English
Based on โHow music rewires and impacts the human body | Michael Spitzer: Full Interviewโ from Big Think Watch the original video
The Ancient Echo: How Music Forged Humanity and Continues to Reshape Our Minds
For millennia, long before the first written word or even the earliest human civilizations, music has been an intrinsic part of the human experience. Itโs a universal language, a primal force that shaped our ancestors, continues to define our cultures, and profoundly rewires our brains. Michael Spitzer, Professor of Music at the University of Liverpool and author of The Musical Human: A History of Life on Earth, challenges conventional wisdom by tracing musicโs origins not just to human history, but to the very dawn of life, revealing its astonishing depth and enduring power.
Spitzerโs journey into musicโs past began with a profound realization: if Yuval Harariโs Sapiens blew minds by charting human history, then music, which predates Homo sapiens by at least a million years, deserved an even grander narrative. Eschewing the traditional โcomposer wrote what piece whenโ approach, Spitzer sought the โbigger picture,โ recognizing music as a global, universal phenomenon that transcends even our own species, echoing in the songs of birds and whales.
The Deep Roots of Human Melody: A Prehistory of Sound
Unraveling musicโs prehistory is a formidable task. Before Edisonโs phonograph in 1877, no sound recordings exist. Instruments made of organic materials like wood, skin, or gut biodegrade over time. Yet, traces remain. The earliest evidence points to โlithic instrumentsโโrocks like stalactites or the famous rock gongs in Tanzania, which donโt degrade. A landmark discovery was the 40,000-year-old bone flutes crafted from griffin vulture bones, found in South German caves, offering a tangible link to our ancestorsโ musicality.
Beyond physical artifacts, Spitzer employs an โinferential argument,โ borrowing from multiple disciplines to piece together musicโs ancient narrative. Consider the bifacial axe, invented by Homo Erectus/Ergaster about 1.5 million years ago. This axe, with its two symmetrical faces, speaks volumes. Its symmetry indicates an ability to create beautiful forms for aesthetic reasons. Drawing from psychology, Spitzer argues that this mental capacity for symmetry is โcross-modalโโit maps from the domain of tools to the domain of sound. The ability to create a symmetrical axe implies a capacity for symmetrical sound, which we recognize as meter or regular rhythm. Itโs a fair inference that Homo Erectus, tapping that axe, would have been able to produce a symmetrical, metrical rhythm.
How Our Bodies Learