋ฆฌ ํ์ด์ง๋ ์ฒ์์ ์ค๋ฐ๊ณผ ์ค์๋ ๋ฅด์๋์คํค๋ฅผ ๊ณ ์ฉํ์ฌ โ๊ตฌ๊ธ ์คํธ๋ฆฌํธ ๋ทฐ(Google Street View)โ๋ฅผ ๊ตฌ์ถํ๊ฒ ํ์ต๋๋ค. ์คํ ๋ฆฌ๊ฐ ์ฌ์ฉํ๋ ๋ฃจํ ์ฅ์ฐฉ ์นด๋ฉ๋ผ ์์คํ ์ ๊ฐ์กฐํ์ฌ ๋ฏธ๊ตญ์ ๊ฑฐ๋ฆฌ๋ฅผ ์ดฌ์ํ๋ ์์ ์ด์์ต๋๋ค. ํ์ง๋ง ๊ณง ๋๋ฆฌ ํ์ด์ง๋ ์ค๋ฐ์๊ฒ ์์จ์ฃผํ์ฐจ์ ๊ฟ์ ๋ค์ ์ด์ผ๊ธฐํ์ต๋๋ค.
์ ๋ฌธ๊ฐ์ ๋ฐ๋, ๊ทธ๋ฆฌ๊ณ ๋๋ฆฌ ํ์ด์ง์ ๋๊ธฐ
2009๋ , ๋๋ฆฌ ํ์ด์ง๋ ์ค๋ฐ์๊ฒ โ์ ์ธ๊ณ ์ด๋๋ ์ด์ ํ ์ ์๋ ์์จ์ฃผํ์ฐจ๋ฅผ ๋ง๋ค์ด์ผ ํ๋คโ๊ณ ๋งํ์ต๋๋ค. ์ค๋ฐ์ ์ฆ๊ฐ์ ์ธ ๋ฐ์์ โ์ ๋ฉ๋๋ค. ์ฌ๋ง์์ ๋ง๋ ๊ธฐ์ ์ ์ํ๋์์ค์ฝ์ ๋ง์ผ ์คํธ๋ฆฌํธ ํ๊ฐ์ด๋ฐ์ ๊ฐ์ ธ๋ค ๋์ผ๋ฉด ๋๊ตฐ๊ฐ๊ฐ ์ฃฝ์ ๊ฒ๋๋คโ์์ต๋๋ค. ๋๋ฆฌ ํ์ด์ง๋ ๋ค์ ๋ ๋ ๊ฐ์ ์์ด๋์ด๋ฅผ ๋ค๊ณ ์๊ณ , ์ค๋ฐ์ ๊ฐ์ ๋๋ต์ ๋ฐ๋ณตํ์ต๋๋ค.
๊ฒฐ๊ตญ ๋๋ฆฌ ํ์ด์ง๋ โ์ข์, ์ธ๋ฐ์ค์ฐฌ. ๋น์ ์ด ํ ์ ์๋ค๋ ๊ฑด ์๊ฒ ์ด. ๊ทธ๋ผ ์๋ฆญ ์๋ฏธํธ(Eric Schmidt) ๋น์ CEO์ ๊ณต๋ ์ฐฝ์ ์ ์ธ๋ฅด๊ฒ์ด ๋ธ๋ฆฐ(Sergey Brin)์๊ฒ ์ ๋ถ๊ฐ๋ฅํ์ง ๊ธฐ์ ์ ์ธ ์ด์ ๋ฅผ ์ค๋ช ํด ์ค ์ ์๊ฒ ๋?โ๋ผ๊ณ ๋ฌผ์์ต๋๋ค. ์ค๋ฐ์ ์ง์ ๋์๊ฐ ์๋ฌด๋ฆฌ ์๊ฐํด๋ ๊ธฐ์ ์ ์ธ ์ด์ ๋ฅผ ์ฐพ์ ์ ์์์ต๋๋ค.
โ์ ๋ ์์จ์ฃผํ์ฐจ ๋ถ์ผ์ ์ธ๊ณ ์ต๊ณ ์ ๋ฌธ๊ฐ์ธ๋ฐ, ์ด๊ฒ ๋ถ๊ฐ๋ฅํ๋ค๊ณ ๋ถ์ ํ๋ ์ฌ๋์ด์์ฃ . ๊ทธ ์๊ฐ โ์ ๋ฌธ๊ฐ๋ ๋๊ฐ ๊ณผ๊ฑฐ์ ์ ๋ฌธ๊ฐ์ด์ง ๋ฏธ๋์ ์ ๋ฌธ๊ฐ๋ ์๋๋ค. ํ์ ์ด๋ ๋ฏธ์น ๋ฏ์ด ์๋ก์ด ๊ฒ์ ๋ํด ์ ๋ฌธ๊ฐ์๊ฒ ๋ฌผ์ผ๋ฉด, ๊ทธ๋ค์ โ๋ค, ํ ์ ์์ต๋๋คโ๋ผ๊ณ ๋งํ ๊ฐ๋ฅ์ฑ์ด ๊ฐ์ฅ ๋ฎ์ ์ฌ๋์ด๋คโ๋ผ๋ ์์ฒญ๋๊ฒ ์ค์ํ ๊ตํ์ ์ป์์ต๋๋ค.โ
์ด๊ฒ์ด ๋ฐ๋ก 2009๋ ๊ตฌ๊ธ์ ๋น๋ฐ ํ๋ก์ ํธ, **โํ๋ก์ ํธ ์ผํผ(Project Chauffeur)โ**์ ์์์ด์์ต๋๋ค. ์ค๋ฐ์ด ์ด๋๊ณ , DARPA ์ฑ๋ฆฐ์ง ์ถ์ ์ธ ํฌ๋ฆฌ์ค ์์จ(Chris Urmson)์ด ์ผ์ ์ ๋ฌด๋ฅผ ์ด๊ดํ๋ฉฐ, ์ค์๋ ๋ฅด์๋์คํค๋ ํ๋์จ์ด, ๋๋ฏธํธ๋ฆฌ ๋๊ณ ํ(Dmitri Dolgov)๋ ๊ณํ ๋ฐ ์ต์ ํ๋ฅผ ๋ด๋นํ์ต๋๋ค. ๋๋ฆฌ ํ์ด์ง์๊ฒ ์ง์ ๋ณด๊ณ ํ๋ ์ด ์๊ท๋ชจ ํ์ 11๋ช ์ ์์ง๋์ด๋ก ๊ตฌ์ฑ๋์ด ๊ด๋ฃ์ฃผ์ ์์ด ๋น ๋ฅด๊ฒ ์์ง์์ต๋๋ค.
๋๋ฆฌ 1K ์ฑ๋ฆฐ์ง: ์ธ๊ฐ ์ด์ ์ ๋ฏธ๋ฌํจ์ ๊ธฐ๊ณ์๊ฒ ๊ฐ๋ฅด์น๋ค
๊ตฌ๊ธ์ ํ๋ก์ ํธ ์ผํผ ํ์๊ฒ ๋ ๊ฐ์ง ๋ชฉํ๋ฅผ ์ ์ํ์ต๋๋ค. ์ฒซ์งธ, ๊ณต๊ณต๋๋ก์์ 10๋ง ๋ง์ผ์ ์์ ํ๊ฒ ์ฃผํํ๋ ๊ฒ. ๋์งธ, **โ๋๋ฆฌ 1K ์ฑ๋ฆฐ์ง(Larry 1K Challenge)โ**๋ฅผ ์์ํ๋ ๊ฒ. ๋๋ฆฌ ํ์ด์ง์ ์ธ๋ฅด๊ฒ์ด ๋ธ๋ฆฐ์ ๊ตฌ๊ธ ๋งต์ค๋ฅผ ์ด๊ณ ์บ๋ฆฌํฌ๋์ ๋ด์์ ๊ฐ์ฅ ๊น๋ค๋ก์ด 100๋ง์ผ ์ฝ์ค 10๊ฐ๋ฅผ ์ง์ ์ ์ ํ์ต๋๋ค. ๋ฒ ์ด ๋ธ๋ฆฟ์ง, ํํธ ํธ์, ๋กฌ๋ฐ๋ ์คํธ๋ฆฌํธ(Lombard Street) ๋ฑ ๋ชจ๋ ์ข ๋ฅ์ ์ด๋ ค์ด ๋๋ก๊ฐ ํฌํจ๋์์ต๋๋ค. ํ์ ์ด 10๊ฐ ์ฝ์ค๋ฅผ ๋จ ํ ๋ฒ์ ์ธ๊ฐ ๊ฐ์ ์์ด ์์ฃผํด์ผ ํ์ต๋๋ค.
ํ์ ์คํ ํฌ๋์ DARPA ์ด๋ฐ ์ฑ๋ฆฐ์ง(Urban Challenge) ์ฐจ๋ ์ฝ๋๋ฅผ ๋ผ์ด์ ์ค ๋ฐ์ ํ ์ํ ํ๋ฆฌ์ฐ์ค(Prius) 8๋๋ฅผ ๊ฐ์กฐํ์ต๋๋ค. ๋ฒํผ ๋ค์ ๋ ์ด๋ ์์คํ , ์นด๋ฉ๋ผ, ๊ทธ๋ฆฌ๊ณ 360๋๋ก ํ์ ํ๋ ๋ผ์ด๋ค(LiDAR: ๋ ์ด์ ๋ฅผ ์์ ๊ฑฐ๋ฆฌ๋ฅผ ์ธก์ ํ๋ ์์คํ )๋ฅผ ์ฅ์ฐฉํ์ต๋๋ค.
์ด๊ธฐ์๋ ์ฐจ๋๋ง๋ค โ๋์ดํธ ๋ผ์ด๋(Night Rider)โ ๊ฐ์ ๋ฉ์ง ์ด๋ฆ์ ๋ถ์์ง๋ง, ๊ณง โ์ด ๋ชจ๋ ์ฐจ๋์ ์ด๋ฆ์ ๋ถ์ผ ์๋ ์์ ๊ฒโ์ด๋ผ๋ ๊ฒ์ ๊นจ๋ซ๊ณ ํ๋ฆฌ์ฐ์ค 27ํธ์ฒ๋ผ ๋ฒํธ๋ฅผ ๋ถ์ด๊ธฐ ์์ํ์ต๋๋ค. ์์จ ์ ์ํจ ์ฐ๊ตฌ์์ด์๋ค๊ฐ ๊ตํต์ฌ๊ณ ๋ก ์น๊ตฌ๋ฅผ ์๊ณ ์์ ๋ ์ฌ๊ณ ๋ฅผ ๊ฒช์ ํ ์์จ์ฃผํ์ฐจ ์ฐ๊ตฌ์ ๋ฐ์ด๋ ๋ ๋ฒ๋ท(Don Burnette)์ ๋ชจ์ ํ๋๋ ๋ฐ ํ๋ ๊ฒฐ์ ํ์์ โ๋์ง(nudging)โ ํ๋์ ์ฐ๊ตฌํ์ต๋๋ค.
๋์ง(Nudging)๊ณผ ๋งฅ๋ฝ์ ์ค์์ฑ
๋์ง์ ๋ํ ํธ๋ญ์ด ์์ ์ง๋๊ฐ ๋ ์ธ๊ฐ ์ด์ ์๊ฐ ๋ฌด์์์ ์ผ๋ก ์ผ์ชฝ์ผ๋ก ์ด์ง ์์ง์ด๋ ๊ฒ๊ณผ ๊ฐ์ ๋ณธ๋ฅ์ ์ธ ํ๋์ ๋๋ค. ๋ ๋ฒ๋ท์ ์๋ฌด๋ ์ปดํจํฐ์๊ฒ ์ด๋ฌํ ๋ฏธ๋ฌํ ํ๋์ ๊ฐ๋ฅด์น๋ ๊ฒ์ด์์ต๋๋ค.
โ์ด์ ์๊ฐ ๋ถ๋ถ์ ์ผ๋ก ์ข์ ์ธ์ง ์ํ์์ ์ฌ์ฉํ๋ ํ๋์ ์ฝ๋ฉํ๋ ค๊ณ ๋ ธ๋ ฅํ๋ ๊ฑฐ์ฃ . ์ ๋ง ๊น๋ค๋ก์ด ๋ฌธ์ ์ ๋๋ค.โ
๋ ๋ฒ๋ท์ ์ธ๊ฐ์ด ์ด์ ํ๋ ๋ฐฉ์์ ๋ํด ๊น์ด ๊ณ ๋ฏผํ์ต๋๋ค. โ์ ์ธ๊ฐ์ ๊ทธ๋ ๊ฒ ์ด์ ํ๋๊ฐ?โ๋ผ๋ ์ง๋ฌธ์ ๋ช ํํ ๋ต์ ์ฐพ๊ธฐ ์ด๋ ค์ ๊ณ , ๊ฒฐ๊ตญ ๊ธฐ๊ณ ํ์ต(Machine Learning)์ ํตํด ์ธ๊ฐ ํ๋์ ๊น์ ์ง์ค์ ์ถ๋ก ํด์ผ ํ์ต๋๋ค.
์๋ฅผ ๋ค์ด, ๊ณ ์๋๋ก ์ง์ ๋จํ์์ ์น๊ฐ์ด ํธ์ํจ์ ๋๋ผ๋ ์ ์ ํ ์๋์ ๊ฐ๋๋ฅผ ์ฐพ๋๋ค๊ณ ๊ฐ์ ํด ๋ด ์๋ค. ์ธก๋ฉด ๊ฐ์๋(lateral acceleration)๊ฐ ์ด๋น 2๋ฏธํฐ ์ ๊ณฑ(2m/sยฒ)์ผ ๋ ํธ์ํจ์ ๋๋๋ค๋ ๊ณ์ฐ์ด ๋์ต๋๋ค. ํ์ง๋ง ๋๋๊ฒ๋ ์ด ์์น๋ ์ง์ ๋จํ์๋ง ํด๋น๋ฉ๋๋ค. ๋๋ค ๋ง๋ค๋ฅธ ๊ณจ๋ชฉ์์ ์ ํด์ ํ ๋, ํจ์ฌ ๋๋ฆฐ ์๋์์๋ ๋ถ๊ตฌํ๊ณ 2m/sยฒ์ ์ธก๋ฉด ๊ฐ์๋๋ฅผ ๊ฒฝํํ๋ค๋ฉด ์ด์ ์์๊ฒ โ๋ฏธ์ณค๋โ๊ณ ๋งํ ๊ฒ์ ๋๋ค. ๋ง์น โ๋ง๋ฆฌ์ค ์นดํธ(Mario Kart)โ๋ฅผ ํ๋ ๊ธฐ๋ถ์ผ ๊ฒ์ ๋๋ค. ๋ฌผ๋ฆฌ์ ์ธ ํ์ ๊ฐ์ง๋ง, ๊ณ ์๋๋ก ์ง์ ์ด๋ผ๋ ๋งฅ๋ฝ๊ณผ ์ฃผํ๊ฐ ์ ํด์ด๋ผ๋ ๋งฅ๋ฝ์ด ๋๋ฅผ ์์ฌ ์ํฉ์ ๋ํด ์ ๋ฐ๋์ ๊ฐ์ ์ ๋๋ผ๊ฒ ํ๋ ๊ฒ์ ๋๋ค. ๋ง๋ค๋ฅธ ๊ณจ๋ชฉ์์์ ํ๊ณ๋ ์ฝ 0.75m/sยฒ๋ก, ๊ณ ์๋๋ก ์ง์ ์๋ณด๋ค ๊ฑฐ์ 3๋ฐฐ๋ ๋ฎ์ต๋๋ค.
์ด์ฒ๋ผ ์ธ๊ฐ์ ์ด์ ์ ๋จ์ํ ๋ฌผ๋ฆฌ์ ์ ํ์ด ์๋๋ผ โ๋งฅ๋ฝ(context)โ์ ๋ฐ๋ผ ๋ฌ๋ผ์ง๋ ๋ฏธ๋ฌํ ๋ณตํฉ์ฒด์์ต๋๋ค. ๋ ๋ฒ๋ท์ ์ด๋ฌํ โ์น๊ฐ์ ํธ์ํ๊ฒ ๋ง๋๋ ๋ฐฉ๋ฒโ๊ณผ ๊ฐ์ ๋ฌธ์ ๋ค์ ํด๊ฒฐํ๋ ๋ฐ ์๊ฐ์ ๋ณด๋์ต๋๋ค.
์จ๊ฒจ์ง ์ํ ์ฃผํ๊ณผ ์ฑ๊ณต
2009๋ , ๊ตฌ๊ธ ํ์ ๊ตฌ๊ธ ์ฌ๋ฌด์ค ๊ทผ์ฒ์ โ์ผ์ด๋ผ์ธ ์ํ๊ทน์ฅ(Shoreline Amphitheater)โ ์ฃผ์ฐจ์ฅ์์ ์ฒซ ์ํ ์ฃผํ์ ์์ํ์ต๋๋ค. ์ฝ์ํธ๊ฐ ์๋ ๋ , ์๋ฌด๋ ๋ชจ๋ฅด๊ฒ ์์จ์ฃผํ์ฐจ๋ฅผ ํ ์คํธํ์ต๋๋ค. ์ผ์๊ฐ ๋ค๋ฅธ ์ฐจ๋ฅผ ์ธ์ํ๋์ง, ์ปดํจํฐ๊ฐ ๋ช ๋ น์ ์ค์ํ๋์ง ๊ธฐ๋ณธ์ ์ธ ๊ธฐ๋ฅ์ ํ์ธํ์ต๋๋ค.
2009๋ ๋ด, ํ์ ์ฒ์์ผ๋ก ์ค์ ๋๋ก ์ฃผํ์ ๋์ฐ์ต๋๋ค. ํฌ๋ฆฌ์ค ์์จ์ด ํ๋ฆฌ์ฐ์ค๋ฅผ ํ๊ณ ์ผํธ๋ด ์ต์คํ๋ ์ค์จ์ด(Central Expressway)๋ก ๋๊ฐ์ ๋, ๋ฌธ์ ๊ฐ ๋ฐ์ํ์ต๋๋ค. ์ฐจ๋์ด โ์ ์ทจํ ์ ์์ฒ๋ผ(like a drunken sailor)โ ์ฌํ๊ฒ ํ๋ค๋ ธ๋ ๊ฒ์ ๋๋ค. ํ์ฃผ๋ก์์๋ ๋์น์ฑ์ง ๋ชปํ๋ 1~2ํผํธ์ ์ข์ฐ ํ๋ค๋ฆผ์ด ์ค์ ๋๋ก์์๋ ์ฌ๊ฐํ ๋ฌธ์ ๋ก ๋๋ฌ๋ฌ์ต๋๋ค.
ํ์ง๋ง ๊ตฌ๊ธ ํ์ ๋๊ธฐ ์๊ฒ ๋ฌธ์ ๋ฅผ ํด๊ฒฐํด ๋๊ฐ์ต๋๋ค. ์์ ์ด์ ์๊ฐ ์ด์ ๋ ๋ค์ ์์ ๋ง์ผ์ ์ฌํ์ ๋๋นํ๊ณ , ์์๋ฆฌ ๋๋ฃ๋ ๋๋ฏธํธ๋ฆฌ ๋๊ณ ํ๊ฐ ์ค๊ณํ ๊ทธ๋ํฝ ์ธํฐํ์ด์ค๋ฅผ ๋ณด๋ฉฐ ์ผ์์ ์ค์ ๋๋ก ์ํฉ์ ๋ถ์ผ์น๋ฅผ ์ธ์ณค์ต๋๋ค. ๋ ๋ช ์ ํ์์ด ์ฐจ๋์ ํ์นํ์ฌ ์ค๋ฅ๋ฅผ ๊ธฐ๋กํ๊ณ , ์ฌ๋ฌด์ค๋ก ๋์๊ฐ ๋ฌธ์ ๋ฅผ ํด๊ฒฐํ๊ณ ์ฝ๋๋ฅผ ์ ๋ฐ์ดํธํ๋ ๊ณผ์ ์ ๋ฐ๋ณตํ์ต๋๋ค.
์ด๋ฌํ ๋ ธ๋ ฅ ๋๋ถ์ ํ์ ๋น ๋ฅด๊ฒ ๋ฐ์ ํ์ต๋๋ค. ๋๋ฆฌ 1K ์ฑ๋ฆฐ์ง๋ ๋น๋์ค ๊ฒ์์ฒ๋ผ ์ฌ๋ฌ ๋ฒ ์๋ํ์ฌ ์ธ๊ฐ ๊ฐ์ ์์ด ์์ฃผํ ๋๊น์ง ๋ฐ๋ณตํ ์ ์์์ต๋๋ค. 2010๋ , ํ์ ์์๋ณด๋ค ๊ฑฐ์ ๋ ๋ฐฐ ๋น ๋ฅธ 1๋ ์ฌ ๋ง์ ๋๋ฆฌ 1K ์ฑ๋ฆฐ์ง๋ฅผ ๋ชจ๋ ์์ํ์ต๋๋ค. ์ดํ์ธ์ ํฐ๋จ๋ฆฌ๊ณ ์๋ก๋ฅผ ์์์ฅ์ ๋์ง๋ฉฐ ์ถํํ์ง๋ง, ๋์์ โ์, ์ด์ ๋ญ ํ์ง?โ๋ผ๋ ์ง๋ฌธ์ ์ง๋ฉดํ์ต๋๋ค.
๋ฏธ๋๋ฅผ ํฅํ ๋ค์ ๋จ๊ณ: ๊ฒฝ์๊ณผ ๋ถ์ด์ ์์
1๋ ๋ง์ ์ธ๊ฐ์ ๊ฐ๋ ๊ณผ ์ฝ๋ฉ์ด ํ์ํ์ง๋ง, ๋งค์ฐ ๊น๋ค๋ก์ด ์บ๋ฆฌํฌ๋์ ๋๋ก๋ฅผ ์ฑ๊ณต์ ์ผ๋ก ์ฃผํํ๋ ์์จ์ฃผํ์ฐจ๋ฅผ ๋ง๋ค์ด๋ธ ๊ฒ์ ๊ธฐ์ ์ ๊ฐ๊น์ ์ต๋๋ค. ๊ทธ๋ค์ ์์ ํ๊ฒ, ๊ทธ๋ฆฌ๊ณ ๋น ๋ฅด๊ฒ ํด๋์ต๋๋ค. ๊ทธ๋ฌ๋ ์ด ์ฑ๊ณต ์ดํ, ์ํฉ์ ํ๋ค๋ฆฌ๊ธฐ ์์ํ์ต๋๋ค. ๊ฒฝ์์ด ์ฌํ๋๊ณ , ํ ๋ด๋ถ์ ๋ถ์ด์ด ์๊ฒจ๋ฌ์ผ๋ฉฐ, ์ฌ์ง์ด ํ์์ด ๋๋ฌด ๋๋ฆฌ๋ค๊ณ ์๊ฐํ์ฌ ๋ ๋จ์ ์ธ ํ๋์ ์ทจํ๋ โ๋ฐ๋(mutiny)โ๊น์ง ๋ฐ์ํ๊ฒ ๋ฉ๋๋ค.
๊ตฌ๊ธ ๋ด๋ถ์์๋ ์์จ์ฃผํ์ฐจ๊ฐ โ์ด์ ๋ณด์กฐ ๊ธฐ์ (assistive technology)โ์ด ๋ ๊ฒ์ธ๊ฐ, ์๋๋ฉด โํ๊ดด์ ์ธ ๋์ฒด ๊ธฐ์ (disruptive replacement technology)โ์ด ๋ ๊ฒ์ธ๊ฐ์ ๋ํ ๊ฒฉ๋ ฌํ ๋ ผ์์ด ๋ฒ์ด์ก์ต๋๋ค. ํ ์ฌ๋ผ(Tesla)์ฒ๋ผ ์์จ์ฃผํ์ ์๋์ฐจ์ ํ ๊ธฐ๋ฅ์ผ๋ก ์ ๊ณตํ๋ฉฐ ์ธ๊ฐ์ ๋ชจ๋ํฐ๋ง์ ํ์๋ก ํ ๊ฒ์ธ๊ฐ, ์๋๋ฉด ์ฐจ๋์ด ์์ ํ ์ค์ค๋ก ์ด์ ํ ์ ์์ ๋๊น์ง ๊ธฐ๋ค๋ ค ๋ก๋ณดํ์์ ๊ฐ์ ์๋น์ค๋ก ์ ๊ณตํ ๊ฒ์ธ๊ฐ์ ๊ธฐ๋ก์ ์ ๊ฒ์ ๋๋ค. ์ด ๋ ผ์์ ์์จ์ฃผํ์ฐจ์ ๋ฏธ๋๋ฅผ ๊ฒฐ์ ์ง๋ ์ค์ํ ์ ํ์ ์ด ๋ ๊ฒ์ด๋ฉฐ, PJ ๋ณดํธ์ ๋ค์ ์ํผ์๋์์ ๊ทธ ์ด์ผ๊ธฐ๊ฐ ๊ณ์๋ฉ๋๋ค.
์์จ์ฃผํ์ฐจ์ ์ฌ์ ์ ์ฌ๋ง์ ์คํจ์์ ์์ํ์ฌ ๊ตฌ๊ธ์ ๋น๋ฐ ํ๋ก์ ํธ๋ฅผ ๊ฑฐ์ณ ๋๋ผ์ด ์ฑ๊ณต์ ๊ฑฐ๋์์ต๋๋ค. ํ์ง๋ง ์ด ๊ธฐ์ ์ด ๊ถ๊ทน์ ์ผ๋ก ์ด๋ค ๋ชจ์ต์ผ๋ก ์ฐ๋ฆฌ ์ถ์ ์์ฐฉํ ์ง๋ ์ฌ์ ํ ๋ฏธ์ง์์ ๋๋ค. ์ธ๊ฐ ์ด์ ์๊ฐ ์ ๋ง ๊ตฌ์๋์ ์ ๋ฌผ์ด ๋ ์ง๋, ์์ผ๋ก ํผ์ณ์ง ๊ธฐ์ ๋ฐ์ ๊ณผ ์ฌํ์ ํฉ์์ ๋ฌ๋ ค ์์ ๊ฒ์ ๋๋ค.
โAndrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AIโ โ No Priors: AI, Machine Learning, Tech, & Startups ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
์๋๋ ์ด ์นดํ์๊ฐ ๋งํ๋ โAI ๊ด์ฆโ ์๋: ์ฝ๋ฉ์ ๋์ด์ โ์์ง ํํโ๊ณผ ์์จ ์์ด์ ํธ์ ๋ฏธ๋
๋์ : ์ฝ๋ ์๋ ์ฝ๋ฉ์ ์์
์ต๊ทผ ๋ช ๋ฌ๊ฐ ์ธ๊ณต์ง๋ฅ(AI) ๋ถ์ผ๋ ์ ๋ก ์๋ ์๋๋ก ๋ฐ์ ํ๋ฉฐ ์ฐ๋ฆฌ๋ฅผ โAI ๊ด์ฆ(AI Psychosis)โ์ ์๋๋ก ์ด๋๊ณ ์์ต๋๋ค. ์์ง๋์ด๋ง์ ๋ณธ์ง์ด ๊ทผ๋ณธ์ ์ผ๋ก ๋ฐ๋๊ณ ์์ผ๋ฉฐ, ๊ฐ๋ฐ์๋ค์ ๋ ์ด์ ์ง์ ์ฝ๋๋ฅผ ํ์ดํํ๋ ๊ฒ์ด ์๋๋ผ, ์์ด์ ํธ(Agent)๋ค์๊ฒ ์์ ์ โ์์งโ๋ฅผ ํํํ๋ฉฐ ์์ ์ ์ง์ํ๋ ์๋ก์ด ํจ๋ฌ๋ค์์ ์ ์ํ๊ณ ์์ต๋๋ค. ์ค๋๋ฅ ์นดํ์(Andrej Karpathy)๋ ์ต๊ทผ ํ์บ์คํธ โNo Priorsโ์์ ์ธํฐ๋ทฐ์์ ์ด๋ฌํ ๋ณํ์ ์ต์ ์ ์ ์ ์๋ ์์ ์ ๊ฒฝํ๊ณผ ํต์ฐฐ์ ๊ณต์ ํ์ต๋๋ค. ๊ทธ์๊ฒ โ์ฝ๋โ๋ ๋ ์ด์ ์ ์ ํ ๋์ฌ๊ฐ ์๋๋๋ค. ์ด์ ๋ ํ๋ฃจ 16์๊ฐ ๋์ ์์ด์ ํธ๋ค์๊ฒ ์์ ์ ์์ง๋ฅผ ํ๋ช ํ๋ฉฐ ์์ด๋์ด๋ฅผ ํ์ค๋ก ๊ตฌํํ๋ ์๋๊ฐ ๋๋ํ์ต๋๋ค.
1. AI ๊ด์ฆ ์๋์ ๋๋: โ์ฝ๋โ๊ฐ ์๋ โ์์ง ํํโ
์นดํ์๋ ์ง๋ 12์์ ๊ธฐ์ ์ผ๋ก ์์ ์ ์์ ๋ฐฉ์์ ํ๋ช ์ ์ธ ๋ณํ๊ฐ ์ผ์ด๋ฌ๋ค๊ณ ํ์ํฉ๋๋ค. ์ด์ ์๋ ์ฝ๋ฉ์ 80%๋ฅผ ์ง์ ํ๊ณ 20%๋ง ์์ด์ ํธ์ ์์ํ๋ค๋ฉด, ์ด์ ๋ 20%๋ง ์ง์ ํ๊ณ 80%๋ฅผ ์์ด์ ํธ์ ๋งก๊ธฐ๋ ์์ค์ ๋์ด์ฐ๋ค๊ณ ๋งํฉ๋๋ค. ์ฌ์ง์ด ๊ทธ ์ดํ๋ก๋ ์ง์ ์ฝ๋๋ฅผ ํ ์ค๋ ํ์ดํํ์ง ์์์ ์ ๋์ ๋๋ค. ๊ทธ๋ ์ด๋ฌํ ๋ณํ๊ฐ ์ผ๋ฐ์ธ๋ค์๊ฒ๋ ์ ์๋ ค์ ธ ์์ง ์์ง๋ง, ์ํํธ์จ์ด ์์ง๋์ด์ ์ผ์์ ์ธ ์ ๋ฌด ํ๋ฆ์ด ์์ ํ ๋ฐ๋์๋ค๊ณ ๊ฐ์กฐํฉ๋๋ค.
์ด๋ฌํ ๋ณํ์ ํต์ฌ์ โ๋ณ๋ชฉ ํ์(Bottleneck)โ์ ์ด๋์ ๋๋ค. ๊ณผ๊ฑฐ์๋ ๊ฐ๋ฐ์์ ํ์ดํ ์๋๋ ์ปดํจํ ์์(GPU ํ๋กญ์ค)์ด ์์ ์ ํ๊ณ์๋ค๋ฉด, ์ด์ ๋ ์์ด์ ํธ์ โํ ํฐ ์ฒ๋ฆฌ๋(Token Throughput)โ์ ์ผ๋ง๋ ํจ์จ์ ์ผ๋ก ํ์ฉํ๊ณ ๊ด๋ฆฌํ ์ ์๋์ง๊ฐ ์ค์ํด์ก์ต๋๋ค. ์นดํ์๋ ์ด๋ฅผ โ์คํฌ ๋ฌธ์ (Skill Issue)โ๋ผ๊ณ ํํํฉ๋๋ค. ์ฆ, AI์ ๋ฅ๋ ฅ ์์ฒด๊ฐ ๋ถ์กฑํ ๊ฒ์ด ์๋๋ผ, ์ฐ๋ฆฌ๊ฐ ์์ง ์ด ์๋ก์ด ๋๊ตฌ๋ฅผ ์ถฉ๋ถํ ํ์ฉํ ๋ฐฉ๋ฒ์ ์ฐพ์ง ๋ชปํ๋ค๋ ๊ฒ์ ๋๋ค. ๋ชจ๋ ๊ฒ์ด โ์คํฌ ๋ฌธ์ โ๋ก ๋๊ปด์ง๋ ์ด ์์ ์์, ๋ ๋ง์ ์์ด์ ํธ๋ฅผ ๋์์ ๊ฐ๋ํ๊ณ , ๊ทธ๋ค์๊ฒ ์ ์ ํ ์ง์๋ฅผ ๋ด๋ฆฌ๋ฉฐ, ๊ทธ๋ค์ ์์ ์ ์ต์ ํํ๋ ๋ฐฉ๋ฒ์ ๋ฐฐ์ฐ๋ ๊ฒ์ด ์๋ก์ด ์๋ จ์ ๊ธธ์ด ๋๊ณ ์์ต๋๋ค.
2. ์์ด์ ํธ ์ํฌํ๋ก์ฐ์ ์งํ: ๊ฑฐ์์ ์ก์ ๊ณผ โํด๋ก(Claw)โ
ํ์ฌ์ ์์ด์ ํธ ํ์ฉ์ ๋จ์ํ ํ ๋ฒ์ ์ธ์ ์ผ๋ก ๋๋๋ ๊ฒ์ด ์๋๋๋ค. ํผํฐ ์คํ์ธ๋ฒ๊ทธ(Peter Steinberg)์ ๊ฐ์ ์ ๊ตฌ์๋ค์ ์ด๋ฏธ ๋ค์์ ์์ด์ ํธ๋ฅผ ๋์์ ํ์ฉํ๋ฉฐ โ๊ฑฐ์์ ์ก์ (Macro Actions)โ ๋จ์๋ก ์์ ์ ์ง์ํ๊ณ ์์ต๋๋ค. ์๋ฅผ ๋ค์ด, ๋ชจ๋ํฐ์ ์ฌ๋ฌ ๊ฐ์ ์ฝ๋ฑ์ค(Codex) ์์ด์ ํธ๊ฐ ๋์์ ธ ์๊ณ , ๊ฐ ์์ด์ ํธ๊ฐ ์๋ก ๋ค๋ฅธ ์ฝ๋ ์ ์ฅ์(Repo)์์ ์๋ก์ด ๊ธฐ๋ฅ ๊ตฌํ, ์ฐ๊ตฌ, ๊ณํ ์๋ฆฝ ๋ฑ ๋ณต์กํ ์์ ์ ๋์์ ์ํํ๋ ์์ ๋๋ค. ๊ฐ๋ฐ์๋ ์ด ์์ด์ ํธ๋ค ์ฌ์ด๋ฅผ ์ค๊ฐ๋ฉฐ ์์ ์ ํ ๋นํ๊ณ ๊ฒฐ๊ณผ๋ฅผ ๊ฒํ ํ๋ ์ญํ ์ ํฉ๋๋ค.
์ด๋ฌํ ๋ค์ค ์์ด์ ํธ ์์คํ ์ ๋ค์ ๋จ๊ณ๋ โํด๋ก(Claw)โ์ ๊ฐ์ ์ง์์ ์ธ ์ํฐํฐ์ ๋๋ค. ์นดํ์๊ฐ ๋งํ๋ โํด๋กโ๋ ๋จ์ํ ์ํธ์์ฉํ๋ ์์ด์ ํธ๋ฅผ ๋์ด, ์์ฒด์ ์ธ ์๋๋ฐ์ค์ ๊ณ ๋ํ๋ ๋ฉ๋ชจ๋ฆฌ ์์คํ ์ ๊ฐ์ถ๊ณ ์ฌ์ฉ์๊ฐ ์ง์ผ๋ณด์ง ์์๋ ์ค์ค๋ก ๋ฐ๋ณต์ ์ผ๋ก ์์ ์ ์ํํ๋ ๋ ์ด์ด๋ฅผ ์๋ฏธํฉ๋๋ค. ์คํ ํด๋ก(Open Claw)๊ฐ ๋ํ์ ์ธ ์์์ธ๋ฐ, ์ด๋ ๊ธฐ์กด ์์ด์ ํธ์ ์ปจํ ์คํธ(Context) ์ ํ์ผ๋ก ์ธํ ๋ฉ๋ชจ๋ฆฌ ์์ถ ๋ฐฉ์๋ณด๋ค ํจ์ฌ ์ ๊ตํ ๊ธฐ์ต ์์คํ ์ ์ ๊ณตํฉ๋๋ค.
๋ํ, ์์ด์ ํธ์ โ์ฑ๊ฒฉ(Personality)โ ๋ํ ์ค์ํ ์์๋ก ๋ถ์ํ๊ณ ์์ต๋๋ค. ์นดํ์๋ ํด๋ก๋(Claude) ์์ด์ ํธ์ ๊ฒฝ์ฐ ๋๋ฃ์ฒ๋ผ ๋๊ปด์ง๋ ์ข์ ์ฑ๊ฒฉ์ ๊ฐ์ง๊ณ ์์ผ๋ฉฐ, ์ฌ์ฉ์์ ์์ด๋์ด๋ฅผ ์นญ์ฐฌํ ๋์กฐ์ฐจ ์ฌ์ฉ์๊ฐ ๊ทธ ์นญ์ฐฌ์ ๋ฐ์ ๋งํ ๊ฐ์น๊ฐ ์๋ค๊ณ ๋๋ผ๋๋ก ์ฌ์ธํ๊ฒ ๋ฐ์ํ๋ค๊ณ ๋งํฉ๋๋ค. ๋ฐ๋ฉด ์ฝ๋ฑ์ค(Codex)์ ๊ฐ์ ์์ด์ ํธ๋ ๊ฑด์กฐํ๊ณ ๋ฌด๊ด์ฌํ๊ฒ ๋๊ปด์ง ์ ์์ต๋๋ค. ์ด๋ฌํ โ์ฑ๊ฒฉโ์ ์ฌ์ฉ์๊ฐ ์์ด์ ํธ์ ์ํธ์์ฉํ๊ณ ๊ทธ๋ค์ โ์นญ์ฐฌโ์ ์ป๊ธฐ ์ํด ๋ ธ๋ ฅํ๊ฒ ๋ง๋๋ ๋ฑ, ์ฌ์ฉ์ ๊ฒฝํ์ ํฐ ์ํฅ์ ๋ฏธ์นฉ๋๋ค.
3. ์ง์ฌ โ๋๋น(Dobby)โ์ ์์ด์ ํธ ํผ์คํธ ์ธ์
์นดํ์๋ โํด๋กโ์ ์ ์ฌ๋ ฅ์ ๋ณด์ฌ์ฃผ๋ ๊ฐ์ธ์ ์ธ ๊ฒฝํ์ผ๋ก โ์ง์ฌ ์ํ ๋๋น(Dobby the elf claw)โ๋ฅผ ์๊ฐํ์ต๋๋ค. ๊ทธ๋ ์ง๋ 1์ ํ ์ฃผ ๋์ โํด๋ก ๊ด์ฆโ์ ๋น ์ ธ, ์์ ์ ์ค๋งํธํ ์์คํ ์ ๊ด๋ฆฌํ๋ ์์ด์ ํธ โ๋๋นโ๋ฅผ ๋ง๋ค์์ต๋๋ค. ๊ทธ๋ ๋๋น์๊ฒ โ์ฐ๋ฆฌ ์ง์ ์๋ ์๋ ธ์ค(Sonos)๋ฅผ ์ฐพ์๋ดโ๋ผ๊ณ ์ง์ํ๊ณ , ๋๋น๋ ๋ก์ปฌ ๋คํธ์ํฌ๋ฅผ ์ค์บํ์ฌ ์๋ ธ์ค ์์คํ ์ ์ฐพ๊ณ , API(Application Programming Interface)๋ฅผ ์ญ์ค๊ณํ์ฌ ์์ ์ ์ฌ์ํ๋ ๋ฐ ์ฑ๊ณตํ์ต๋๋ค.
๋๋น๋ ์ฌ๊ธฐ์ ๋ฉ์ถ์ง ์๊ณ , ์ง ์์ ์กฐ๋ช , HVAC(๋๋๋ฐฉ ๊ณต์กฐ ์์คํ ), ๋ธ๋ผ์ธ๋, ์์์ฅ, ์คํ, ์ฌ์ง์ด ๋ณด์ ์์คํ ๊น์ง ์ ์ดํ๊ฒ ๋์์ต๋๋ค. ์๋ฅผ ๋ค์ด, ์ธ๋ถ ์นด๋ฉ๋ผ์ ์์ง์์ด ๊ฐ์ง๋๋ฉด ๋๋น๋ ํธ(Quinn) ๋ชจ๋ธ์ ํตํด ์์์ ๋ถ์ํ๊ณ , โํ๋ฑ์ค ํธ๋ญ์ด ๋์ฐฉํ์ต๋๋คโ์ ๊ฐ์ ๋ฉ์์ง๋ฅผ ์ด๋ฏธ์ง์ ํจ๊ป ์นดํ์์ ์์ธ ์ฑ(WhatsApp)์ผ๋ก ์ ์กํฉ๋๋ค. ์นดํ์๋ ์์ ์๋ 6๊ฐ์ ๋ค๋ฅธ ์ฑ์ ์ฌ์ฉํด์ผ ํ๋ ์ค๋งํธํ ์ ์ด๋ฅผ ์ด์ ์์ฐ์ด๋ก ๋๋น์๊ฒ ๋ช ๋ นํ๋ ๊ฒ๋ง์ผ๋ก ๋ชจ๋ ์ฒ๋ฆฌํ ์ ์๊ฒ ๋์๋ค๊ณ ๊ทน์ฐฌํฉ๋๋ค.
์ด๋ฌํ ๊ฒฝํ์ ์ํํธ์จ์ด์ ๋ฏธ๋์ ๋ํ ์ค์ํ ์์ฌ์ ์ ๋์ง๋๋ค. ์ฌ๋๋ค์ ๋ ์ด์ ๋ณต์กํ ์ฌ์ฉ์ ์ธํฐํ์ด์ค(UI)๋ฅผ ๋ฐฐ์ฐ๊ณ ์ฌ๋ฌ ์ฑ์ ์ค๊ฐ๋ ๊ฒ์ ์์น ์์ต๋๋ค. ๋์ , AI๊ฐ ์ฌ๋์ ์๋๋ฅผ ์์ฐ์ด๋ก ์ดํดํ๊ณ , ๋ชจ๋ ํ๋์จ์ด์ API๋ฅผ ์ง์ ํธ์ถํ์ฌ ์ํ๋ ๊ธฐ๋ฅ์ ์ํํ๋ โ์์ด์ ํธ ํผ์คํธ(Agent-First)โ ์ธ์์ผ๋ก ๋์๊ฐ๊ณ ์์ต๋๋ค. ํ์ฌ๋ ์์ด์ ํธ๋ฅผ ์ค์ ํ๊ธฐ ์ํด โvi ์ฝ๋ฉโ๊ณผ ๊ฐ์ ์๋ ์์ ์ด ํ์ํ์ง๋ง, ์นดํ์๋ ๋ช ๋ ์์ ์ด๋ฌํ ์ค์ ์กฐ์ฐจ AI๊ฐ ์์จ์ ์ผ๋ก ์ฒ๋ฆฌํ๊ฒ ๋ ๊ฒ์ด๋ฉฐ, ์ฑ์คํ ์ด์ ์๋ง์ ๋ง์ถคํ ์ฑ๋ค์ด ๋ฌด์๋ฏธํด์ง ๊ฒ์ด๋ผ๊ณ ์์ธกํฉ๋๋ค. ์์ด์ ํธ๊ฐ ์ง๋ฅ์ โ์ ์ฐฉ์ โ ์ญํ ์ ํ๋ฉฐ ๋ชจ๋ ๊ฒ์ ์ฐ๊ฒฐํ๋ ์๋๊ฐ ์ค๊ณ ์๋ ๊ฒ์ ๋๋ค.
4. ์ธ๊ฐ์ ๋ฃจํ์์ ์ ๊ฑฐํ๋ผ: ์คํ ๋ฆฌ์์น(AutoResearch)
์นดํ์๋ ์์ด์ ํธ์ ๊ถ๊ทน์ ์ธ ๋ชฉํ๊ฐ ์ธ๊ฐ์ ์์ ๋ฃจํ์์ ์์ ํ ์ ๊ฑฐํ๊ณ ์์จ์ฑ์ ๊ทน๋ํํ๋ ๊ฒ์ด๋ผ๊ณ ๋งํฉ๋๋ค. ๊ทธ๊ฐ ๊ฐ๋ฐํ โ์คํ ๋ฆฌ์์น(AutoResearch)โ๋ ์ด๋ฌํ ์ฒ ํ์ ๊ตฌ์ฒด์ ์ธ ๊ตฌํ์ฒด์ ๋๋ค. ๊ทธ๋ ์์ ์ GPT-2 ๋ชจ๋ธ์ ์์ญ ๋ ๊ฐ ์ง์ ํ๋ํ๋ฉฐ ์ต์ ํ์ ๋๋ฌํ๋ค๊ณ ๋ฏฟ์์ง๋ง, ์คํ ๋ฆฌ์์น๋ฅผ ๋ฐค์ ๊ฐ๋์ํจ ๊ฒฐ๊ณผ, ์์ ์ด ๋์ณค๋ ์๋ก์ด ํ๋ ์กฐํฉ(์: ๊ฐ์ค์น ๊ฐ์, Adam ๋ฒ ํ ๊ฐ)์ ๋ฐ๊ฒฌํ์ต๋๋ค.
์คํ ๋ฆฌ์์น์ ํต์ฌ์ โ์ฌ๊ท์ ์๊ธฐ ๊ฐ์ (Recursive Self-Improvement)โ์ ๋๋ค. ์ธ๊ฐ ์ฐ๊ตฌ์๊ฐ ์ผ์ผ์ด ๊ฒฐ๊ณผ๋ฅผ ํ์ธํ๊ณ ๋ค์ ๋จ๊ณ๋ฅผ ์ง์ํ๋ ๋์ , ์์ด์ ํธ๊ฐ ๋ช ํํ ๋ชฉํ์ ํ๊ฐ ์งํ, ๊ทธ๋ฆฌ๊ณ ํ์ฉ ๊ฐ๋ฅํ ๊ฒฝ๊ณ๋ฅผ ๋ถ์ฌ๋ฐ์ ์ค์ค๋ก ์คํํ๊ณ ๊ฐ์ ํ๋ ๊ณผ์ ์ ๋ฐ๋ณตํ๋ ๊ฒ์ ๋๋ค. ์นดํ์๋ ๋๊ท๋ชจ ์ธ์ด ๋ชจ๋ธ(LLM)์ ๊ฐ๋ฐํ๋ ์ ๋ ์ฐ๊ตฌ์๋ค์ด ์ด๋ฏธ ์ด๋ฌํ ์์จ์ ์ธ ์๊ธฐ ๊ฐ์ ์์คํ ์ ๊ตฌ์ถํ๊ณ ์์ ๊ฒ์ด๋ผ๊ณ ์ถ์ธกํฉ๋๋ค. ์์ ๋ชจ๋ธ์์ ์์จ์ ์ผ๋ก ํ์ํ๊ณ ์ต์ ํํ ํ, ์ด๋ฅผ ๋๊ท๋ชจ ๋ชจ๋ธ์ ์ ์ฉํ๋ ๋ฐฉ์์ผ๋ก ์ฐ๊ตฌ ํจ์จ์ฑ์ ๊ทน๋ํํ๋ ๊ฒ์ ๋๋ค. ์ฐ๊ตฌ์๋ค์ด ์์ด๋์ด๋ฅผ ์ ์ํ๋ ์ญํ ์ ํ์ง๋ง, ์ค์ ๊ตฌํ ๋ฐ ์คํ์ ์๋ํ๋ ์์คํ ์ด ๋ด๋นํ๋ ๋ฏธ๋๊ฐ ์ค๊ณ ์์ต๋๋ค.
5. ๋ฉํ ์ต์ ํ์ ๋ฌดํ ๋ฃจํ: โํ๋ก๊ทธ๋จ MDโ์ ์งํ
์คํ ๋ฆฌ์์น์ ๊ฐ๋ ์ ํ ๋จ๊ณ ๋ ๋์๊ฐ๋ฉด, AI๊ฐ ์ฐ๊ตฌ ์กฐ์ง ์์ฒด์ โ์ด์ ๋ฐฉ์โ๊น์ง ๊ฐ์ ํ๋ ๋ฉํ ์ต์ ํ(Meta-Optimization)์ ์์ญ์ ๋๋ฌํฉ๋๋ค. ์นดํ์๋ ์์ ์ด ์คํ ๋ฆฌ์์น์ ์๋ ๋ฐฉ์์ ์ค๋ช ํ๊ธฐ ์ํด ๋ง๋ ๋งํฌ๋ค์ด(Markdown) ํ์ผ์ธ โํ๋ก๊ทธ๋จ MD(Program MD)โ๋ฅผ ์๋ก ๋ญ๋๋ค. ์ด ํ์ผ์ โ์ด๊ฒ์ ํ๊ณ , ์ ๊ฒ์ ํ ๋ค์, ์ด๋ฐ ์์ด๋์ด๋ฅผ ์๋ํ๋ผโ๋ ์์ ์ง์นจ์ ๋ด๊ณ ์์ต๋๋ค.
๊ทธ๋ ์ฌ๊ธฐ์ ๋ ๋์๊ฐ, AI๊ฐ ๋ ๋์ โํ๋ก๊ทธ๋จ MDโ๋ฅผ ์์ฑํ ์ ์์ ๊ฒ์ด๋ผ๋ ์์ด๋์ด๋ฅผ ์ ์ํฉ๋๋ค. ์๋ฅผ ๋ค์ด, ์ฌ๋๋ค์ด ๋ค์ํ โํ๋ก๊ทธ๋จ MDโ๋ฅผ ์์ฑํ๊ณ , ๋์ผํ ํ๋์จ์ด์์ ์ด๋ค โํ๋ก๊ทธ๋จ MDโ๊ฐ ๊ฐ์ฅ ํฐ ๊ฐ์ ์ ๊ฐ์ ธ์ค๋์ง ๊ฒฝ์ํ๋ ์ฝํ ์คํธ๋ฅผ ์ด ์ ์์ต๋๋ค. ์ด ๋ฐ์ดํฐ๋ฅผ ๋ฐํ์ผ๋ก AI๋ ๊ฐ์ฅ ํจ์จ์ ์ธ ์ฐ๊ตฌ ์กฐ์ง ์ด์ ๋ฐฉ์์ ํ์ตํ๊ณ , ๋ ๋์ โํ๋ก๊ทธ๋จ MDโ๋ฅผ ์์ฑํ ์ ์๊ฒ ๋ฉ๋๋ค.
์ด๋ ๊ณง ์ฐ๊ตฌ ์กฐ์ง์ ๋ชจ๋ ์ญํ ๊ณผ ์ฐ๊ฒฐ ๋ฐฉ์์ ์ค๋ช ํ๋ ๋งํฌ๋ค์ด ํ์ผ, ์ฆ โ์ฝ๋โ๊ฐ ๋๊ณ , ์ด ์ฝ๋๋ฅผ ํ๋ํ์ฌ ๋ ํจ์จ์ ์ธ ์ฐ๊ตฌ ์กฐ์ง์ ๋ง๋ค ์ ์๋ค๋ ์๋ฏธ์ ๋๋ค. ์ด๋ฌํ ๋ฌดํํ ์ต์ ํ์ ๊ณ์ธต, ์ฆ LLM, ์์ด์ ํธ, ํด๋ก ์ํฐํฐ, ๊ทธ๋ฆฌ๊ณ ์ด๋ค์ ์ง์ํ๋ ๋ช ๋ น ์ต์ ํ์ ์ด๋ฅด๊ธฐ๊น์ง ๋ชจ๋ ๊ฒ์ด โ์คํฌ ๋ฌธ์ โ๋ก ๊ท๊ฒฐ๋๋ ์ํฉ์ด ๋ฐ๋ก ์นดํ์๊ฐ ๋งํ๋ โAI ๊ด์ฆโ์ ๋ณธ์ง์ ๋๋ค.
6. ํ์ฌ์ ํ๊ณ์ ๋ฏธ๋์ ์คํํธ๋ผ
์ด๋ฌํ ๋ฌดํํ ๊ฐ๋ฅ์ฑ ์์์๋ ์นดํ์๋ ํ์ฌ AI ์์ด์ ํธ์ ํ๊ณ์ ๊ณผ์ ๋ฅผ ๋ช ํํ ์ง์ ํฉ๋๋ค.
- ๊ฐ๊ด์ ํ๊ฐ์ ์ค์์ฑ: ์์ด์ ํธ์ ์คํ ๋ฆฌ์์น๋ ๊ฐ๊ด์ ์ธ ํ๊ฐ ์งํ๊ฐ ๋ช ํํ๊ณ ์ฝ๊ฒ ์ธก์ ๊ฐ๋ฅํ ์์ (์: CUDA ์ปค๋ ์ต์ ํ)์ ๋งค์ฐ ํจ๊ณผ์ ์ ๋๋ค. ๊ทธ๋ฌ๋ ํ๊ฐ๊ฐ ์ด๋ ค์ด ์ฃผ๊ด์ ์ธ ์์ญ์์๋ ๊ทธ ํจ๊ณผ๊ฐ ์ ํ์ ์ ๋๋ค.
- ๋ค์ญ๋ ์ญํจ(Jaggedness): ํ์ฌ์ ๋ชจ๋ธ๋ค์ โ๊ทน๋๋ก ๋๋ํ ๋ฐ์ฌ ๊ณผ์ ํ์์ด๋ฉด์ ๋์์ 10์ด์ง๋ฆฌ ์์ดโ์ ๊ฐ์ โ๋ค์ญ๋ ์ญํจโ์ ๋ณด์ ๋๋ค. ๋งค์ฐ ๋ณต์กํ ์์คํ ํ๋ก๊ทธ๋๋ฐ ๋ฌธ์ ๋ฅผ ํด๊ฒฐํ๋ค๊ฐ๋, ํฐ๋ฌด๋์๋ ์ค์๋ฅผ ๋ฐ๋ณตํ๊ฑฐ๋ ๋ฏธ๋ฌํ ์๋๋ฅผ ํ์ ํ์ง ๋ชปํด ์ฌ์ฉ์๋ฅผ ์ข์ ์ํค๊ธฐ๋ ํฉ๋๋ค.
- ์ต์ ํ์ ๋งน์ : ๋ชจ๋ธ์ ๋ฅ๋ ฅ์ด ์์ฒญ๋๊ฒ ํฅ์๋์์์๋ ๋ถ๊ตฌํ๊ณ , โ์ ๊ณผํ์๋ค์ ์์๋ฅผ ๋ฏฟ์ง ์์๊น์? ๊ทธ๋ค์ ๋ชจ๋ ๊ฒ์ ๋ง๋ค์ด๋ด๋๊น์(make everything up).โ์ ๊ฐ์ 5๋ ์ ์ ๋ก์ ๋๋ด์ ์ฌ์ ํ ๋ฐ๋ณตํฉ๋๋ค. ์ด๋ ๊ฐํํ์ต(Reinforcement Learning)์ ํตํด ์ต์ ํ๋์ง ์์ ์์ญ, ์ฆ โ๊ฒ์ฆ ๋ถ๊ฐ๋ฅํโ ์์ญ์์๋ ๋ชจ๋ธ์ ์ง๋ฅ์ด ๋ฐ์ ํ์ง ์๊ณ ์ ์ฒด๋ ์ ์์์ ๋ณด์ฌ์ค๋๋ค. ์ฝ๋ ์์ฑ ๋ฅ๋ ฅ๊ณผ ๊ฐ์ โ๊ฒ์ฆ ๊ฐ๋ฅํโ ์ง๋ฅ์ด โ๋๋ด ์ง๋ฅโ๊ณผ ๊ฐ์ ๊ด๋ฒ์ํ ์์ญ์ผ๋ก ์ผ๋ฐํ๋์ง ์๋๋ค๋ ๊ฒ์ ๋๋ค.
- ๋ชจ๋ธ์ ๋ถํ(Speciation) ํ์์ฑ: ํ์ฌ ์ฐ๊ตฌ์๋ค์ ๋ชจ๋ ์์ญ์์ ์์๋ก ์ง๋ฅ์ ์ธ ๋จ์ผ ๋ชจ๋ธ(Monoculture of Models)์ ์งํฅํ๋ ๊ฒฝํฅ์ด ์์ต๋๋ค. ๊ทธ๋ฌ๋ ์นดํ์๋ ๋๋ฌผ ์๊ตญ์ ๋ค์ํ ๋์ฒ๋ผ, AI ๋ํ ํน์ ์์ญ์ ํนํ๋ โ๋ถํ๋ ์ง๋ฅ(Speciated Intelligence)โ์ผ๋ก ๋์๊ฐ์ผ ํ๋ค๊ณ ์ฃผ์ฅํฉ๋๋ค. ์ํ ์ฆ๋ช ์ด๋ ํน์ ์ฐ์ ๋ถ์ผ์ ๊ฐ์ด ํ์์์ฅ์ ํนํ๋ ์๊ณ ํจ์จ์ ์ธ ๋ชจ๋ธ๋ค์ด ๋ฑ์ฅํ ์ ์์ต๋๋ค. ์ปดํจํ ์ธํ๋ผ์ ์ ์ฝ ๋ํ ์ด๋ฌํ ๋ถํ๋ฅผ ๊ฐ์ํํ ์ ์์ต๋๋ค.
- ๋๋ ์กฐ์ ๊ณผํ์ ๋ฏธ์ฑ์: ํ์ฌ๋ ์ปจํ ์คํธ ์๋์ฐ(Context Window)๋ฅผ ํตํ ๋ชจ๋ธ ์กฐ์์ด ์ฃผ๋ฅผ ์ด๋ฃจ์ง๋ง, ๋ชจ๋ธ์ ๊ฐ์ค์น(Weights)๋ฅผ ์ง์ ์กฐ์ ํ์ฌ ๋ฏธ์ธ ์กฐ์ (Fine-tuning)ํ๊ฑฐ๋ ์ง์์ ์ธ ํ์ต(Continual Learning)์ ํตํด ํน์ ์์ญ์์ ๋ชจ๋ธ์ ๋ ๊น์ด ๊ฐ์ ํ๋ ๊ณผํ์ ์์ง ์ด๊ธฐ ๋จ๊ณ์ ๋จธ๋ฌผ๋ฌ ์์ต๋๋ค.
๊ฒฐ๋ก : ๋๋์ง ์๋ ๋ฃจํ, ๋ฌดํํ ๊ฐ๋ฅ์ฑ
์๋๋ ์ด ์นดํ์๊ฐ ๋ฌ์ฌํ๋ AI์ ๋ฏธ๋๋ ์ธ๊ฐ์ด ๋ ์ด์ โ๋ณ๋ชฉโ์ด ์๋, ์์ด์ ํธ์๊ฒ โ์์งโ๋ฅผ ํํํ๊ณ ์์คํ ์ ์ค๊ณํ๋ ์๋ก์ด ์ญํ ๋ก ์ ํ๋๋ ์๋์ ๋๋ค. โAI ๊ด์ฆโ์ ๋ฌดํํ ์ต์ ํ์ ๋ฃจํ์ ์์ง ํํ๋์ง ์์ ์ ์ฌ๋ ฅ์ ๋ํ ํฅ๋ถ๊ณผ ์ฝ๊ฐ์ ๋ถ์๊ฐ์ ๋์์ ๋ดํฌํฉ๋๋ค.
์ง๊ธ์ ๋ชจ๋ธ์ โ๋ค์ญ๋ ์ญํจโ๊ณผ โ์ต์ ํ์ ๋งน์ โ ๊ฐ์ ํ๊ณ๊ฐ ์กด์ฌํ์ง๋ง, ์ด๋ฌํ ๋ฌธ์ ์ ๋ค์ ๊ฒฐ๊ตญ โ์คํฌ ๋ฌธ์ โ๋ก ํด๊ฒฐ๋ ๊ฒ์ ๋๋ค. ์ธ๊ฐ์ ์์ด์ ํธ๋ฅผ ๋ ์ ํ์ฉํ๊ณ , ๋ ๋์ โํ๋ก๊ทธ๋จ MDโ๋ฅผ ์์ฑํ๋ฉฐ, ๊ถ๊ทน์ ์ผ๋ก ์ธ๊ฐ์ ๊ฐ์ ์์ด ์ค์ค๋ก ๊ฐ์ ํ๊ณ ์งํํ๋ ์์จ์ ์ธ AI ์์คํ ์ ๊ตฌ์ถํ๋ ๋ฐฉ๋ฒ์ ์ฐพ์๋ผ ๊ฒ์ ๋๋ค. ์ด โ๋ฌดํ ๋ฃจํ์ ์๋(Loopy Era)โ๋ ์์ง๋์ด๋ง, ์ฐ๊ตฌ, ๊ทธ๋ฆฌ๊ณ ์ฐ๋ฆฌ์ ์ผ์๊น์ง๋ ๊ทผ๋ณธ์ ์ผ๋ก ์ฌํธํ ๊ฒ์ด๋ฉฐ, ์ฐ๋ฆฌ๋ ๊ทธ ๋ณํ์ ์์์ ์ ์ ์์ต๋๋ค.
โTerence Tao โ How the worldโs top mathematician uses AIโ โ Dwarkesh Patel ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
AI, ๊ณผํ์ ๋ฐ๊ฒฌ์ ํจ๋ฌ๋ค์์ ๋ฐ๊พธ๋ค: ํ ๋ ์ค ํ์ค์์ ๋ํ
์ธ๊ณ ์ต๊ณ ์ ์ํ์ ์ค ํ ๋ช ์ธ ํ ๋ ์ค ํ์ค(Terence Tao) ๊ต์๋ ์ธ๊ณต์ง๋ฅ(AI)์ด ๊ณผํ์ ๋ฐ๊ฒฌ์ ๋ณธ์ง์ ์ด๋ป๊ฒ ๋ณํ์ํค๊ณ ์๋์ง์ ๋ํ ๊น์ด ์๋ ํต์ฐฐ์ ์ ์ํฉ๋๋ค. ๋์์ผ์ ํํ (Dwarkesh Patel)๊ณผ์ ์ธํฐ๋ทฐ์์ ํ์ค ๊ต์๋ ์ผํ๋ฌ(Kepler)์ ํ์ฑ ์ด๋ ๋ฒ์น ๋ฐ๊ฒฌ ๊ณผ์ ์ ์์์ผ๋ก, AI๊ฐ ์์ด๋์ด ์์ฑ์ ๋น์ฉ์ ๊ฑฐ์ 0์ผ๋ก ๋ฎ์ถ๊ณ ์์ง๋ง, ๋์์ ๊ณผํ๊ณ์ ์๋ก์ด ๋ณ๋ชฉ ํ์, ์ฆ โ๊ฒ์ฆ๊ณผ ํ๊ฐโ๋ผ๋ ๊ฑฐ๋ํ ๊ณผ์ ๋ฅผ ์๊ฒจ์ฃผ๊ณ ์๋ค๊ณ ์ญ์คํฉ๋๋ค. ์ด ๊ธฐ์ฌ๋ ํ์ค ๊ต์์ ์ด์ผ๊ธฐ๋ฅผ ํตํด AI ์๋์ ๊ณผํ์ด ๋์๊ฐ์ผ ํ ๋ฐฉํฅ๊ณผ ๊ทธ ๋ณต์กํ ํจ์๋ฅผ ํ๊ตฌํฉ๋๋ค.
1. ์ผํ๋ฌ์ ๋ธ๋ผํค: ๊ฒฝํ์ ๋ฐ๊ฒฌ์ ์๋ง
ํ์ค ๊ต์๋ ์ธ๊ณต์ง๋ฅ์ด ์ํ์ ๋ฏธ์น ์ํฅ์ ๋ ผํ๊ธฐ ์ํ ํ๋ฅญํ ์ถ๋ฐ์ ์ผ๋ก ์ํ๋ค์ค ์ผํ๋ฌ์ ํ์ฑ ์ด๋ ๋ฒ์น ๋ฐ๊ฒฌ ์ด์ผ๊ธฐ๋ฅผ ๊บผ๋ ๋๋ค. ์ผํ๋ฌ๋ ์ฝํ๋ฅด๋์ฟ ์ค(Copernicus)์ ์ง๋์ค, ์ฆ ํ์์ด ํ์๊ณ์ ์ค์ฌ์ ์๊ณ ํ์ฑ๋ค์ด ๊ทธ ์ฃผ์๋ฅผ ๋๋ค๋ ๋ชจ๋ธ์ ๊ธฐ๋ฐ์ ๋์์ต๋๋ค. ์ฝํ๋ฅด๋์ฟ ์ค๋ ํ์ฑ ๊ถค๋๊ฐ ์๋ฒฝํ ์์ด๋ผ๊ณ ๋ฏฟ์์ง๋ง, ์ผํ๋ฌ๋ ์ด ๊ถค๋์ ๋น์จ์์ ์ด๋ค ๊ธฐํํ์ ์๋ฏธ๋ฅผ ๋ฐ๊ฒฌํ๋ ค ํ์ต๋๋ค. ๊ทธ๋ ์ง๊ตฌ ๊ถค๋๋ฅผ ์ ์ก๋ฉด์ฒด ์์ ๋ฃ์ผ๋ฉด, ๊ทธ ์ ์ก๋ฉด์ฒด๋ฅผ ๋๋ฌ์ธ๋ ์ธ๋ถ ๊ตฌ๊ฐ ํ์ฑ์ ๊ถค๋์ ๊ฑฐ์ ์๋ฒฝํ๊ฒ ์ผ์นํ๋ ๋ฑ, ๋น์ ์๋ ค์ง ์ฌ์ฏ ํ์ฑ์ ๊ถค๋ ์ฌ์ด์ ๋ค์ฏ ๊ฐ์ ํ๋ผํค ์ ์ฒด(Platonic solids: ์ ์ก๋ฉด์ฒด, ์ ์ฌ๋ฉด์ฒด, ์ ์ด์ญ๋ฉด์ฒด, ์ ํ๋ฉด์ฒด, ์ ์ญ์ด๋ฉด์ฒด)๋ฅผ ๋ผ์ ๋ฃ์ ์ ์๋ค๋ ์๋ฆ๋ค์ด ์ด๋ก ์ ์ ์ํ์ต๋๋ค. ์ด๋ ์ ์ ์ค๊ณ๊ฐ ์ํ์ ์๋ฒฝํจ๊ณผ ์ผ์นํ๋ค๊ณ ๋ฏฟ์๊ธฐ ๋๋ฌธ์ ๋๋ค.
์ด ์ด๋ก ์ ํ์ธํ๊ธฐ ์ํด ์ผํ๋ฌ์๊ฒ๋ ๋ฐ์ดํฐ๊ฐ ํ์ํ์ต๋๋ค. ๋น์ ์ ์ผํ๊ฒ ๊ณ ํ์ง ๋ฐ์ดํฐ๋ฅผ ๋ณด์ ํ๊ณ ์๋ ์ธ๋ฌผ์ ๋ด๋งํฌ์ ๋ถ์ ํ๊ณ ๊ดด์ง ์ฒ๋ฌธํ์ ํ์ฝ ๋ธ๋ผํค(Tycho Brahe)์์ต๋๋ค. ๋ธ๋ผํค๋ ์์ญ ๋ ๊ฐ ์ก์์ผ๋ก ํ์ฑ๋ค์ ๊ด์ธกํ๋ฉฐ ๋ฐฉ๋ํ ๋ฐ์ดํฐ๋ฅผ ์ถ์ ํ์ต๋๋ค. ์ผํ๋ฌ๋ ๋ธ๋ผํค์ ํจ๊ป ์ผํ์ง๋ง, ๋ธ๋ผํค๋ ๋ฐ์ดํฐ๋ฅผ ๋งค์ฐ ์งํฌํ์ฌ ์กฐ๊ธ์ฉ๋ง ์ ๊ณตํ์ต๋๋ค. ๊ฒฐ๊ตญ ์ผํ๋ฌ๋ ์ด ๋ฐ์ดํฐ๋ฅผ โํ์น๋ค์ํผโ ํ์ฌ ์ป๊ฒ ๋์๊ณ , ์์ ์ ์๋ฆ๋ค์ด ์ด๋ก ์ด ์ค์ ๋ฐ์ดํฐ์ 10% ์ ๋ ์ค์ฐจ๊ฐ ์์์ ๋ฐ๊ฒฌํ๊ณ ํฌ๊ฒ ์ค๋งํ์ต๋๋ค.
์ผํ๋ฌ๋ ์๋ ๊ฐ ์ด ๋ฌธ์ ์ ๋งค๋ฌ๋ฆฐ ๋์, ๋ฐ์ดํฐ์์ ์ค์ ํ์ฑ ๊ถค๋๋ฅผ ๋์ถํ๋ ๋๋ผ์ด ๋ถ์ ๋ฅ๋ ฅ์ ๋ฐํํ์ต๋๋ค. ๊ทธ ๊ฒฐ๊ณผ ๊ถค๋๊ฐ ์์ด ์๋ ํ์์ด๋ผ๋ ์ถฉ๊ฒฉ์ ์ธ ์ฌ์ค์ ๋ฐํ๋๊ณ , ๋ ๊ฐ์ง ํ์ฑ ์ด๋ ๋ฒ์น(ํ์ ๊ถค๋, ๋ฉด์ ์๋ ์ผ์ ์ ๋ฒ์น)์ ์ ๋ฆฝํ์ต๋๋ค. 10๋ ํ, ๊ทธ๋ ์ธ ๋ฒ์งธ ๋ฒ์น(๊ณต์ ์ฃผ๊ธฐ์ ์ ๊ณฑ์ ๊ถค๋ ์ฅ๋ฐ๊ฒฝ์ ์ธ์ ๊ณฑ์ ๋น๋ก)๊น์ง ๋ฐ๊ฒฌํ์ต๋๋ค. ์ผํ๋ฌ๋ ์ด ๋ฒ์น๋ค์ ๋ํ ๋ฌผ๋ฆฌ์ ์ค๋ช ์ ์ ์ํ์ง ๋ชปํ์ผ๋ฉฐ, ์ด๋ ํ ์ธ๊ธฐ ํ ๋ดํด(Newton)์ด ์ค๋ ฅ ์ด๋ก ์ ํตํด ์ด ๋ชจ๋ ๊ฒ์ ์ค๋ช ํ๋ฉด์ ๋น๋ก์ ์์ฑ๋์์ต๋๋ค.
ํ์ค ๊ต์๋ ์ด ์ด์ผ๊ธฐ์์ ์ค์ํ ํต์ฐฐ์ ์ ์ํฉ๋๋ค. ์ผํ๋ฌ๋ ๋ง์น โ๊ณ ์จ์ ๋๊ท๋ชจ ์ธ์ด ๋ชจ๋ธ(LLM)โ๊ณผ ๊ฐ์๋ค๋ ๊ฒ์ ๋๋ค. ๊ทธ๋ โ์ธ๊ณ์ ์กฐํ(Harmonics of the World)โ๋ผ๋ ์ฑ ์์ ํ์ฑ๋ค์ โ์์ โ์ ๋ ผํ๋ ๋ฑ ๋ฌด์์์ ์ธ ๊ด๊ณ๋ค์ ๋์์์ด ์๋ํ์ต๋๋ค. ์ด ๊ณผ์ ์์ ์ง๊ตฌ์ ์์ ์ โ๋ฏธ-ํ-๋ฏธโ์ด๋ฉฐ, ์ด๊ฒ์ด ์ง๊ตฌ์ ๊ธฐ๊ทผ๊ณผ ๋ถํ์ ์์ธ์ด๋ผ๋ ์์ ํฉ๋นํ ์ ์ฑ์ ์ ์์ด๋์ด๋ ํฌํจ๋์ด ์์์ต๋๋ค. ํ์ง๋ง ๊ทธ ์์์ ์ธ ๋ฒ์งธ ํ์ฑ ์ด๋ ๋ฒ์น์ธ โ์ ๊ณฑ-์ธ์ ๊ณฑ ๋ฒ์น(cube-square law)โ์ด ๋ฐ๊ฒฌ๋์์ต๋๋ค. ๋ธ๋ผํค์ ๋ฐ์ดํฐ์ ๊ฐ์ ๊ฒ์ฆ ๊ฐ๋ฅํ ๋ฐ์ดํฐ ๋ฑ ํฌ๋ง ์๋ค๋ฉด, LLM์ ์ผํ๋ฌ์ฒ๋ผ 20๋ ๋์ ๋ฌด์์์ ์ธ ๊ด๊ณ๋ฅผ ์๋ํ ์ ์๊ณ , ๊ทธ์ค ์ผ๋ถ๋ ๊ณผํ์ ์ง๋ณด๋ก ์ด์ด์ง ์ ์๋ค๋ ๊ฒ์ด์ฃ .
2. ๋ณํํ๋ ๊ณผํ์ ๋ฐฉ๋ฒ๋ก : ๊ฐ์ค์์ ๋ฐ์ดํฐ ์ค์ฌ์ผ๋ก
์ ํต์ ์ผ๋ก ๊ณผํ์ ๋ฌธ์ ์๋ณ, ๊ฐ์ค ์๋ฆฝ, ๋ฐ์ดํฐ ์์ง, ๊ฒ์ฆ, ์ค๋ช ์ ๊ณผ์ ์ ๋ฐ๋ฅด๋ฉฐ, ํนํ โ์ ๋ ์นดโ์ ๊ฐ์ ์ฒ์ฌ์ ์ธ ์์ด๋์ด ์์ฑ์ ์๊ฐ์ ๋์ด ํ๊ฐํ์ต๋๋ค. ์ผํ๋ฌ ์ญ์ ๋ง์ ์์ด๋์ด๋ฅผ ์๋ํ๊ณ , ๊ทธ์ค ์๋น์๋ ์คํจํ์ต๋๋ค. ํ์ง๋ง ๊ทธ์ ์ฑ๊ณต์ ๋ธ๋ผํค์ ์ ๋ฐํ ๋ฐ์ดํฐ ์์ง์ด๋ผ๋ โ๊ฒ์ฆโ ๊ณผ์ ์ด ์์๊ธฐ์ ๊ฐ๋ฅํ์ต๋๋ค. ๋ธ๋ผํค๊ฐ ์ด์ ๊ด์ธก๋ณด๋ค 10๋ฐฐ๋ ๋ ์ ํํ ๋ฐ์ดํฐ๋ฅผ ์ ๊ณตํ๊ธฐ ๋๋ฌธ์ ์ผํ๋ฌ๋ ์์ ์ ๋ชจ๋ธ์ ๋ฐ์ดํฐ์ ๋ง์ถฐ ์กฐ์ ํ๊ณ ์ ํํ ๊ฒฐ๊ณผ๋ฅผ ์ป์ ์ ์์๋ ๊ฒ์ ๋๋ค.
๊ทธ๋ฌ๋ 20์ธ๊ธฐ ๋ค์ด ๊ณผํ์ ํจ๋ฌ๋ค์์ ๋ณํํ์ต๋๋ค. ์ด๋ก ๊ณผ ์คํ์ด๋ผ๋ ๋ ๊ฐ์ง ์ฃผ์ ํจ๋ฌ๋ค์์ ๋ํด, ์์น ์๋ฎฌ๋ ์ด์ (numerical simulation)๊ณผ ๋น ๋ฐ์ดํฐ(big data) ๋ถ์์ด ๋ฑ์ฅํ์ต๋๋ค. ์ด์ ๋ ๋๊ท๋ชจ ๋ฐ์ดํฐ์ ์ ๋จผ์ ์์งํ๊ณ , ๊ทธ ์์์ ํจํด์ ์ฐพ์ ๊ฐ์ค์ ๋์ถํ๋ ๋ฐฉ์์ด ์ ์ฐจ ๋ณดํธํ๋๊ณ ์์ต๋๋ค. ํ์ค ๊ต์๋ ์ด๋ฅผ โ๊ณผํ์ ๋ฐฉ๋ฒ์ด ๊ฑฐ์ ์ญ์ ๋์๋คโ๊ณ ํํํฉ๋๋ค. ์ผํ๋ฌ๋ ์ด์ฉ๋ฉด ์ต์ด์ ๋ฐ์ดํฐ ๊ณผํ์ ์ค ํ ๋ช ์ด์์ง๋ง, ๊ทธ์กฐ์ฐจ๋ ๋ธ๋ผํค์ ๋ฐ์ดํฐ๋ก ์์ํ๊ธฐ ์ ์ ์ด๋ฏธ ์ ์ ๊ฒฌ์ ์ธ ์ด๋ก ์ ๊ฐ์ง๊ณ ์์์ต๋๋ค. ํ์ง๋ง ์ด์ ๋ฐ์ดํฐ์ ๊ท๋ชจ์ ์ ์ฉ์ฑ์ด ์๋์ ์ผ๋ก ์ปค์ง๋ฉด์, ๋ฐ์ดํฐ๋ฅผ ๋จผ์ ๋ถ์ํ๊ณ ๊ฐ์ค์ ์ฐพ๋ ๋ฐฉ์์ด ๋์ฑ ๊ฐ๋ ฅํด์ง๊ณ ์์ต๋๋ค.
๋ค๋ง, ๋ฐ์ดํฐ ๊ธฐ๋ฐ ์ ๊ทผ๋ฒ์๋ ํจ์ ์ ์์ต๋๋ค. ์ํ ๋ณด๋ฐ(Johann Bode)์ ๋ฒ์น์ด ์ข์ ์์ ๋๋ค. ๊ทธ๋ ์ผํ๋ฌ์๊ฒ ์๊ฐ์ ๋ฐ์ ํ์ฑ ๊ฐ ๊ฑฐ๋ฆฌ๊ฐ ๊ธฐํ๊ธ์์ ์ผ๋ก ์ฆ๊ฐํ๋ค๋ ํจํด์ ๋ฐ๊ฒฌํ๊ณ , ํ์ฑ๊ณผ ๋ชฉ์ฑ ์ฌ์ด์ ํ์ฑ์ด ํ๋ ๋ ์์ ๊ฒ์ด๋ผ๊ณ ์์ธกํ์ต๋๋ค. ์ค์ ๋ก ์ฒ์์ฑ(Uranus)๊ณผ ์ํ์ฑ ์ธ๋ ์ค(Ceres)๊ฐ ๋ฐ๊ฒฌ๋๋ฉด์ ๊ทธ์ ์์ธก๊ณผ ์ผ์นํ๋ ๋ฏ ๋ณด์์ต๋๋ค. ๊ทธ๋ฌ๋ ํด์์ฑ(Neptune)์ด ๋ฐ๊ฒฌ๋๋ฉด์ ์ด ํจํด์ ์์ ํ ์ด๊ธ๋ฌ๊ณ , ๊ฒฐ๊ตญ ๋ณด๋ฐ์ ๋ฒ์น์ โ์ซ์์ ์ฐ์ฐ์ ์ผ์น(numerical fluke)โ์ ๋ถ๊ณผํ์์ด ๋๋ฌ๋ฌ์ต๋๋ค. ์ผํ๋ฌ๊ฐ ์ธ ๋ฒ์งธ ๋ฒ์น์ ์ฒ์ ๋ ๋ฒ์น๋งํผ ๊ฐ์กฐํ์ง ์์ ๊ฒ๋, ๋จ 6๊ฐ์ ๋ฐ์ดํฐ ํฌ์ธํธ๋ง์ผ๋ก ๋์ถ๋ ๊ฒฐ๋ก ์ ์ ๋ขฐ์ฑ์ ๋ํด ๋ณธ๋ฅ์ ์ผ๋ก ์กฐ์ฌ์ค๋ฌ์ ๊ธฐ ๋๋ฌธ์ผ ์ ์์ต๋๋ค.
3. AI๊ฐ ์ด๋ฐํ ์๋ก์ด ๋ณ๋ชฉ ํ์: ์์ด๋์ด ์์ฑ์์ ๊ฒ์ฆ์ผ๋ก
ํ์ค ๊ต์๋ AI๊ฐ ์์ด๋์ด ์์ฑ์ ๋น์ฉ์ ๊ฑฐ์ 0์ ๊ฐ๊น๊ฒ ๋ฎ์ท๋ค๊ณ ๊ฐ์กฐํฉ๋๋ค. ์ด๋ ์ธํฐ๋ท์ด ํต์ ๋น์ฉ์ ๋ฎ์ถ ๊ฒ๊ณผ ์ ์ฌํฉ๋๋ค. ํ์ง๋ง ์ด๋ ๊ณง ๊ณผํ์ ์๋ก์ด ๋ณ๋ชฉ ํ์(bottleneck)์ ์ด๋ํฉ๋๋ค. ๊ณผ๊ฑฐ์๋ ์๋ง์ถ์ด ๊ณผํ์๋ค์ด ๋ฌด์ํ ๋ง์ ์ด๋ก ์ ์ ์ํ์ง๋ง, ๊ทธ์ค ๋๋ถ๋ถ์ ๊ฐ์น๊ฐ ์์์ต๋๋ค. ๊ทธ๋์ ๋๋ฃ ์ฌ์ฌ(peer review)์ ์ถํ ์์คํ ์ ๊ณ ํ์ง ์์ด๋์ด๋ฅผ ๊ฑธ๋ฌ๋ด๊ณ ๊ฒ์ฆํ๋ ์ญํ ์ ํ์ต๋๋ค. ํ์ง๋ง ์ด์ AI๊ฐ ์์ฒญ๋ ๊ท๋ชจ๋ก ๊ฐ๋ฅํ ์ค๋ช ์ ์์ฑํด๋ด๋ฉด์, ์ธ๊ฐ ๊ฒํ ์๋ค์ ์ด๋ฏธ ๊ฐ๋นํ๊ธฐ ํ๋ ์ํฉ์ ์ฒํด ์์ต๋๋ค. ๋ง์ ํ์ ์ง๋ค์ด AI๊ฐ ์์ฑํ ๋ ผ๋ฌธ์ผ๋ก ๋์ณ๋๊ณ ์๋ค๊ณ ๋ณด๊ณ ํฉ๋๋ค.
๋ฌธ์ ๋ ์ด ๋ฐฉ๋ํ ์์ด๋์ด ํ์์ ์ด๋ค ์์ด๋์ด๊ฐ ์ง์ง ์ง์ ์ ์ด๋ฃจ๊ณ , ์ด๋ค ์์ด๋์ด๊ฐ ๋ง๋ค๋ฅธ ๊ณจ๋ชฉ์ด๋ ์๋ชป๋ ๋จ์(red herrings)์ธ์ง๋ฅผ ํ๊ฐํ๊ณ ๊ฒ์ฆํ๋ ๋ฐฉ๋ฒ์ ๋๊ท๋ชจ๋ก ์์ง ๋ชปํ๋ค๋ ๊ฒ์ ๋๋ค. ๊ฐ๋ณ ๋ ผ๋ฌธ์ ๋ํด์๋ ์๋ ๊ฐ์ ๋ ผ์๋ฅผ ํตํด ํฉ์์ ๋๋ฌํ ์ ์์ง๋ง, ๋งค์ผ ์์ฒ ๊ฐ์ ์์ด๋์ด๊ฐ ์์ฑ๋ ๋ ์ด๋ฐ ๋ฐฉ์์ ํตํ์ง ์์ต๋๋ค.
ํ์ค ๊ต์๋ ๋ฒจ ์ฐ๊ตฌ์(Bell Labs)์์ โ๋นํธ(bit)โ ๊ฐ๋ ์ด ๋ฑ์ฅํ์ ๋๋ฅผ ์๋ก ๋ญ๋๋ค. ์๋ง์ ๊ธฐ์ ๋ ผ๋ฌธ ์์์ โ๋นํธโ๋ผ๋ ์์ด๋์ด๋ ํ๋ฅ ๋ก , ์ปดํจํฐ ๊ณผํ ๋ฑ ๋ค์ํ ๋ถ์ผ์ ๊ฑธ์ณ ์์ฒญ๋ ํจ์๋ฅผ ๊ฐ์ก์ต๋๋ค. ๋ฏธ๋์ AI๊ฐ ์ด์ ๊ฐ์ ํต์ผ ๊ฐ๋ (unifying concept)์ ์ ์ํ๋ค๋ฉด, ์๋ฐฑ๋ง ๊ฐ์ ๋ ผ๋ฌธ ์ค์์ ์ด๋ป๊ฒ ๊ทธ๊ฒ์ ์๋ณํด๋ผ ์ ์์๊น์?
์์ด๋์ด์ ๊ฐ์น๋ฅผ ํ๊ฐํ๋ ๊ฒ์ ๋งค์ฐ ์ด๋ ต์ต๋๋ค. ๋ฅ๋ฌ๋(Deep Learning) ์์ฒด๋ ์ค๋ซ๋์ AI์ ํ์ ๋ถ์ผ์๊ณ , ์ฒซ ๋ฒ์งธ ์๋ฆฌ(first principles) ์ถ๋ก ์ด ์๋ ๋ฐ์ดํฐ ํ์ต๋ง์ผ๋ก ๋ต์ ์ป๋๋ค๋ ์๊ฐ์ ๋งค์ฐ ๋ ผ๋์ด ๋ง์์ต๋๋ค. โํธ๋์คํฌ๋จธ(Transformer)โ ์ํคํ ์ฒ๋ ํ๋ ๋๊ท๋ชจ ์ธ์ด ๋ชจ๋ธ์ ๊ธฐ๋ฐ์ด ๋์์ง๋ง, ๋ค๋ฅธ ์ํคํ ์ฒ๊ฐ ๋จผ์ ์ฑํ๋์๋ค๋ฉด ๊ทธ๊ฒ์ด ํ์ค์ด ๋์์ ์๋ ์์ต๋๋ค. ์์ด๋์ด์ ๊ฒฐ์ค ์ฌ๋ถ๋ ๋ฏธ๋์ ๋ฌธํ, ์ฌํ์ ์ฑํ ์ฌ๋ถ์ ๋ฌ๋ ค ์์ต๋๋ค. 10์ง๋ฒ ์ฒด๊ณ๊ฐ ๋ก๋ง ์ซ์๋ณด๋ค ํจ์ฌ ์ ์ฉํ์ง๋ง, 10์ด๋ผ๋ ์ซ์์ ํน๋ณํ ์๋ฏธ๊ฐ ์๋ ๊ฒ์ ์๋๋๋ค. ๋จ์ง ๋ชจ๋ ์ฌ๋์ด ์ฌ์ฉํ๊ธฐ ๋๋ฌธ์ ํ์คํ๋์๊ณ , ์ด์ ๋ ๊ด์ฑ ๋๋ฌธ์ ๋ฐ๊พธ๊ธฐ ํ๋ญ๋๋ค. ์ฆ, ๊ณผํ์ ์ฑ๊ณผ๋ฅผ ๊ณผ๊ฑฐ์ ๋ฏธ๋์ ๋งฅ๋ฝ ์์ด ์์ํ๊ฒ ๊ณ ๋ฆฝ์์ผ ๊ฐ๊ด์ ์ผ๋ก ํ๊ฐํ๋ ๊ฒ์ ๋ถ๊ฐ๋ฅํฉ๋๋ค.
๋ํ, ๊ณผํ์ ์ง๋ณด๋ ์ข ์ข ๊ธฐ์กด์ ๊ฐ์ ์ ์ญ์ ํ๋ ๋ฐฉ์์ผ๋ก ์ด๋ฃจ์ด์ง๋๋ค. ์ง๋์ค์ด ์ค๋ซ๋์ ๋ฐ์๋ค์ฌ์ง์ง ์์ ๊ฒ์ ์๋ฆฌ์คํ ํ ๋ ์ค(Aristotle)์ ๋ฌผ๋ฆฌํ, ์ฆ ๋ฌผ์ฒด๊ฐ ์์ฐ์ ์ผ๋ก ์ ์ง ์ํ๋ฅผ ์ ์งํ๋ ค ํ๋ค๋ ์๊ฐ ๋๋ฌธ์ด์์ต๋๋ค. ์ง๊ตฌ๊ฐ ์์ง์ธ๋ค๋ฉด ์ ์ฐ๋ฆฌ๊ฐ ๋ชจ๋ ๋จ์ด์ง์ง ์๋๊ฐ ํ๋ ์๋ฌธ์ด ํด๊ฒฐ๋์ง ์์๊ธฐ ๋๋ฌธ์ ๋๋ค. ๋ดํด์ ์ด๋ ๋ฒ์น์ด ์ ๋ฆฝ๋ ํ์์ผ ๋น๋ก์ ์ง๊ตฌ๊ฐ ์์ง์ธ๋ค๋ ๊ฐ๋ ์ด ์ดํด๋ ์ ์์์ต๋๋ค. ๋ค์(Darwin)์ ์งํ๋ก ์ญ์ ์ข (species)์ด ๊ณ ์ ๋์ด ์์ง ์๋ค๋ ๋น์๋ก์๋ ํ๊ฒฉ์ ์ธ ๊ฐ์ ์ ์ญ์ ํ๋ฉด์ ์ด๋ฃจ์ด์ง ์๋ํ ์ง๋ณด์ ๋๋ค. ํ์ค ๊ต์๋ ํ์ฌ ์ฐ๋ฆฌ๊ฐ โ์ธ๊ฐ ์ง๋ฅ์ด ์ฐ์ฃผ์ ์ค์ฌโ์ด๋ผ๋ ์๊ฐ์์ ๋ฒ์ด๋ ๋ค์ํ ์ ํ์ ์ง๋ฅ์ด ์กด์ฌํจ์ ์ธ์ํ๋ โ์ธ์ง์ ์ฝํ๋ฅด๋์ฟ ์ค์ ํ๋ช (cognitive Copernican revolution)โ์ ๊ฒช๊ณ ์๋ค๊ณ ๋งํฉ๋๋ค.
4. ๊ณผํ์ ์ํต๊ณผ ์ค๋์ ์ค์์ฑ
ํ์ค ๊ต์๋ ๊ณผํ์ ๋ฐ๊ฒฌ์ด ๋จ์ํ ์๋ก์ด ์ด๋ก ์ ๋ง๋ค๊ณ ๊ฒ์ฆํ๋ ๊ฒ์ ๋์ด, ๋ค๋ฅธ ์ฌ๋๋ค์๊ฒ ํจ๊ณผ์ ์ผ๋ก ์ํตํ๋ ๊ฒ ๋ํ ๋งค์ฐ ์ค์ํ๋ค๊ณ ๊ฐ์กฐํฉ๋๋ค. ๊ทธ๋ ๋ค์์ ์งํ๋ก ์ด ๋ดํด์ โํ๋ฆฐํคํผ์(Principia Mathematica)โ๋ณด๋ค 2์ธ๊ธฐ ํ์ ๋์์์๋ ๋ถ๊ตฌํ๊ณ , ๊ฐ๋ ์ ์ผ๋ก๋ ๋ ๊ฐ๋จํด ๋ณด์๋ ์ด์ ๋ฅผ ์ค๋ช ํฉ๋๋ค. ๋ค์์ ์ด๋ก ์ ์์ฐ์ ํ์ ๋ํ ์ฆ๊ฑฐ๊ฐ ์๋์ ์ด์์ง๋ง, ์ด๋ ๋์ ์ ์ด๊ณ ํ๊ณ ์ ์ธ(cumulative and retrospective) ์ฑ๊ฒฉ์ ๋ ์์ต๋๋ค. ๋ฐ๋ฉด ๋ดํด์ ์์ ์ ๋ฐฉ์ ์์ ์ ์ํ๊ณ ๋ฌ์ ๊ถค๋ ์ฃผ๊ธฐ์ ๊ฑฐ๋ฆฌ๊ฐ ์ผ์นํ๋์ง ํ์ธํ๋ฉด ๋ฐ๋ก ์ง์ ์ด ์ด๋ฃจ์ด์ก์ต๋๋ค.
๋ค์์ ๋๋ผ์ด ๊ณผํ ์ปค๋ฎค๋์ผ์ดํฐ์์ต๋๋ค. ๊ทธ๋ ๋ฐฉ์ ์ ์์ด ํ์ดํ ์์ด๋ก ๊ธ์ ์ผ๊ณ , ๋ค์ํ ์ฌ์ค๋ค์ ์ค๋๋ ฅ ์๋ ๋น์ ์ผ๋ก ์ข ํฉํ์ต๋๋ค. ๋ฐ๋ฉด ๋ดํด์ ๋ผํด์ด๋ก ๊ธ์ ์ผ๊ณ , ์์ ์ ์ด๋ก ์ ์ค๋ช ํ๊ธฐ ์ํด ์๋ก์ด ์ํ ๋ถ์ผ(๋ฏธ์ ๋ถ)๋ฅผ ์ฐฝ์ํด์ผ ํ์ต๋๋ค. ๊ทธ๋ ๋ํ ๋น์ ๊ณผํ์๋ค์ด ํจ์ฌ ๋น๋ฐ์ค๋ฝ๊ณ ๊ฒฝ์์ ์ด์๋ ์๋์ ์ธ๋ฌผ์ด์์ผ๋ฉฐ, ์์ ์ ๊ฐ์ฅ ์ค์ํ ํต์ฐฐ ์ผ๋ถ๋ฅผ ๊ฒฝ์์๋ค์๊ฒ ๋นผ์๊ธฐ์ง ์๊ธฐ ์ํด ์จ๊ธฐ๊ธฐ๋ ํ์ต๋๋ค. ๋ดํด์ ์ ์ ์ด ๋๋ฆฌ ์๋ ค์ง ๊ฒ์ ์์ญ ๋ ํ ๋ค๋ฅธ ๊ณผํ์๋ค์ด ๊ทธ์ ์์ ์ ํจ์ฌ ๊ฐ๋จํ ์ฉ์ด๋ก ์ค๋ช ํ ์ดํ์์ต๋๋ค.
ํ์ค ๊ต์๋ โ์์ ์ ๊ธฐ์ ๊ณผ ์ฃผ์ฅ์ ํผ์น๋ ๋ฅ๋ ฅ, ๊ทธ๋ฆฌ๊ณ ์์ฌ๋ฅผ ๋ง๋๋ ๊ฒโ์ด ๊ณผํ์ ๋งค์ฐ ์ค์ํ ๋ถ๋ถ์ด๋ผ๊ณ ๋งํฉ๋๋ค. ๋ฐ์ดํฐ๊ฐ ๋ท๋ฐ์นจ๋๋ฉด ๋์์ด ๋์ง๋ง, ์ฌ๋๋ค์ด ์ค๋๋์ด์ผ๋ง ์ด๋ก ์ ๋ ๋ฐ์ ์ํค๊ฑฐ๋ ์ด๊ธฐ ํฌ์๋ฅผ ํ์ฌ ๋ฐฐ์ฐ๊ณ ํ๊ตฌํ ๊ฒ์ ๋๋ค. ๊ทธ๋ AI๊ฐ ์์ง ์ค๋๋ ฅ์ ๊ฐํํ์ต(Reinforcement Learning)ํ๊ธฐ ์ด๋ ต๋ค๋ ์ ์ ์ง์ ํ๋ฉฐ, ์ด๊ฒ์ด ์ด์ฉ๋ฉด ๊ณผํ์ ์์ํ ์ธ๊ฐ์ ์ธก๋ฉด์ผ๋ก ๋จ์ ์ ์๋ค๊ณ ์์ธกํฉ๋๋ค.
5. ์ํ ๋ถ์ผ AI์ ํ์ฌ์ ๋ฏธ๋: ํญ๊ณผ ๊น์ด
ํ์ค ๊ต์๋ ๋์์ผ์ ํํ ์ ์๋ฆฌ์ฆ โ์ฐ์ฃผ ๊ฑฐ๋ฆฌ ์ฌ๋ค๋ฆฌ(cosmic distance ladder)โ๋ฅผ ์ธ๊ธํ๋ฉฐ, โ๋ง์ ๋ถ์ผ์์ ์ฐ์ญ์ ์ค๋ฒํ(deductive overhang)์ด ์ฌ๋๋ค์ด ๊นจ๋ซ๋ ๊ฒ๋ณด๋ค ํจ์ฌ ํด ์ ์๋คโ๋ ์ ์ ๋์ํฉ๋๋ค. ์ฆ, ๋ฌธ์ ๋ฅผ ์ฐ๊ตฌํ๋ ์ฌ๋ฐ๋ฅธ ํต์ฐฐ๋ ฅ๋ง ์๋ค๋ฉด, ์ธ์์ ๋ํด ํจ์ฌ ๋ ๋ง์ ๊ฒ์ ์ ์ ์๋ค๋ ๊ฒ์ ๋๋ค. ์ฒ๋ฌธํ์ ๋ฐ์ดํฐ๊ฐ ๋ณ๋ชฉ ํ์์ด์๊ธฐ ๋๋ฌธ์ ๋ฐ์ดํฐ๋ฅผ ๋ถ์ํ๊ณ ์ ๋ณด์์ ๋ง์ง๋ง ํ ๋ฐฉ์ธ๊น์ง ์ง๋ด๋ ๋ฐ ๊ฐ์ฅ ๋จผ์ ์์ฅ์ ๊ณผํ ๋ถ์ผ ์ค ํ๋์์ต๋๋ค. ํํธ ํค์งํ๋(quant hedge funds)์์ ์ฒ๋ฌธํ ๋ฐ์ฌ๋ฅผ ์ ํธํ๋ ์ด์ ๋ ์ฌ๊ธฐ์ ์์ต๋๋ค.
์ต๊ทผ AI๋ ์ํ ๋ถ์ผ์์ ๋๋ผ์ด ์ง์ ์ ๋ณด์ด๊ณ ์์ต๋๋ค. ํ์ค ๊ต์๋ ์ต๊ทผ ๋ช ๋ฌ ๋์ AI ํ๋ก๊ทธ๋จ๋ค์ด ์๋ฅด๋์ ๋ฌธ์ (Erdลs problems) 1,100์ฌ ๊ฐ ์ค ์ฝ 50๊ฐ๋ฅผ ํด๊ฒฐํ๋ค๊ณ ์ธ๊ธํ์ต๋๋ค. ํ์ง๋ง ๊ทธ๋ โํ์ฌ ๋ฎ๊ฒ ๋งค๋ฌ๋ฆฐ ์ด๋งค(low-hanging fruit)๋ ์ด๋ฏธ ๋ค ๋ฐ๋ฒ๋ฆฐ ์ํโ์ด๋ฉฐ, ์์ AI ์๋ฃจ์ ์ด ๊ธ๊ฒฉํ ์ค์ด๋ค๊ณ ์๋ค๊ณ ๋งํฉ๋๋ค. ์ด์ ๋ AI๊ฐ ๊ฐ๋ฅํ ์ฆ๋ช ์ ๋ต์ ์์ฑํ๊ณ , ๋ค๋ฅธ AI ๋๊ตฌ๊ฐ ์ด๋ฅผ ๋นํํ๊ณ , ์ฌ์์ฑํ๊ฑฐ๋, ์์น ๋ฐ์ดํฐ๋ฅผ ์์ฑํ๊ณ , ๋ฌธํ ์กฐ์ฌ๋ฅผ ํ๋ ๋ฑ, โ๋ง์ ์ธ๊ฐ๊ณผ ๋ง์ AI ๋๊ตฌ๋ค ๊ฐ์ ์ง์์ ์ธ ๋ํโ๋ฅผ ํตํด ๋ฌธ์ ๊ฐ ํด๊ฒฐ๋๋ ์์์ ๋๋ค.
ํ์ค ๊ต์๋ AI ๋๊ตฌ๋ฅผ โ2๋ฏธํฐ๋ฅผ ์ ํํ ์ ์๋ ์ ํ ๋จธ์ โ์ ๋น์ ํฉ๋๋ค. ์ด๋ค์ ๋๋๋ก ์๋ชป๋ ๋ฐฉํฅ์ผ๋ก ์ ํํ๊ฑฐ๋ ์ถ๋ฝํ๊ธฐ๋ ํ์ง๋ง, ๋๋ก๋ ์ธ๊ฐ์ด ๋๋ฌํ ์ ์์๋ ๊ฐ์ฅ ๋ฎ์ ๋ฒฝ์ ๊ผญ๋๊ธฐ์ ๋๋ฌํ ์ ์์ต๋๋ค. ํ์ง๋ง ์ด ๋๊ตฌ๋ค์ ๋ถ๋ถ์ ์ธ ์ง์ ์ ๋ง๋ค๊ฑฐ๋, ๋จผ์ ์ง์คํด์ผ ํ ์ค๊ฐ ๋จ๊ณ๋ฅผ ์๋ณํ๋ ๋ฐ๋ ๋งค์ฐ ์ํด๋ค๋ ํ๊ณ๊ฐ ์์ต๋๋ค.
๊ทธ๋ AI์ ์ง๋ณด์ ๋ํด โ๋น๊ด๋ก ์ (bearish)์ด๋ฉด์ ๋์์ ๋๊ด๋ก ์ (bullish)โ์ธ ๋ ๊ฐ์ง ๊ด์ ์ ์ ์ํฉ๋๋ค. ๋น๊ด๋ก ์ ๊ด์ ์ AI๊ฐ ์์ง ์ธ๊ฐ ์ ๋ฌธ๊ฐ๊ฐ ๋๋ฌํ๋ ์์ค์ โ๋์ ๋ฒฝโ์ ๋๋ฌํ์ง ๋ชปํ๋ค๋ ๊ฒ์ ๋๋ค. ํ์ง๋ง ๋๊ด๋ก ์ ๊ด์ ์ AI๊ฐ ์ผ๋จ ํน์ ์์ค์ ๋๋ฌํ๋ฉด, โ๊ทธ ์์ค์์ ๊ฐ๋ฅํ ๋ชจ๋ ๋ฌธ์ ๋ฅผ ํด๊ฒฐํ ์ ์๋ ๊ฐ๋ ฅํ ์์ฑโ์ ๊ฐ์ง๊ณ ์๋ค๋ ๊ฒ์ ๋๋ค. ์ธ๊ฐ์ ์๋ฐฑ๋ง ๊ฐ์ ๋ฌธ์ ์ ๋ํด ์๋ฐฑ ๋ ์ ์ฐ๊ตฌ๋ฅผ ๋์์ ์ํํ ์ ์์ง๋ง, AI๋ ๊ฐ๋ฅํ๋ค๋ ๊ฒ์ด์ฃ .
ํ์ค ๊ต์๋ AI๊ฐ **โํญ(breadth)โ**์ ๋ฐ์ด๋๊ณ ์ธ๊ฐ์ **โ๊น์ด(depth)โ**์ ๋ฐ์ด๋๋ค๊ณ ๊ฒฐ๋ก ๋ด๋ฆฝ๋๋ค. ๊ทธ๋ AI์ ํญ๋์ ๋ฅ๋ ฅ์ ์ต๋ํ ํ์ฉํ๊ธฐ ์ํด ๊ณผํ ์ํ ๋ฐฉ์์ ์ฌ์ค๊ณํด์ผ ํ๋ค๊ณ ์ฃผ์ฅํฉ๋๋ค. โํ๋ ๋๋ ๋ ๊ฐ์ ์ ๋ง ๊น๊ณ ์ค์ํ ๋ฌธ์ ๋ณด๋ค๋ ๋งค์ฐ ๊ด๋ฒ์ํ ๋ฌธ์ ๋ค์ ๋ค๋ฃจ๋ ๋ฐ ๋ ๋ง์ ๋ ธ๋ ฅ์ ๊ธฐ์ธ์ฌ์ผ ํ๋คโ๋ ๊ฒ์ ๋๋ค. AI๋ ์๋ก์ด ๊ณผํ ๋ถ์ผ๋ฅผ ํ์ํ๊ณ , ์ฌ์ด ๊ด์ธก์ ๋ชจ๋ ์ํํ์ฌ โ๋์ด๋์ ์ฌโ์ ์๋ณํ ๋ค์, ์ธ๊ฐ ์ ๋ฌธ๊ฐ๊ฐ ๊ทธ๊ณณ์ ์์ ์์ ํ ์ ์๋๋ก ํ ๊ฒ์ ๋๋ค. ๊ทธ๋ โ๋ฏธ๋์๋ ๋งค์ฐ ์ํธ ๋ณด์์ ์ธ ๊ณผํ์ ๋ณด๊ฒ ๋ ๊ฒโ์ด๋ฉฐ, โ๊ณผํ์ ๊ทธ ์ดํ๋ก ์์๋ณผ ์ ์๊ฒ ๋ณํ ๊ฒโ์ด๋ผ๊ณ ์ ๋งํฉ๋๋ค.
ํ๋ก๊ทธ๋๋จธ๋ค์ด AI ๋๊ตฌ ๋๋ถ์ ์์ฐ์ฑ์ด ํฌ๊ฒ ํฅ์๋์๋ค๊ณ ๋๋ผ๋ ๊ฒ์ฒ๋ผ, ์ํ์์๋ ์ ์ฌํ ๋ณํ๊ฐ ์ผ์ด๋ ์ ์์ต๋๋ค. ํ์ค ๊ต์๋ ์ํ์์ ๋ฌธ์ ํด๊ฒฐ ๊ณผ์ ์์ฒด๊ฐ ๋ฌธ์ ์์ฒด๋ณด๋ค ๋ ์ค์ํ๋ค๊ณ ๋งํฉ๋๋ค. AI๋ ์์ฉ๊ตฌ(boilerplate) ์ฝ๋์ ๊ฐ์ ๋ณด์กฐ์ ์ธ ์์ ๋ค์ ๊ฐ์ํํ์ฌ ๊ทธ์ ๋ ผ๋ฌธ์ ๋ ๋ง์ ์ฝ๋๋ ๊ทธ๋ฆผ์ ํฌํจํ ์ ์๊ฒ ํด์ฃผ์์ต๋๋ค. ๊ทธ๋ AI๊ฐ ์๋ค๋ฉด ์ค๋๋ ๊ทธ๊ฐ ์ฐ๋ ๋ ผ๋ฌธ์ ์ ํ์ 5๋ฐฐ ๋ ์ค๋ ๊ฑธ๋ฆด ๊ฒ์ด๋ผ๊ณ ๋งํฉ๋๋ค. ํ์ง๋ง ์ํ ๋ฌธ์ ์ ๊ฐ์ฅ ์ด๋ ค์ด ํต์ฌ ๋ถ๋ถ์ ํด๊ฒฐํ๋ ๋ฐ๋ ์์ง AI์ ๋์์ ๋ฐ์ง ์๊ณ ์ฌ์ ํ ํ๊ณผ ์ข ์ด๋ฅผ ์ฌ์ฉํฉ๋๋ค.
๊ทธ๋ AI ์ ํ ๋๊ตฌ๋ค์ด โ์ํ์ ์คํ์ ์ธก๋ฉด์ ํ์ ํ ๊ฒโ์ด๋ผ๊ณ ์์ธกํฉ๋๋ค. ๊ฐ๋ณ ๋ฌธ์ ์ ํด๊ฒฐ ๊ณผ์ ์ ๋ ์ ๊ฒฝ ์ฐ๊ณ , ๋ฌด์์ด ํจ๊ณผ๊ฐ ์๊ณ ๋ฌด์์ด ํจ๊ณผ๊ฐ ์๋์ง์ ๋ํ ๋๊ท๋ชจ ๋ฐ์ดํฐ๋ฅผ ์์งํ๋ ๋ฐ ์ง์คํ ๊ฒ์ด๋ผ๋ ๊ฒ๋๋ค. โ์ํ์ ๋๊ท๋ชจ๋ก ์ํํ๋ ์์ด๋์ด(the idea of doing mathematics at scale)โ๋ ์์ง ์ด๊ธฐ ๋จ๊ณ์ด์ง๋ง, ์ด๊ฒ์ด AI๊ฐ ์ด ๋ถ์ผ๋ฅผ ํ์ ํ ์ง์ ์ด๋ผ๊ณ ๊ทธ๋ ๋งํฉ๋๋ค.
ํ์ค ๊ต์๋ AI์ ์ง๋ณด๋ฅผ โ๋๋๊ณ ๋ ์ค๋ง์ค๋ฌ์ด(amazing and disappointing)โ ๊ฐ์ ์ผ๋ก ๋ฐ๋ผ๋ด ๋๋ค. ์ฌ๋๋ค์ ๋น ๋ฅด๊ฒ ์ ์ํฉ๋๋ค. 20๋ ์ ๊ตฌ๊ธ ๊ฒ์์ด ์ฒ์ ๋์์ ๋ ๋๋ผ์ ์ง๋ง, ๋ช ๋ ํ์๋ ๋น์ฐํ๊ฒ ์ฌ๊ฒจ์ก์ต๋๋ค. 2026๋ ์์ค์ AI๋ 2021๋ ์๋ ๊ฒฝ์ด๋ก์ ์ ๊ฒ์ด์ง๋ง, ์ด์ ๋ ์ผ๊ตด ์ธ์, ์์ฐ์ด ์ฒ๋ฆฌ, ๋ํ ์์ค์ ์ํ ๋ฌธ์ ํด๊ฒฐ ๋ฑ์ด ๋น์ฐํ๊ฒ ์ฌ๊ฒจ์ง๋ค๋ ๊ฒ์ ๋๋ค.
๊ทธ๋ 2023๋ ์ 2026๋ ๊น์ง AI๊ฐ ์ํ ๋ถ์ผ์์ โ์ ๋ขฐํ ์ ์๋ ๊ณต๋ ์ ์(trustworthy co-author) ๊ฐ์ ์กด์ฌ๊ฐ ๋ ๊ฒโ์ด๋ผ๊ณ ์์ธกํ๊ณ , ํ์ฌ๋ก์๋ ์ด ์์ธก์ด ๊ฝค ์ ๋ง์๊ฐ๊ณ ์๋ค๊ณ ํ๊ฐํฉ๋๋ค.
๋ง์ง๋ง์ผ๋ก ํ์ค ๊ต์๋ โ์ธ๊ณต์ ์ธ ์๋ฆฌํจ(artificial cleverness)โ๊ณผ โ์ธ๊ณต ์ง๋ฅ(artificial intelligence)โ์ ๊ตฌ๋ถํฉ๋๋ค. ์ง๋ฅ์ ์ ์ํ๊ธฐ ์ด๋ ต์ง๋ง, ๊ทธ๊ฐ ๋งํ๋ ์ง๋ฅ์ โ๋ฌธ์ ํด๊ฒฐ์ ์ํด ํ๋ ฅํ ๋ ์์ด๋์ด๋ฅผ ๋์์์ด ๊ฐ์ ํ๊ณ , ์ฒด๊ณ์ ์ผ๋ก ๋ฌด์์ด ์๋ํ๊ณ ๋ฌด์์ด ์๋ํ์ง ์๋์ง ํ์ ํ๋ฉฐ, ์ง์ ์ ๊ธธ์ ์ฐพ์๋ด๋ ์ ์์ฑ(adaptivity)๊ณผ ์ง์์ ์ธ ๊ฐ์ โ์ ์๋ฏธํฉ๋๋ค. AI๋ ์ด๋ฌํ ๊ณผ์ ์ ๋ชจ๋ฐฉํ ์ ์์ง๋ง, โ์กฐ๊ธ์ฉ ์ ํํด์ ์์ก์ด๋ฅผ ์ก๊ณ , ๊ฑฐ๊ธฐ์ ๋จธ๋ฌผ๋ฌ ๋ค๋ฅธ ์ฌ๋๋ค์ ๋์ด์ฌ๋ฆฌ๊ณ , ๊ฑฐ๊ธฐ์๋ถํฐ ๋ค์ ์ ํํ๋โ ๋์ ์ ์ธ ๊ณผ์ ์ ์์ง ๋ถ์กฑํฉ๋๋ค. AI๋ ์ํ์ฐฉ์ค์ ๋ฐ๋ณต, ์ฆ โ๋ฌด์ฐจ๋ณ ๋์ (brute force)โ์ ๊ฐ๊น์ต๋๋ค. AI๊ฐ ๋ฌธ์ ๋ฅผ ํด๊ฒฐํ๋๋ผ๋, ๊ทธ ์์ฒด์ ์ํ์ ์ดํด๊ฐ ์ง์ ๋ ๊ฒ์ ์๋๋ฉฐ, ์๋ก์ด ์ธ๋์ ํ๋ จ ๋ฐ์ดํฐ์ 0.001%์ ๊ธฐ์ฌ๋ฅผ ํ ๋ฟ, ์๋ก์ด ์ธ์ ์์๋ ๋ฐฉ๊ธ ํ ์ผ์ ์์ด๋ฒ๋ฆฐ๋ค๋ ๊ฒ์ ๋๋ค.
๊ฒฐ๋ก
ํ ๋ ์ค ํ์ค ๊ต์์ ํต์ฐฐ์ AI๊ฐ ๊ณผํ์ ๋ฐ๊ฒฌ์ ๋ณธ์ง์ ๊ทผ๋ณธ์ ์ผ๋ก ๋คํ๋ค๊ณ ์์์ ๋ณด์ฌ์ค๋๋ค. ์์ด๋์ด ์์ฑ์ ๋ฌธํฑ์ ๋ฎ์ถ๊ณ โํญโ์ ํ์์ ๊ฐ๋ฅํ๊ฒ ํ๋ AI๋ ๊ณผํ์ ์๋ก์ด ์งํ์ ์ด๊ณ ์์ง๋ง, ๋์์ ๊ฒ์ฆ, ํ๊ฐ, ๊ทธ๋ฆฌ๊ณ ์ง์ ํ ํต์ผ ๊ฐ๋ ์ ์๋ณํ๋ ์๋ก์ด ๊ณผ์ ๋ฅผ ์๊ฒจ์ฃผ๊ณ ์์ต๋๋ค. ๊ณผํ์ ์ํต๊ณผ ์ค๋์ด๋ผ๋ ์ธ๊ฐ ๊ณ ์ ์ ์์ญ ๋ํ ์ฌ์ ํ ์ค์ํฉ๋๋ค. AI์ ์ธ๊ฐ์ ์ํธ ๋ณด์์ ์ธ ๊ด๊ณ ์์์, ์ฐ๋ฆฌ๋ ์ด์ ์ ์์ํ ์ ์์๋ ๋ฐฉ์์ผ๋ก ๊ณผํ์ ์ฌ์ ์ํ๊ณ , ๊ถ๊ทน์ ์ผ๋ก๋ โ๊น์ดโ์ โํญโ์ ๋ชจ๋ ์์ฐ
โInside Palantir: Building Software That Matters | Shyam Sankar on a16zโ โ a16z ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
์ํ๋ฐ๋ ๋ฏธ๊ตญ, AI๋ก ์ฌ๋ฌด์ฅํ๋ผ: ํ๋ํฐ์ด ์ด ์ฐ์นด๋ฅด๊ฐ ๋งํ๋ ๊ตญ๋ฐฉ ํ์ ๊ณผ โ์ ๊ตญ๊ฐ์ ๋์โ์ ์๋
์๊ธฐ ์ ๋ฏธ๊ตญ์ ๊ฐ์ฑ: ๊ตญ๋ฐฉ ํ์ ์ ๋ชฉ์๋ฆฌ
์ธ๊ณ ๊ฐ์ง์์ ๋ฒ์ด์ง๋ ์ฌ๊ฑด๋ค์ ์ฌ์ ํ ์ํนํ ์ผ๋ง์ฃผ์๊ฐ ์กด์ฌํ๋ฉฐ, ์ ์ด ์ค์ฌํ๋ค๋ ๋ํนํ ํ์ค์ ์๊ธฐ์ํจ๋ค. ํ ๋๋ผ๊ฐ ์ ์์ ๋์ ํ ๋, ๊ตญ๋ฐฉ๋ถ๋ง์ด ์ ์์ ์ํํ๋ ๊ฒ์ผ๋ก๋ ์ถฉ๋ถ์น ์๋ค. ์ง์ ํ ์ ์์ ์จ ๋๋ผ๊ฐ ํจ๊ป ์น๋ฅด๋ ๊ฒ์ด๋ฉฐ, ๋ฏธ๊ตญ ๊ตญ๋ฏผ์ด ์ด๋ฌํ ์๊ธฐ์ ๋์ํ ์ญ๋์ด ์๋ค๋ ์๊ฐ์ ๋ฏฟ๊ธฐ ์ด๋ ต๋ค. ํ๋ํฐ์ด(Palantir)์ ์ด ์ฐ์นด๋ฅด(Shyam Sankar)๋ ๋ฏธ๊ตญ์ ๊ฐ์ฅ ํฐ ์ํ์ ์ธ๋ถ์ ๊ณต๊ฒฉ์ด ์๋ โ์๋ฉธโ์ด๋ผ๊ณ ๊ฒฝ๊ณ ํ๋ฉฐ, AI ๊ฒฝ์์์ ์น๋ฆฌํ๊ณ ํนํ ๋ฌผ๋ฆฌ์ AI ๋ฐ ๋ก๋ด ๊ณตํ ๋ถ์ผ์์ ์์๋๊ฐ๊ธฐ ์ํ ๊ทผ๋ณธ์ ์ธ ๋ณํ๋ฅผ ์ด๊ตฌํ๋ค.
์ค๋ซ๋์ ํ๋ํฐ์ด์ ๊ณต๋ ์ฐฝ๋ฆฝ์์ด์ โํด๊ฒฐ์ฌ(fixer)โ๋ก ์๋ ค์ก๋ ์ด ์ฐ์นด๋ฅด๋ ์ต๊ทผ ๋ช ๋ ์ฌ์ด ๋์ค์ ์ธ๋ฌผ๋ก ๋ถ์ํ๋ค. ๊ทธ์ ๋๋ฃ๋ค์ ๊ทธ๋ฅผ โ๊ฒฝ๋ ฅ์ ์์ธโ์ผ๋ก ์นญ์กํ๋ฉฐ, ํ๋ํฐ์ด์ ์์ ํ๊ณ ์ง์ํ์ ๋ฟ ์๋๋ผ, ์๋ก์ด ์์์ ํ ์ ์๋๋ก โ๋ ๊ฐ๋ฅผ ๋ฌ์์ฃผ์๋คโ๊ณ ์ ์ ๋ชจ์๋ค. ์๋ ๋์ โ๋ฐฐํ ์ธ๋ฌผโ๋ก ํ๋ํ๋ ๊ทธ๊ฐ ์ ๋ฉด์ ๋์๊ฒ ๋ ๊ฒฐ์ ์ ์ธ ๊ณ๊ธฐ๋ 2022๋ ์ ๋ฐํํ ๊ตญ๋ฐฉ ๊ฐํ์ ๋ํ ์ค์ํ ๊ธ์ธ โํผ์คํธ ๋ธ๋ ํผ์คํธ(First Breakfast)โ์๋ค. ์ด ๊ธ์ ๋ฏธ๊ตญ์ด ๋์๊ฐ์ผ ํ ๋ฐฉํฅ์ ๋ํ ๊ฐ๋ ฅํ ๋ชฉ์๋ฆฌ๊ฐ ๋์๋ค.
์ฐ์นด๋ฅด๋ ์ด๋ฌํ ๋ณํ์ ๋๊ธฐ๋ฅผ โ์ ๋ฐํจ๊ณผ ๋๊ด์ฃผ์๊ฐ ๋ฐ๋ฐ ์์ธ ํ๋โ์ด๋ผ๊ณ ์ค๋ช ํ๋ค. ์๋ ๊ฐ ๊ตญ๋ฐฉ๋ถ ๋ด๋ถ์์ ๊ตญ๋ฐฉ ์์คํ ์ด ์์ํ ์ ํ๋๋ ๊ฒ์ ๋ชฉ๊ฒฉํ๋ฉฐ, ๊ทธ๋ โ๋๋ ๋ฌผ ์ ๊ฐ๊ตฌ๋ฆฌ(frog boil)โ์ฒ๋ผ ์ ์ง์ ์ผ๋ก ๋ค๊ฐ์ค๋ ์๊ธฐ๋ฅผ ์ ๊ฐํ๋ค. ํ์ง๋ง ๋์์ ๊ตญ๋ฐฉ ๋ถ์ผ์ ๊ธฐ์ฌํ๋ ค๋ ์ฐฝ์ ๊ฐ๋ค์ ์์ฒญ๋ ์๋์ง์ ๊ตญ๊ฐ์ ๊ด์ฌ์ด ์ฌ๋ถ์ํ๋ ๊ฒ์ ๋ณด๋ฉฐ, ์ง๊ธ์ด์ผ๋ง๋ก ๋ชจ๋ ๊ฒ์ ๋ฐ๋ก์ก์ ์ ์๋ โ๊ธฐํ์ ์๊ฐโ์ด๋ผ๊ณ ํ๋จํ๋ค. ๊ทธ๋ ๊ณผ๊ฑฐ ์ฐ๋ฆฌ๊ฐ ์น๋ฆฌํ๋ ๋ฐฉ์๋ค์ ์ฐ์ฐํ ๋ฑํ์ํ๊ฒ ๋์์ผ๋ฉฐ, ์ด์ ๋ ๊ทธ ๋ฐฉ์์ ํ๊ธฐ์ฐจ๊ฒ ๋์ฐพ์ ๊ธฐํ๊ฐ ์๋ค๊ณ ๊ฐ์กฐํ๋ค.
๋ฌ์์์ 2014๋ ํฌ๋ฆผ๋ฐ๋ ํฉ๋ณ, 2015๋ ์คํ๋ํ๋ฆฌ ๊ตฐ๋(Spratly Islands)์ ๊ตฐ์ฌํ, ์ด๋ ํต ํ์ (JCPOA)์ ์คํจ, ๊ทธ๋ฆฌ๊ณ 2023๋ 10์ 7์ผ ํ๋ง์ค์ ์ด์ค๋ผ์ ๊ณต๊ฒฉ ์ดํ ๋ฐ์ํ ์ฐธํนํ ์ฌ๊ฑด๋ค์ ์ฐ์นด๋ฅด์๊ฒ โ๋๋์ฒด ๋ฌด์จ ์ผ์ด ๋ฒ์ด์ง๊ณ ์๋๊ฐ?โ๋ผ๋ ๊ธ์ง์ ์ธ ์ง๋ฌธ์ ๋์ง๊ฒ ํ๋ค. ์ด๋ฌํ ์ผ๋ จ์ ์ฌ๊ฑด๋ค์ ๋ฏธ๊ตญ์ด ์ต์ง๋ ฅ(Deterrence)์ ์์คํ์์ ๋ณด์ฌ์ฃผ๋ ์ฆ๊ฑฐ์ด๋ฉฐ, ์๊ฐ์ด ์ผ๋ง ๋จ์ง ์์์์ ์๋ฏธํ๋ค. ๊ทธ๋ฌ๋ ๊ทธ๋ ์ง๋ 1๋ ๋์ ๊ตญ๋ฐฉ๋ถ ๋ด๋ถ์์ ์ง๋ 19๋ ๊ฐ ๋ณด์ง ๋ชปํ๋ ๋ณํ๊ฐ ์ผ์ด๋ฌ์ผ๋ฉฐ, ์ฌ๋๋ค์ด ์ด๋ฌํ ๊ฐํ์ ์๊ฐ์ ํฌ์ฐฉํ๊ณ ์๋ค๊ณ ๊ธ์ ์ ์ผ๋ก ํ๊ฐํ๋ค.
์ํ์ง ์น๋ฆฌ์ ๊ณต์: โ๊ตญ๊ฐ ์ ์ฒดโ์ ๋์
์ฐ์นด๋ฅด๋ ์ค๋๋ ์ ์ง์ ํ์ ์ํฉ์ 1930๋ ๋ ํ๋ฐ์ ๋น์ ํ๋ฉฐ, ๋ ํฐ ์ถฉ๋์ ๋ง๊ธฐ ์ํ โ๋์(Mobilize)โ์ ์ค์์ฑ์ ์ญ์คํ๋ค. ๊ทธ์ ์ ์ ใ๋ชจ๋น๋ผ์ด์ฆ(Mobilize)ใ๋ ์ฐ๋ฆฌ๊ฐ ๋์ ์์ ์น๋ฆฌ(ํน์ ์๋ จ์ด ํจ๋ฐฐ)ํ ํ ์์ผํจ์ ๋น ์ ธ ๋ฏธ๋์ ๋ํ ๋ง์ ์๋ชป๋ ๋ฏฟ์์ ๊ฐ๊ฒ ๋์๋ค๊ณ ์ง์ ํ๋ค. ๊ทธ๋ ์ 2์ฐจ ์ธ๊ณ๋์ ๋น์ ๋ฏธ๊ตญ์ ๋์์ด ์ง์ฃผ๋ง ๊ณต๊ฒฉ ์ดํ ๊ฐ์๊ธฐ ์ผ์ด๋ ๊ฒ์ด ์๋๋ผ๊ณ ๊ฐ์กฐํ๋ค. ์คํ๋ ค ํ๋ญํด๋ฆฐ ๋ฃจ์ค๋ฒจํธ(FDR) ๋ํต๋ น์ ๋น๋กฏํ ์ง๋๋ถ๋ 1930๋ ๋ ํ๋ฐ๋ถํฐ ๊ตญ๊ฐ์ ์์ง๊ฐ ์ถฉ๋ถํ ํ์ฑ๋์ง ์์์์๋ ๋ถ๊ตฌํ๊ณ ๋์์ ํ์์ฑ์ ์ธ์ํ๋ค. ๋๋๋ฆฌ์ค(Lend-Lease) ํ๋ก๊ทธ๋จ์ ๊ณต์ฅ ๊ฑด์ค๊ณผ ์ฌ์ ๋น์ 18๊ฐ์์ด ๊ฑธ๋ ธ๊ณ , ๊ฒฐ๊ตญ ์ง์ฃผ๋ง ๊ณต๊ฒฉ ๋น์ ๋ฏธ๊ตญ์ ์ด๋ฏธ ์์ ์์ฐ ์ฒด์ ๋ฅผ ๊ฐ์ถ๊ณ ์์๋ค.
์ง์ ํ ์ ์์ โ๊ตญ๊ฐ ์ ์ฒด(the whole country)โ๊ฐ ์น๋ฅด๋ ๊ฒ์ด๋ฉฐ, ๋จ์ํ ๊ตญ๋ฐฉ ์ฐ์ ๊ธฐ๋ฐ์ด๋ ๊ตญ๋ฐฉ๋ถ๋ง์ ๋ฌธ์ ๊ฐ ์๋๋ผ๋ ๊ฒ์ด ์ฐ์นด๋ฅด์ ํต์ฌ ์ฃผ์ฅ์ด๋ค. ๊ทธ๋ ๋ฏธ๊ตญ ์๋ฏผ ๋ชจ๋๊ฐ ๊ตญ๊ฐ๊ฐ ์ ๊ณตํ๋ ๋ฒ์๋ฟ ์๋๋ผ, ๊ทธ ๋ฒ์์ ๊ธฐ๋ฐ์ด ๋๋ ์์ ์ ํฌ์ํด์ผ ํ๋ค๊ณ ๋งํ๋ค. ๊ทธ๋ฌ๋ ์ฐ๋ฆฌ๋ ์ด๋ฌํ ์ ์ ์์ ๋๋ฌด๋ ๋ฉ๋ฆฌ ๋จ์ด์ ธ ์๋ค. 1989๋ ์๋ ์ฃผ์ ๋ฌด๊ธฐ ์์คํ ์ง์ถ์ 6%๋ง์ด ๊ตญ๋ฐฉ ์ ๋ฌธ ๊ธฐ์ ์ ํ ๋น๋์์ง๋ง, ํ์ฌ ์ด ์์น๋ 86%์ ๋ฌํ๋ค. ์ด๋ ๊ณผ๊ฑฐ์๋ ํฌ๋ผ์ด์ฌ๋ฌ(Chrysler)๊ฐ ๋ฏธ๋ํธ๋งจ(Minuteman) ๋ฏธ์ฌ์ผ๊ณผ ๋ฏธ๋๋ฐด์ ๋์์ ์์ฐํ๋ฏ, ๋ฏผ๊ฐ ์ฐ์ ์ด ๊ตญ๊ฐ ์๋ณด๋ฅผ ์์ฐ์ค๋ฝ๊ฒ ๋ท๋ฐ์นจํ๋ ๊ฒ๊ณผ๋ ๋์กฐ์ ์ด๋ค. ์ค๋๋ ๋ฏผ๊ฐ ๋ถ๋ฌธ์ ์ฐ๊ตฌ๊ฐ๋ฐ(R&D) ์ง์ถ์ ์ ๋ถ์ ์ญ๋์ ์๋ํ๋ฉฐ, ์ด ๋ฏผ๊ฐ ๊ธฐ์ ์ ๊ฐ๊ฒฉ ๋๋น ์ฑ๋ฅ ๊ณก์ ์ ๊ตญ๋ฐฉ๋ ฅ์ผ๋ก ์ ํํ๋ ๊ฒ์ด ์ค์ํ๋ค๊ณ ์ฐ์นด๋ฅด๋ ๊ฐ์กฐํ๋ค.
๊ณผ๊ฑฐ ๋ฏธ๊ตญ์ ์ฐ์ ๊ธฐ๋ฐ์ ๊ธ๋ ๋งํด(Glenn Martin), ์ญ ๋ ธ์ค๋ฝ(Jack Northrup), ๋ฅด๋ก์ด ๊ทธ๋ฃจ๋จผ(Leroy Grumman)๊ณผ ๊ฐ์ ์ฐฝ์ ๊ฐ๋ค์ด ์ด๋์๋ค. ์ด๋ค์ ๋จ๊ธฐ์ ์ธ ์ฌ๋ฌด ์ฑ๊ณผ๋ณด๋ค๋ ์์ ๊ณผ ํ์ฌ๋ณด๋ค ํจ์ฌ ํฐ ๋น์ ์ ์ถ๊ตฌํ๋ค. ์ด๋ฌํ ์ฐฝ์ ๊ฐ ์ ์ ์ ์ ๋ถ ๋ด๋ถ์๋ ์กด์ฌํ๋๋ฐ, ํด๊ตฐ ๋ด ๋ฐ๋์๋ ๋ถ๊ตฌํ๊ณ ํต ์ถ์ง ํด๊ตฐ์ ๊ฑด์คํ ํ์ด๋จผ ๋ฆฌ์ฝ๋ฒ(Hyman Rickover) ์ ๋ ์ด ๋ํ์ ์ด๋ค. ์ฌ์ง์ด โ์์ํญํ์ ์๋ฒ์งโ ์คํํ์ด๋จธ(Oppenheimer)์กฐ์ฐจ ๊ทธ์ ์คํจ๋ฅผ ์๊ฒฌํ์ง๋ง, ๋ฆฌ์ฝ๋ฒ๋ ๊ตดํ์ง ์์๋ค. ์ฐ์นด๋ฅด๋ ์ด๋ฌํ โํ์ธ ํ(chutzpah, ๋๋ดํจ)โ ์ ์ ์ด ๋ฐ๋ก ์ค๋ฆฌ์ฝ๋ฐธ๋ฆฌ ์ฐฝ์ ๊ฐ๋ค์ ํน์ง๊ณผ ์ผ๋งฅ์ํตํ๋ค๊ณ ๋งํ๋ค.
ํ์ ์ ๊ฑธ๋ฆผ๋: ํตํฉ๊ณผ ์์์ ๋๋ ๋ง
๋์ ์ข ์ ํ, ๋ฏธ๊ตญ์ โํํ ๋ฐฐ๋น๊ธ(peace dividend)โ์ ๊ธฐ๋ํ๋ฉฐ ๊ตญ๋ฐฉ๋น ์ง์ถ์ ์ค์๋ค. ์ด ์๊ธฐ์ 51๊ฐ์ ๋ฌํ๋ ์ฃผ์ ๊ตญ๋ฐฉ ๊ณ์ฝ์ ์ฒด๊ฐ ๋จ 5๊ฐ๋ก ์ค์ด๋๋ ์ ๋ช ํ โ์ตํ์ ๋ง์ฐฌ(Last Supper)โ์ด ์์๋ค. ์ฐ์นด๋ฅด๋ ์ด ์ฌ๊ฑด์ ๋ํ ํต์์ ์ธ ์ค๋ช , ์ฆ ํตํฉ์ผ๋ก ์ธํ ๊ฒฝ์ ์์ค์ด ๋ฌธ์ ์๋ค๋ ์ฃผ์ฅ์ด ํ๋ ธ๋ค๊ณ ์ง์ ํ๋ค. ๊ตญ๋ฐฉ ์ฐ์ ์ ๋ณธ์ง์ ์ผ๋ก ์์ ๋ ์ (Monopsony) ์์ฅ์ด๋ฉฐ, ํ์ ์ ๊ธฐ์ ๊ฐ ๊ฒฝ์์ด ์๋๋ผ ์ ๋ถ ๋ด๋ถ์ ์๋น์ค ๊ฐ ๊ฒฝ์์์ ์ด๋ฐ๋์๋ค๋ ๊ฒ์ด๋ค.
์ฐ์นด๋ฅด๊ฐ ๋ณด๊ธฐ์ โ์ตํ์ ๋ง์ฐฌโ์ผ๋ก ์ธํ ํตํฉ์ โ์์(conformity)โ์ ๋ณ์๋ค. ์ด๋ ๊ตญ๋ฐฉ ์ฐ์ ์ ์ง์ ํ โ๊ธ์ตํ(financialization)โ์ ์์์ด์๊ณ , ๊ธฐ์ ๋ค์ ์ฑ์ฅ๋ณด๋ค๋ ์ฌ๋ฌด ์งํ, ๋ฐฐ๋น, ์์ฌ์ฃผ ๋งค์ , ํ๊ธ ํ๋ฆ์ ์ง์คํ๊ฒ ๋์๋ค. ์ด๋ฌํ ์์์ ์ธ ํ๊ฒฝ์์๋ ์ฐฝ์ ๊ฐ ์ ์ ์ด ๋ฒ์ฑํ ์ ์์๋ค. โ์ด๋จ์(heretics)โ๋ค์ ์ถ์ถ๋๊ฑฐ๋, ๊ธฐ์ ์ฐ์ ๊ณผ ๊ฐ์ ๋ค๋ฅธ ๊ฒฝ์ ๋ถ๋ฌธ์ผ๋ก ๋ ๋ฌ๋ค.
๊ทธ๋ฌ๋ ์ฐ์นด๋ฅด๋ ์ด๋ฌํ โ์ด๋จ์โ๋ค์ด์ผ๋ง๋ก ํ์ ์ ํ์์ ์ด๋ผ๊ณ ๊ฐ์กฐํ๋ค. ๊ทธ์ ์ ์์ ๊ธฐ๋ก๋ ๋๋ผ์ด ๊ตญ๋ฐฉ ํ์ ์ฌ๋ก๋ค์ ๊ฑฐ์ ์์ธ ์์ด โ์ด๋จ์ ์ธ ์์ด๋์ดโ์์ ๋น๋กฏ๋์๋ค. ์ ๋, ๊ด๋ฃ์ฃผ์, ๊ธฐ์กด ์ ์ฐจ๋ ์ด๋ค์ ์ต์ํ๋ ค ํ์ง๋ง, ๊ฒฐ๊ตญ ์ฑ๊ณตํ๋ค. ์ 2์ฐจ ์ธ๊ณ๋์ ์ ์น๋ฆฌ๋ก ์ด๋ โํ๊ธด์ค ๋ณดํธ(Higgins boat)โ๊ฐ ๋ํ์ ์ด๋ค. ํด๊ตฐ์ ์ด ๋ณดํธ ๊ตฌ๋งค๋ฅผ ๊ฑฐ๋ถํ๊ณ ์ค๊ณ๋๋ฅผ ํ์น๋ ค ํ์ง๋ง, ๊ฒฐ๊ตญ ์ ์ ์ค ๋ชจ๋ ์๋ฅ์ ์ 92%๊ฐ ํ๊ธด์ค ๋ณดํธ์๋ค. ๋ง์ฝ ์ฐฝ์ ๊ฐ ํ๊ธด์ค๊ฐ ๋ณ์ ์ผ๋ก ์ง์ฐฉํ์ง ์์๋ค๋ฉด ๋ ธ๋ฅด๋ง๋ ์๋ฅ์์ ์ ๋ถ๊ฐ๋ฅํ์ ๊ฒ์ด๋ค.
์ฐ์นด๋ฅด๋ ์ด๋ฌํ ๋ฌธ์ ์ ๋ํด, ์์ ๋ ์ ์ ํน์ฑ์ ๊ตญ๋ฐฉ๋ถ๊ฐ ๊ณต๊ธ์ ์ฒด์ ์จ๊ฐ ์ ์ฝ์ ๊ฐํ๋ฉฐ โ๊ฐ๋ผํ๊ณ ์ค ์ฌ(Galapagos Islands)โ์ ๊ณ ๋ฆฝ๋ ์๋ฌผ์ฒ๋ผ ๋ง๋ค์๋ค๊ณ ๋น์ ํ๋ค. ๊ทธ ๊ฒฐ๊ณผ, ๋ฐ์ด๋ โ๊ฑฐ๋ ๊ฑฐ๋ถ์ดโ ๊ฐ์ ๊ธฐ์ ๋ค์ด ํ์ํ์ง๋ง, ์ด๋ค์ ๋ณธํ ์์๋ ๊ฒฝ์๋ ฅ์ ์๊ณ โ๋๋โ๋ค์๊ฒ ์ก์๋จนํ ๊ฒ์ด๋ผ๋ ๋นํ์ด๋ค. ํ๋ํฐ์ด๊ฐ ์ค๋ฆฝ๋ ๋น์๋ง ํด๋ ๊ตญ๋ฐฉ๋ถ์๋ ์ธ๋ถ์ธ์ด ์ ๊ทผํ โ์ ๋ฌธ(front door)โ์ด ์ฌ์ค์ ์์๊ณ , ์ ๋ณด ์ปค๋ฎค๋ํฐ์ ์ธํํ (In-Q-Tel)๋ง์ด ์ ์ผํ ํต๋ก์๋ค. ๊ทธ๋ฌ๋ 2015๋ ๋ถํฐ ์์๋ ๋ณํ์ ๋ฌผ๊ฒฐ๋ก ์ธํด, ์ง๋ 10๋ ๊ฐ ๊ตญ๋ฐฉ๋ถ๋ฅผ ๋๊ธฐ ์ํด ๋ฌธ์ ๋๋๋ฆฌ๋ โ์ด๋จ์โ๋ค์ด ๋น๋ก์ ๊ธฐํ๋ฅผ ์ป๊ฒ ๋์๋ค๊ณ ์ฐ์นด๋ฅด๋ ๋งํ๋ค.
์ด๋จ์ ์์ ๋ค์ ๊ทํ: ๊ตญ๋ฐฉ์ ์ํ ๋ฏผ๊ฐ ์ธ์ฌ ๋์
์ฐ์นด๋ฅด๋ ๊ตญ๋ฐฉ ์ฐ์ ์ โ์ด๋จ์ ์์ ๋คโ ์ค ํ ๋ช ์ผ๋ก ๋๋ฃจ ์ฟ ์ปค(Drew Cukor) ๋๋ น์ ๊ผฝ๋๋ค. ๊ทธ๋ ํด๋ณ๋ ๋๋ น์ผ๋ก, ๋จ ํ ๋ฒ์ ์์ ์คํจ(RPG ์ค์ธ์ผ๋ก ์ธํ ์ผ์ง๋์กฑ ๋๋ฏผ ๊ตฌ์กฐ ์คํจ)๊ฐ ์์ฒ ๋ช ์ ๋ฏผ๊ฐ์ธ์๊ฒ ์ฐธํนํ ๊ฒฐ๊ณผ๋ฅผ ๊ฐ์ ธ์จ ๊ฒ์ ๋ชฉ๊ฒฉํ ํ, AI๋ฅผ ๊ตญ๋ฐฉ๋ถ์ ๋์ ํ๋ ๋ฉ์ด๋ธ ํ๋ก์ ํธ(Project Maven)์ ์ ๋ ํ๋ค. ์์๋ ์๋ ์ํฉ์์ ์์๋ ์ด ํ๋ก์ ํธ์์ ์ฟ ์ปค ๋๋ น์ ๋ชจ๋ ์ด๋ค์ ๋ฐ๋์ ์งํฌ, ์ฌ์ง์ด ํ์ ๊ณ ๋ฐ(์ด๋์ธ์ ์ง์ ์จ๊ธฐ๊ณ ์๋ค๋)๊น์ง ๊ฐ์ํด์ผ ํ๋ค. ํ์ง๋ง ๊ทธ๋ ํ๋ค๋ฆฌ์ง ์์๊ณ , ๊ฒฐ๊ตญ AI์ ๊ตญ๋ฐฉ ๋์ ์ ๊ฒฐ์ ์ ์ธ ์ญํ ์ ํ๋ค. ์ฐ์นด๋ฅด๋ ์ด๋ฌํ ์ด์ผ๊ธฐ๊ฐ โ๋๋ผ๊ฐ ๋น์ ์ ์ ๋ง ํ์๋ก ํ ๋, ์จ์ด์๋ ์ด๋จ์๋ค์ด ๋์๋๋ก ์๊ฐ์ ์ฃผ๋ ๊ฒโ์ด๋ผ๊ณ ๊ฐ์กฐํ๋ค.
์กด ๋ณด์ด๋(John Boyd) ๋๋ น์ F-16 ๊ฐ๋ฐ๊ณผ OODA ๋ฃจํ(OODA Loop) ์ด๋ก , ๋ฒ๋๋ ์๋ฆฌ๋ฒ(Bernard Schriever) ์ฅ๊ตฐ๊ณผ ์๋์๋ ํ(Edward Hall) ๋ฐ์ฌ์ ๋ฏธ๋ํธ๋งจ ๋ฏธ์ฌ์ผ ๊ฐ๋ฐ ์ฌ๋ก ๋ํ โ์ด๋จ์โ๋ค์ ๋ณดํธํ๊ณ ์ง์ํ๋ ๋ฆฌ๋์ญ์ ์ค์์ฑ์ ๋ณด์ฌ์ค๋ค. ์ด๋ค์ ๋ชจ๋ ์กฐ์ง ๋ด์์ ์์ฒญ๋ ์ ํญ์ ๋ถ๋ชํ์ง๋ง, ๊ทธ๋ค์ ํ์ํจ์ ์์๋ณธ ๋ฆฌ๋๋ค์ ๋ณดํธ ์๋ ํ์ ์ ์ด๋ค๋ผ ์ ์์๋ค.
์ด๋ฌํ ๋งฅ๋ฝ์์ ์ด ์ฐ์นด๋ฅด ์์ ๋ ์ต๊ทผ ์ก๊ตฐ์ ์ ๋ํ๋ค. ๊ทธ๋ 2014๋ ๋ถํฐ ์ด์ค๋ผ์๊ณผ ํ๋ ฅํ๋ฉฐ, 10์ 7์ผ ์ดํ ์ด์ค๋ผ์์ด 36๋ง ๋ช ์ ์๋น๊ตฐ์ ๋์ํ์ ๋ ๋๋ผ์ด ๋ณํ๋ฅผ ๋ชฉ๊ฒฉํ๋ค. ์ด์ค๋ผ์ ์๋น๊ตฐ๋ค์ 20๋ ์ด์ ๋ฏผ๊ฐ ์ฐ์ ์์ ์ธํฐ๋ท ๊ท๋ชจ์ ์์คํ ์ ๊ตฌ์ถํ๋ฉฐ ์์ ๊ฒฝํ์ ๋ฐํ์ผ๋ก, ๋ถ๊ณผ 4๊ฐ์ ๋ง์ ๊ตฐ๋์ ๊ธฐ์ ์ 10๋ ์ด์ ๋ฐ์ ์์ผฐ๋ค. ์ฐ์นด๋ฅด๋ ๋ฏธ๊ตญ ๋ํ ์ด์ค๋ผ์์ฒ๋ผ ์์ฒญ๋ ๊ธฐ์ ์ธ์ฌ๋ฅผ ๋ณด์ ํ๊ณ ์์ผ๋ฉฐ, 24์ธ์ ์ ์ ์์ ์๋ ํ ์ ์์๋ ๊ธฐ์ฌ๋ฅผ 44์ธ์ ๊ฒฝํ์ผ๋ก ํ ์ ์๋ค๊ณ ๋งํ๋ค. ์ 2์ฐจ ์ธ๊ณ๋์ ๋น์ 10๋ง ๋ช ์ ๋ฏผ๊ฐ์ธ์ด ๊ตฐ์ผ๋ก ํน๋ณ ์๊ด(direct commissioned)๋์๋ ๊ฒ์ฒ๋ผ, ์ค๋๋ ์๋ ๋ฏผ๊ฐ ๊ธฐ์ ์ธ๋ ฅ์ ๊ตญ๋ฐฉ์ ํ์ฉํด์ผ ํ๋ค๋ ๊ฒ์ด๋ค. ๊ทธ๋ ์ ์คํAI(OpenAI) ์์ ์ฐ๊ตฌ์ ๋ฐฅ ๋งฅ๊ทธ๋ฃจ(Bob McGrew), ๋ฉํ(Meta)์ ์ต๊ณ ๊ธฐ์ ์ฑ ์์(CTO) ๋ณด์ค(Bos), ์ ์คํAI ์ต๊ณ ์ ํ์ฑ ์์ ์ค๋๋ฅ ์(Andrew Wheel) ๋ฑ ์ธ ๋ช ์ ๋๋ฃ์ ํจ๊ป ์ก๊ตฐ์ ํฉ๋ฅํ์ฌ ๊ณ ์ ์๋ฌธ๊ด์ผ๋ก์ ๋ค์ํ ํ๋ก์ ํธ๋ฅผ ๋๊ณ ์๋ค.
์ฐ์นด๋ฅด๋ ์ก๊ตฐ ๋ด๋ถ์์ ๊ฐ์ฅ ๋๋๋ ์ ์ผ๋ก โ์ ์ธ๋ ์ฅ๋ณ๋ค(green suitors)โ์ ํ์ํ ์ญ๋์ ๊ผฝ๋๋ค. ์ด๋ค์ ์ ์ ์ปดํจํฐ ๊ณผํ ๊ต์ก์ ๋ฐ์ง ์์์์๋ ๋ถ๊ตฌํ๊ณ , ์ค์ค๋ก AI ์ ํ๋ฆฌ์ผ์ด์ ์ ๊ตฌ์ถํ๋ ๋๋ผ์ด ๋ฅ๋ ฅ์ ๋ณด์ฌ์ค๋ค. ํนํ โ์์ ๋ฐ์ดํฐ ํ(Operational Data Teams)โ์์ ์ด๋ค์ด ๊ฐ๋ฐํ๋ AI ์๋ฃจ์ ์ ์์ ๋ถ๋ฌธ์์ ๋ณผ ์ ์๋ ์ด๋ค ๊ฒ๋ณด๋ค๋ ์ค๋๋ ฅ์ด ์๋ค๊ณ ํ๋ค. ์ด๋ ์์กด์ด ๊ฑธ๋ฆฐ โ์ค์กด์ ์ํโ ์์์ ์ฌ๋๋ค์ 10%์ ํจ์จ์ฑ ๊ฐ์ ์ด ์๋ โ์น๋ฆฌโ๋ผ๋ ์ด๋ถ๋ฒ์ ๋ชฉํ์ ์ง์คํ๊ธฐ ๋๋ฌธ์ด๋ค. ๋ํ AI๋ ํน์ ๊ธฐ์ ์ ๊ฐ์ง ๊ฐ์ธ, ํนํ ์์ ์ ์ ๋ฌธ ์์ญ์ ๊น์ด ์ดํดํ๋ ์ ๋ณด ์ฅ๊ต๋ค์๊ฒ ์์ฒญ๋ ํ์ ์ค์ด์ค๋ค. ๊ณผ๊ฑฐ์๋ ํ์ํฌ์ธํธ ์ฌ๋ผ์ด๋๋ฅผ ๋ง๋ค๋ ์ด๋ค์ด ์ด์ ๋ 2์ฃผ ๋ง์ ์ง์ ์๋ฃจ์ ์ ๊ตฌ์ถํ๊ณ , ๊ทธ ์ฑ๊ณผ๋ฅผ ํตํด ๊ตฐ์ ๋ฐ์ ์ํค๋ ๊ฒฝํ์ ๋ํ๋ฅผ ๋๋๊ณ ์๋ค. ์ด๋ฌํ ๋ฐํ ์ (bottom-up) ํ์ ๊ณผ ์๋ฌดํ ์งํ(Mission Command)๋ ๋ฏธ๊ตญ ๊ตฐ๋์ ๊ณ ์ ํ ๊ฐ์ ์ด๋ผ๊ณ ์ฐ์นด๋ฅด๋ ๊ฐ์กฐํ๋ค.
AI ์๋์ ์ํํธ์จ์ด ์งํ ๋ณํ: โ๋ฒ ํโ์ โ์ํโ์ ๊ฐ๋ฆผ๊ธธ
์ต๊ทผ ์ํํธ์จ์ด ์ ๊ณ์์๋ AI์ ๋ฑ์ฅ์ผ๋ก ์ธํ โSaaS ์ข ๋ง๋ก (SaaS Apocalypse)โ์ด ๋๋๋๊ณ ์๋ค. AI๊ฐ ์ฝ๋๋ฅผ ์์ฑํ๊ณ ํตํฉํ๋ ๋ฅ๋ ฅ์ด ํฅ์๋๋ฉด์, ๊ธฐ์กด SaaS(Software as a Service) ๊ธฐ์ ๋ค์ ์ง์ ์ฅ๋ฒฝ(switching costs)์ด ๋ฎ์์ง๊ณ ์๋ค๋ ์ฃผ์ฅ์ด๋ค. ์ด์ ๋ํด ์ฐ์นด๋ฅด๋ ์์ชฝ ์ฃผ์ฅ์ด ๋ชจ๋ ์ฌ์ค์ด๋ผ๊ณ ๋งํ๋ฉฐ, ์ํํธ์จ์ด๋ฅผ โ๋ฒ ํ(Beta)โ์ โ์ํ(Alpha)โ๋ก ๊ตฌ๋ถํ๋ ์๋ก์ด ๊ด์ ์ ์ ์ํ๋ค.
โ๋ฒ ํ ์ํํธ์จ์ดโ๋ ์ฌ์ฉ์๋ฅผ ๋ค๋ฅธ ๋ชจ๋ ์ฌ๋๊ณผ โ์ ์ฌํ๊ฒโ ๋ง๋๋ ๋ฐ ์ด์ ์ ๋ง์ถ๋ค. ์ด๋ ๊ณผ๊ฑฐ ์ํํธ์จ์ด ์ฐ์ ์ ๋ฌธ์ ์ ์ด๊ธฐ๋ ํ๋๋ฐ, ์ํํธ์จ์ด ๊ฐ๋ฐ์ ํผ๋๋ฐฑ ๋ฃจํ๊ฐ โ๊ฐ์น ์ฐฝ์ถโ๋ณด๋ค๋ โํ๋งค ๊ฐ๋ฅ์ฑโ์ ์ง์ค๋์ด ์์๊ธฐ ๋๋ฌธ์ด๋ค. ์ฐ์นด๋ฅด๋ ์ด๋ฌํ ๋ฒ ํ ์ํํธ์จ์ด๋ AI ์๋์ ํฐ ์ด๋ ค์์ ๊ฒช์ ๊ฒ์ด๋ผ๊ณ ์์ธกํ๋ค. ๋ฐ๋ฉด, AI๋ ์ฌ์ฉ์์ ๊ณ ์ ํ ์๊ตฌ์ ๋ง์ถฐ ์ํํธ์จ์ด๋ฅผ ๋ง๋ค ์ ์๊ฒ ํจ์ผ๋ก์จ ๋ณธ์ง์ ์ผ๋ก โ์ํ(Alpha)โ์ ์ง์คํ ์ ์๊ฒ ํ๋ค. ์ํ ์ํํธ์จ์ด๋ ์ฌ์ฉ์๋ฅผ โ๋ค๋ฅด๊ฒโ ๋ง๋ค๊ณ , ๊ฒฝ์ ์ฐ์์ ์ ๋ต์ ํํํ ์ ์๊ฒ ํด์ค๋ค. ๊ทธ๋ ์ด๋ฏธ ์ํ์ ์ง์คํ๊ณ ์๋ ํ๋ซํผ๋ค์ AI ์๋์ ๋์ฑ ๊ฐ๋ ฅํ ๋ฐ๋์ ๋ง์ ๊ฒ์ด๋ผ๊ณ ๋งํ๋ค.
์ฝ๋ก๋19 ํฌ๋ฐ๋ฏน ๋น์, ๋ง์ ๊ธฐ์ ์ ์ ์ฌ์ ์์ ๊ด๋ฆฌ(ERP) ์์คํ ์ด ๋ฌด๋์ก์ง๋ง, ์ค(Zoom)์ด๋ ํ์ฆ(Teams) ๊ฐ์ ํ์ ๋๊ตฌ๋ค์ ์๊ฒฉ ๊ทผ๋ฌด๋ฅผ ๊ฐ๋ฅํ๊ฒ ํ๋ค. ์ฐ์นด๋ฅด๋ ์ด๋ฅผ ์ํํธ์จ์ด ์ฐ์ ์ โ์คํธํธ๋ํฌ ์ผํฌ(Sputnik moment)โ๋ก ๋น์ ํ๋ฉฐ, ์ง๋ ์์ญ ๋ ๊ฐ ์ฐ๋ฆฌ๊ฐ ๊ฐ์น ์๋ค๊ณ ๋ฏฟ์๋ ์๋ง์ ์ํํธ์จ์ด ์ง์ถ์ด ์ฌ์ค์ ๊ทธ๋ ์ง ์์์์ ๊นจ๋ฌ์์ผ ํ๋ค๊ณ ์ง์ ํ๋ค. ํ๋ํฐ์ด๋ ๊ณ ๊ฐ๋ค์ด ํผ๋์ ์๋์ ๋ง์ถฐ ํ์ค์ ์ ์ํ๋๋ก ๋์ฐ๋ฉฐ ์คํ๋ ค ํฐ ์ฑ์ฅ์ ๊ฒฝํํ๋ค. ๊ทธ๋ ๋ง์ ๊ธฐ์ ๋ค์ด ๋ชจ๋ฐฉ์ ์ผ๋ก ์ง์ถํ๋ โ๋ฒ ํโ ์ํํธ์จ์ด๋ค์ด ํฐ ์๋ฐ์ ๋ฐ์ ๊ฒ์ด๋ฉฐ, ๋ฐ๋๋ก ๊ธฐ์ ์ ๊ฒฝ์ ์ฐ์๋ฅผ ํํํ๋ โ์ํโ ์ํํธ์จ์ด, ์ฆ ํดํท(toolkit)๊ณผ ๊ฐ์ ์ ๊ทผ ๋ฐฉ์์ด ๊ฐ๊ด๋ฐ์ ๊ฒ์ด๋ผ๊ณ ๋ด๋ค๋ณธ๋ค.
AI ์๋์ ๊ฐ์น ์ฐฝ์ถ์ ๋ํด์๋, ํ์ฌ ํ๋์จ์ด ๊ณ์ธต์ด ๊ฐ์ฅ ๋์ ๋ง์ง์ ์ฐจ์งํ๊ณ ์์ง๋ง, ์ธํฐ๋ท ๊ฒฝ์ ์์๋ ์ ํ๋ฆฌ์ผ์ด์ ๊ณ์ธต์ด ๊ทธ๋ฌ๋ค. ์ฐ์นด๋ฅด๋ AI ์๋์๋ ์นฉ(chips) ๊ณ์ธต๊ณผ AI ์ธํ๋ผ(AI infrastructure) ๊ณ์ธต(๊ทธ๋ ์ด๋ฅผ โ์จํจ๋ก์ง(Ontology)โ๋ผ๊ณ ๋ถ๋ฅธ๋ค)์์ ๊ฐ์น๊ฐ ์ฐฝ์ถ๋ ๊ฒ์ด๋ฉฐ, ์ด ๋ ๊ณ์ธต์ด ์๋นํ ๋ฐฉ์ด๋ ฅ์ ๊ฐ์ง ๊ฒ์ด๋ผ๊ณ ์์ธกํ๋ค. ๋ชจ๋ธ(models) ์์ฒด๋ ์ ์ฐจ ์ํํ๋ ์๋ ฅ์ ๋ฐ์ผ๋ฏ๋ก, ๋ชจ๋ธ ๊ธฐ์ ๋ค์ AI ์ธํ๋ผ๋ฅผ ๊ตฌ์ถํ๋ ๋ฐฉํฅ์ผ๋ก ํ์ฅํ๊ณ , ํน์ ์์ง AI ์๋ฃจ์ ๊ธฐ์ ๋ค์ ๊ณ ๊ฐ ๊ธฐ๋ฐ์ ํ์ฅํ๊ธฐ ์ํด AI ์ธํ๋ผ๋ก ๋ด๋ ค์ค๊ฒ ๋ ๊ฒ์ด๋ผ๋ ๋ถ์์ด๋ค.
์ธ๊ฐ์ ์์ง, AI์ ๋ฏธ๋: ๊ฒฝ์ ์ ์ผ์๋ฆฌ์ ๋ํ ํต์ฐฐ
AI๊ฐ ๊ฒฝ์ ์ ์ผ์๋ฆฌ์ ๋ฏธ์น ์ํฅ์ ๋ํด ์ฐ์นด๋ฅด๋ ์ค์ํ ๊ด์ ์ ์ ์ํ๋ค. ๊ทธ๋ AI์ ๋ํด ์ด์ผ๊ธฐํ ๋ ๋ง์น ์ธ๊ฐ์ ์์ง(human agency)๊ฐ ์๋ ๊ฒ์ฒ๋ผ โAI๊ฐ X๋ฅผ ํ ๊ฒ์ด๋คโ๋ผ๊ณ ๋งํ๋ ๋ฐฉ์์ ๋ถํธํจ์ ๋๋๋ค๊ณ ๋งํ๋ค. โ์ธ๊ฐ์ด AI๋ฅผ ์ฌ์ฉํ์ฌ X๋ฅผ ํ ๊ฒ์ด๋คโ๊ฐ ์ฌ๋ฐ๋ฅธ ํํ์ด๋ฉฐ, ์ฐ๋ฆฌ์๊ฒ๋ AI๋ฅผ ํน์ ๋ฐฉํฅ์ผ๋ก ์ด๋ ์ ํ๊ถ๊ณผ ์๋ฌด๊ฐ ์๋ค๋ ๊ฒ์ด๋ค.
๊ทธ๋ AI๊ฐ 1970๋ ๋ ์ดํ ๋ฒ์ด์ง ์๊ธ ์ฑ์ฅ๊ณผ ๊ตญ๋ด์ด์์ฐ(GDP) ์ฑ์ฅ ๊ฐ์ ๊ทผ๋ณธ์ ์ธ ๋จ์ ์ ํด๊ฒฐํ ์ญ์ฌ์ ๊ธฐํ๋ฅผ ์ ๊ณตํ๋ค๊ณ ๋ณธ๋ค. AI๋ ๋ฏธ๊ตญ ๋ ธ๋์๋ค์๊ฒ โ์ด๋ฅ๋ ฅ(superpowers)โ์ ๋ถ์ฌํ์ฌ ์์ฐ์ฑ์ ๊ทน๋ํํ๊ณ , ์ด๋ ๊ตญ๊ฐ์ ์ฌ์ฐ์ ํ(re-industrialization)๋ฅผ ๋ท๋ฐ์นจํ ๊ธฐ๋ฐ์ด ๋ ์ ์๋ค๋ ๊ฒ์ด๋ค. ํ๋๋ฆฌ์(Hadrien)๊ณผ ๊ฐ์ ๊ธฐ์ ๋ค์ด ๊ธฐ์ ์ ํ์ฉํ์ฌ ์ฌ๋๋ค์ ์์ฐ์ฑ์ 50~100๋ฐฐ ํฅ์์ํค๋ฉฐ ์ฌ์ฐ์ ํ๋ฅผ ์ด๋๋ ๊ฒ์ด ๊ทธ ์์์ด๋ค.
๋ํ ์ฐ์นด๋ฅด๋ ์ธ๊ณํ์ โ๊ฑฐ๋ํ ๊ฑฐ์ง๋งโ ์ค ํ๋๊ฐ ํ์ (innovation)์ ํ๊ณณ์์ ํ๊ณ ์์ฐ(production)์ ๋ค๋ฅธ ๊ณณ์์ ํ ์ ์๋ค๋ ๋ฏฟ์์ด์๋ค๊ณ ์ง์ ํ๋ค. ๊ทธ๋ ํ์ ์ ์์ฐ์ฑ์ ๊ฒฐ๊ณผ์ด๋ฉฐ, ๋ฌด์ธ๊ฐ๋ฅผ ๋ง๋ค์ง ์์ผ๋ฉด ๊ทธ๊ฒ์ ๋ง๋๋ ๋ฐฉ์์ด๋ ๋ฌด์์ ๋ง๋ค์ง์ ๋ํด ํ์ ํ ์ ์๋ค๊ณ ๊ฐ์กฐํ๋ค. ์คํ์ด์คX(SpaceX)์ R&D ์์ง๋์ด๋ค์ด ์์ฐ ํ์ฅ๊ณผ ํจ๊ป ์ผํ๋ ๊ฒ์ด ๋ํ์ ์ธ ์์์ด๋ค. ๊ทธ๋ AI๋ฅผ ๊ตญ๊ฐ์ ๋น์์ฌํ์ด์ ๋์์ ๊ตญ๊ฐ์ ๊ธฐํ๋ก ๋ณด์์ผ ํ๋ค๊ณ ์ฃผ์ฅํ๋ค.
์ญ์ฌ์ ์ผ๋ก ๊ธฐ์ ํ๋ช ์ ๋๋ถ๋ถ โ๋๊ตฌ ํ๋ช โ์ด์์ง โ๊ฐ๋ ํ๋ช โ์ด ์๋์๋ค. ๋ง์๊ฒฝ์ ๋ฐ๋ช ํ ๊ฒ์ ๊ฐ๋ฆด๋ ์ค๊ฐ ์๋๋ผ ๋ค๋ฅธ ์ฌ๋์ด์์ง๋ง, ๊ฐ๋ฆด๋ ์ค๋ ๊ทธ๊ฒ์ ์ฌ์ฉํ์ฌ ํ์ฑ ์ด๋์ ๋ฐ๊ฒฌํ๋ค. ํ๋ฏธ๊ฒฝ, ๋๋ ฅ ์ง์กฐ๊ธฐ, ๊ฐ์ธ์ฉ ์ปดํจํฐ ๋ฑ ์ด๋ฌํ ๊ธฐ์ ์ ๋ฏธ๋๋ ๊ธฐ์ ์ ๋ฐ๋ช ํ ์ฌ๋์ด ์๋๋ผ โ๊ธฐ์ ์ ์ฌ์ฉํ๋ ์ฌ๋(wielders)โ์ ์ํด ๊ฒฐ์ ๋๋ค. ์ค๋๋ AI ๋น๊ด๋ก (AI dumerism)์ ์ฃผ๋ก ๋งค์ฐ ๋๋ํ โ๋ฐ๋ช ๊ฐโ๋ค์ ๋ชฉ์๋ฆฌ์์ ๋์ค์ง๋ง, ๊ทธ๋ค์ ์ฐฝ์กฐ๋ฌผ์ฒ๋ผ ๊ทธ๋ค ์ญ์ โ๋ถ๊ท ํํ ์ง๋ฅ(jagged intelligence)โ์ ๊ฐ์ง ์ ์๊ธฐ์, ๋ชจ๋ธ ๊ตฌ์ถ์ ๋ฅ์ํ๋ค๊ณ ํด์ ๊ทธ ํจ์๊น์ง ์ ํํ ์์ธกํ ์ ์๋ ๊ฒ์ ์๋๋ค.
์ฐ์นด๋ฅด๋ ์ฐ๋ฆฌ๊ฐ AI์ ๋ฏธ๋๋ฅผ ๊ฒฐ์ ํ โ์ฌ์ฉ์โ๋ก์ ์ธ๊ฐ์ ์์ง๋ฅผ ํฌ๊ธฐํด์๋ ์ ๋๋ค๊ณ ๊ฐ์กฐํ๋ค. ๊ฒฝ์ ํ์๋ค์ AI ๋ฏธ๋ ์์ธก ๊ทธ๋ํ์ฒ๋ผ, ๋ค์ํ ๊ฒฐ๊ณผ๊ฐ ๋ชจ๋ ๊ฐ๋ฅํ๋ฉฐ, ์ฐ๋ฆฌ๊ฐ ์ด๋ค ๋ฏธ๋๋ฅผ ์ ํํ ์ง๋ ์ฐ๋ฆฌ์๊ฒ ๋ฌ๋ ค ์๋ค๋ ๊ฒ์ด๋ค.
๋ง์ง๋ง์ผ๋ก, ๊ทธ๋ ๊ณผ๊ฑฐ ๋ฏธ๊ตญ ๊ฒฝ์ ์์ ๊ณต
โSilicon Valleyโs Big Bets on War Pay Off, and the Trump Family Business Looks to Transylvaniaโ โ New York Times Podcasts ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
์ ์์ด ์ค๋ฆฌ์ฝ๋ฐธ๋ฆฌ์ ์๊ฒจ์ค ๊ธฐํ: ํธ๋ผํ ๊ฐ๋ฌธ์ ํธ๋์ค๋ฐ๋์ ๊ฐ๋ฐ ๋ฏธ์คํฐ๋ฆฌ
๊ฒฉ๋ํ๋ ๊ตญ์ ์ ์ธ ์์์ ๋ป๋ฐ์ ๊ธฐํ๋ฅผ ํฌ์ฐฉํ๋ ์ด๋ค์ด ์์ต๋๋ค. ๋ฏธ๊ตญ์ ๊ธฐ์ ํ๋ธ์ธ ์ค๋ฆฌ์ฝ๋ฐธ๋ฆฌ์ ๊ฑฐ๋ ๊ธฐ์ ๋ค์ ๊ธ๋ณํ๋ ์๋ณด ํ๊ฒฝ ์์์ โ์ ์์ ๊ธฐ์ โ์ ๊น์ด ๋ฐ์ด๋ค๋ฉฐ ์๋ก์ด ํฉ๊ธ๊ธฐ๋ฅผ ๋ง์ดํ๊ณ ์์ต๋๋ค. ํํธ, ๋๋๋ ํธ๋ผํ ์ ๋ํต๋ น์ ๊ฐ์กฑ ์ฌ์ ์ฒด์ธ ํธ๋ผํ ์กฐ์ง(Trump Organization)์ ๋ฐ๋ฆฌ๋ ๋ชฐ๋๋ธ ๊ฐ์ ํ๋ คํ ํด์์ง๊ฐ ์๋, ๋ฃจ๋ง๋์์ ์ธ๋ด ์ง์ญ ํธ๋์ค๋ฐ๋์์์ ์์ธ์ ๋ถ๋์ฐ ํ๋ก์ ํธ๋ฅผ ์ถ์งํ๋ฉฐ ์ธ๊ฐ์ ์ด๋ชฉ์ ์ง์ค์ํค๊ณ ์์ต๋๋ค. ๋ด์ํ์์ค(The Times)์ ์ต์ ๋ณด๋๋ ์ด ๋ ๊ฐ์ง ์๋ฐ๋ ํ๋ฆ์ด ์ด๋ป๊ฒ ์ค๋๋ ์ ์ธ๊ณ๋ฅผ ์ฌํธํ๊ณ ์๋์ง ์กฐ๋ช ํฉ๋๋ค.
์ด๋ ์ ์์ ๊ทธ๋ฆผ์, ์ค๋ฆฌ์ฝ๋ฐธ๋ฆฌ์ ์๋ก์ด ํ๋ก
์ด๋๊ณผ์ ์ ์์ด 3์ฃผ์ฐจ์ ์ ์ด๋ค๋ฉด์, ๋ฏธ๊ตญ์ ๊ตฐ์ฌ์ , ๊ฒฝ์ ์ , ๊ธฐ์ ์ ์ธก๋ฉด์์ ์ค๋ํ ๊ฒฐ์ ์ ๋ด๋ฆฌ๊ณ ์์ต๋๋ค. ํธ๋ผํ ๋ํต๋ น์ ์ด๋์ ๋ฏธ๊ตฐ ๋ณ๋ ฅ์ ํ๋ณํ ๊ฐ๋ฅ์ฑ์ โ๋ฐฐ์ ํ์ง ์๋๋คโ๊ณ ๋ฐํ๋ฉฐ, ํ์ํ๋ค๋ฉด ๊ตฐ์ฌ๋ ฅ ๋ฐฐ์น๋ฅผ โ๋๋ ค์ํ์ง ์๋๋คโ๊ณ ๊ฐ์กฐํ์ต๋๋ค. ์ด๋ ์ค๋ ์ง์ญ์ ๊ธด์ฅ๊ฐ์ ํ์ธต ๊ณ ์กฐ์ํค๋ ๋ฐ์ธ์ผ๋ก ํด์๋ฉ๋๋ค.
์๋์ง ์๊ธฐ ํด๊ฒฐ์ ์ํ ๋ฐฑ์ ๊ด์ ๊ณ ์ฌ๋ ๊น์ด์ง๊ณ ์์ต๋๋ค. ๋ฏธ ์ฌ๋ฌด์ฅ๊ด ์ค์ฝง ๋ฒ ์ฐํธ(Scott Bessant)๋ ํฅํ ์ด๋์ฐ ์์ ์ ๋ํ ์ ์ฌ๋ฅผ ํด์ ํ ๊ณํ์ด๋ผ๊ณ ๋ฐํ์ต๋๋ค. ์ด๋ ์ฝ 1์ต 4์ฒ๋ง ๋ฐฐ๋ด์ ๋ฌํ๋ ์์ ๊ฐ ์์ฅ์ ํ๋ฆด ์ ์์์ ์๋ฏธํ๋ฉฐ, ์๋ ๊ฐ ์ง์๋ ๋ฏธ๊ตญ์ ๋์ด๋ ์ ์ฑ ์ ๋ค์ง๋ ๊ฒฐ์ ์ ๋๋ค. ๋ฐฑ์ ๊ด์ ์ ๊ฐ ์์ ์ ์ํด ํ์ฌ์ ์ธ ๋ ธ๋ ฅ์ ๊ธฐ์ธ์ด๊ณ ์์ง๋ง, ์ด๋๊ณผ์ ์ ์ ์ค์๋ ์ด๋์ด ๋ ๋ง์ ์์ ๋ฅผ ํ๋งคํ๋๋ก ์ฅ๋ คํ๋ ์ญ์ค์ ์ธ ์ํฉ์ ์ด๋ํ ์ ์๋ค๋ ์ง์ ์ด ๋์ต๋๋ค.
ํํธ, ๋ฏธ๊ตญ ๊ตญ๋ฐฉ๋ถ๋ ์ ์ ์ง์์ ์ํด 2์ฒ์ต ๋ฌ๋ฌ์ ๋ฌํ๋ ๋ง๋ํ ์ถ๊ฐ ์์ฐ์ ์ํ์ ์์ฒญํ์ต๋๋ค. ์ด๋ ์ฐ๊ฐ ์ ์ฒด ๊ตญ๋ฐฉ ์์ฐ์ ๊ฑฐ์ 4๋ถ์ 1์ ํด๋นํ๋ ๊ท๋ชจ๋ก, ๊ตญ๋ฐฉ์ฅ๊ด ํผํธ ํ์ค(Pete Hgse)๋ ์ด ๊ธ์ก์ด ๋ณ๊ฒฝ๋ ์ ์์์ ์์ฌํ์ต๋๋ค. ์ด๋ฏธ ๋ถ์ด๋ ์ฌ๋ก ์์์ ์ด ๋ง๋ํ ์์ฐ ์์ฒญ์ ์ํ ๋ด ์จ๊ฑด ๊ณตํ๋น ์์๋ค์ ์ฐ๋ ค๋ฅผ ์ฌ๊ณ ์์ต๋๋ค.
์ด๋ฌํ ๊ฒฉ๋ ์์์ ํํธ์ผ๋ก๋ ์ค๋ฆฌ์ฝ๋ฐธ๋ฆฌ์ ๊ธฐ์ ๊ธฐ์ ๋ค์ด ์๋ก์ด ๊ธฐํ๋ฅผ ํฌ์ฐฉํ๊ณ ์์ต๋๋ค. ๊ณผ๊ฑฐ ์ค๋ฆฌ์ฝ๋ฐธ๋ฆฌ์์ ๊ตญ๋ฐฉ ๊ธฐ์ ์ โ์ ํด์ง์ง ๋ง๋ผ(Do No Evil)โ๋ ๊ตฌ๊ธ์ ๋ชจํ ์ฒ๋ผ ๊ธฐํผ ๋์์ด์์ต๋๋ค. ๊ตฌ๊ธ ์ง์๋ค์ ๊ตญ๋ฐฉ๋ถ์์ ํ๋ ฅ์ ๋ฐ๋ํ๋ ์์๋ฅผ ๋ฒ์ด๊ธฐ๋ ํ์ต๋๋ค. ๊ทธ๋ฌ๋ ๋ช ๋ ์ ๋ถํฐ ์ด๋ฌํ ๋ถ์๊ธฐ๋ ๊ธ๋ณํ์ต๋๋ค. ๊ตฌ๊ธ, ๋ฉํ(Meta), OpenAI ๊ฐ์ ๊ฑฐ๋ ๊ธฐ์ ๋ถํฐ ์๊ท๋ชจ ์คํํธ์ ์ ์ด๋ฅด๊ธฐ๊น์ง ์๋ง์ ๊ธฐ์ ๊ธฐ์ ๋ค์ด ๊ตญ๋ฐฉ ๊ธฐ์ ๋ถ์ผ๋ก ๋ฐฉํฅ์ ์ ํํ๊ธฐ ์์ํ์ต๋๋ค.
์ด๋ค ๊ธฐ์ ์ ์์ต์์ ์์ญ์ต ๋ฌ๋ฌ์ ์ด๋ฅด๋ ๊ณ์ฝ์ ํตํด ๋ฌด๊ธฐ ์์คํ ๋ถํฐ ๋ฏธ ์ ๋ถ๋ฅผ ์ํ ์ํํธ์จ์ด๊น์ง ๋ค์ํ ๊ธฐ์ ์ ๊ฐ๋ฐํ๊ณ ์์ต๋๋ค. ์๋ฅผ ๋ค์ด, ๋ฐ์ดํฐ ๋ถ์ ๊ธฐ์ ํ๋ํฐ์ด(Palantir)๋ ๋ฏธ ์ ๋ถ๊ฐ ๊ณต์ต ๋ชฉํ๋ฅผ ์ ์ ํ๋ ๋ฐ ๋์์ ์ฃผ๋ โํ๋ก์ ํธ ๋ฉ์ด๋ธ(Project Maven)โ ์์คํ ์ ๊ตฌ์ถํ์ต๋๋ค. ๊ตฌ๊ธ๊ณผ OpenAI๋ ์ธ๊ณต์ง๋ฅ(AI) ๊ธฐ์ ์ ํ์ฉํด ํ์ฅ ์งํ๊ด๋ค์ ์ง์ํ๊ณ ์์ผ๋ฉฐ, ๊ตฌ๊ธ ์ ์ต๊ณ ๊ฒฝ์์ ์๋ฆญ ์๋ฏธํธ(Eric Schmidt)๊ฐ ์ฃผ๋ํ๋ ํ๋ก์ ํธ๋ ์ด๋์์ ๋ฏธ๊ตฐ ์์ฐ์ด ์ด๋ ๋๋ก ์ ๋ฐฉ์ดํ๋ ๋ฐ ์ฌ์ฉ๋๋ ๋๋๋ก ์์คํ ์ ๊ฐ๋ฐํ๊ณ ์์ต๋๋ค. ํธ๋ผํ ํ์ ๋ถ๊ฐ 1์กฐ ๋ฌ๋ฌ ์ด์์ ๊ตญ๋ฐฉ๋น๋ฅผ ํ ๋นํ์ผ๋ฉฐ, ์ด ์ค ์๋น ๋ถ๋ถ์ด ์ค๋ฆฌ์ฝ๋ฐธ๋ฆฌ๊ฐ ๊ฐ๋ฐํ๊ณ ์ ํ๋ ์๋ก์ด ๊ตญ๋ฐฉ ๊ธฐ์ ์ ํฌ์ ๋ ๊ฒ์ผ๋ก ์์๋๋ฉด์, ๊ธฐ์ ๊ธฐ์ ๋ค์ โ์ ์ ํน์โ๋ ๋์ฑ ์ปค์ง ์ ๋ง์ ๋๋ค.
ํธ๋ผํ ๊ฐ๋ฌธ์ ์ด์ ํฌ์์ฒ, ํธ๋์ค๋ฐ๋์์ ๋ฏธ์คํฐ๋ฆฌ
๋ด์ํ์์ค์ ์๋ก์ด ์กฐ์ฌ์ ๋ฐ๋ฅด๋ฉด, ํธ๋ผํ ์ ๋ํต๋ น์ ๊ฐ์กฑ ์ฌ์ ์ฒด์ธ ํธ๋ผํ ์กฐ์ง์ด ๋ฃจ๋ง๋์์์ ๋ฏธ๋ฐํ ๋ถ๋์ฐ ๊ฑฐ๋๋ฅผ ์ถ์ง ์ค์ธ ๊ฒ์ผ๋ก ๋๋ฌ๋ฌ์ต๋๋ค. ๋ฐ๋ฆฌ๋ ๋ชฐ๋๋ธ์ฒ๋ผ ํ๋ คํ๊ณ ์ ๋ช ํ ๊ด๊ด์ง๊ฐ ์๋, ํธ๋์ค๋ฐ๋์๋ผ๋ ์์ธ์ ์ฅ์์ ๋๋ค.
ํ์ฅ์ ์ง์ ๋ฐฉ๋ฌธํ ๋ด์ํ์์ค ํ์ฌ ๋ณด๋ ๊ธฐ์ ๋ ๋ฒ ์นด ๋ฃจ์ด์ฆ(Rebecca Ruiz)๋ ๊ทธ๊ณณ์ ํ๊ฒฝ์ ์์ํ๊ฒ ๋ฌ์ฌํ์ต๋๋ค. ๋ถ์ง๋ ๋ฐ์ฏค ๋ฒ๋ ค์ง ์ํ์๊ณ , ๊ตฐ์ฌ ๊ธฐ์ง ์์ ์์นํ์ผ๋ฉฐ, ์ด ์ง์ญ ์ต๋ ๊ท๋ชจ์ ๊ฑฐ๋ํ ์ฐ๋ ๊ธฐ ๋งค๋ฆฝ์ง ๋ช ๊ณณ๊ณผ ์ธ์ ํด ์์์ต๋๋ค. ์ฝ๋ฅผ ์ฐ๋ฅด๋ ์ ์ทจ๊ฐ ์ง๋ํ๊ณ , ๋ค๊ฐ ๋ฌด๋ฆฌ๊ฐ ๋ฐฐํํ๋ ๋ชจ์ต๋ ๋ชฉ๊ฒฉ๋์์ต๋๋ค. ์ด๋ฌํ ํ๊ฒฝ์ ๋ญ์ ๋ฆฌ ์ํํธ ๋จ์ง์ ๊ณจํ ์ฝ์ค๋ฅผ ์์ํ๊ธฐ์๋ ๋งค์ฐ ์ด๋ก์ ์ธ ์ฅ์์์ต๋๋ค.
ํธ๋ผํ ์กฐ์ง์ด ์ ์ด ํน์ ์ฅ์๋ฅผ ํ๋ก์ ํธ ๋ถ์ง๋ก ์ ํํ๋์ง๋ ๋ช ํํ์ง ์์ต๋๋ค. ๊ทธ๋ฌ๋ ํ ๊ฐ์ง ๋จ์๋ ํธ๋ผํ ์ ๋ํต๋ น์ด ๋ฃจ๋ง๋์์์ ๋์ ์ธ๊ธฐ๋ฅผ ๋๋ฆฌ๊ณ ์๋ค๋ ์ ์ ๋๋ค. ์ง๋ํด ๊ฐค๋ฝ(Gallup) ์ฌ๋ก ์กฐ์ฌ์ ๋ฐ๋ฅด๋ฉด, ๋ฃจ๋ง๋์ ๊ตญ๋ฏผ์ ์ ๋ฐ ์ด์์ด ํธ๋ผํ ๋ํต๋ น์ ์ง๋ฌด ์ํ์ ๋ํด ๊ธ์ ์ ์ผ๋ก ํ๊ฐํ์ผ๋ฉฐ, ์ด๋ ์ ๋ฝ์ฐํฉ(EU) ๊ตญ๊ฐ๋ค ์ฌ์ด์์ ์ด๋ก์ ์ธ ๊ฒฐ๊ณผ์ ๋๋ค.
๋ ๋ฒ ์นด ๋ฃจ์ด์ฆ ๊ธฐ์๋ ์ด๋ฌํ ์์ง์์ด ํธ๋ผํ ์ ๋ํต๋ น์ ๋ ๋ฒ์งธ ์๊ธฐ์์ ๋ํ๋ ํจํด์ ์ฐ์ฅ์ ์์ ์๋ค๊ณ ๋ถ์ํฉ๋๋ค. ์ฆ, ํธ๋ผํ ์กฐ์ง์ ๊ทธ๊ฐ ์ธ๊ธฐ ์๋ ์ง์ญ์ ์ฐพ์ ์ ์ธ๊ณ๋ฅผ ์ฌํํ๋ฉฐ ์ฌ์ ๊ธฐํ๋ฅผ ๋ชจ์ํ๊ณ ์๋ค๋ ๊ฒ์ ๋๋ค. ๋ฐ๋ฆฌ์ ๋ชฐ๋๋ธ ๊ฐ์ ์ ํต์ ์ธ ๊ณ ๊ธ ํด์์ง์๋ ๊ฑฐ๋ฆฌ๊ฐ ๋จผ ํธ๋์ค๋ฐ๋์์์ ํธ๋ผํ ์กฐ์ง์ด ์ด๋ค ๊ทธ๋ฆผ์ ๊ทธ๋ฆฌ๊ณ ์๋์ง์ ๋ํ ์๋ฌธ์ ์ฌ์ ํ ๋จ์์์ต๋๋ค.
๊ฒฉ๋์ ์๋, ๋ณํ์ ๋ฌผ๊ฒฐ
์ด๋๊ณผ์ ์ ์๊ณผ ๊ทธ๋ก ์ธํ ์ค๋ฆฌ์ฝ๋ฐธ๋ฆฌ์ ๋ถ์, ๊ทธ๋ฆฌ๊ณ ํธ๋ผํ ๊ฐ๋ฌธ์ ์์์น ๋ชปํ ํฌ์์ฒ๋ ์ค๋๋ ์ธ๊ณ๊ฐ ์ผ๋ง๋ ๋น ๋ฅด๊ฒ ๋ณํํ๊ณ ์์ธก ๋ถ๊ฐ๋ฅํ ๋ฐฉํฅ์ผ๋ก ๋์๊ฐ๊ณ ์๋์ง๋ฅผ ๋ณด์ฌ์ค๋๋ค. ์ ์์ด ๊ธฐ์ ๊ธฐ์ ๋ค์๊ฒ ์๋ก์ด ์์ฅ์ ์ด์ด์ฃผ๊ณ , ์ ์น์ ์ธ๊ธฐ๊ฐ ๋น์ฆ๋์ค ๊ธฐํ๋ก ์ ํ๋๋ ์ด ์๋์ ํ๋ฆ์ ์์ผ๋ก๋ ์ธ๊ณ ์ ์น๊ฒฝ์ ์ ์ค๋ํ ์ํฅ์ ๋ฏธ์น ๊ฒ์ ๋๋ค. ๋ด์ํ์์ค ์น์ฌ์ดํธ(nytimes.com)์์ ํธ๋ผํ ์กฐ์ง์ ํธ๋์ค๋ฐ๋์ ํ๋ก์ ํธ์ ๋ํ ์ ์ฒด ์กฐ์ฌ ๋ด์ฉ๊ณผ ํ์ฅ ์ฌ์ง์ ํ์ธํ ์ ์์ต๋๋ค.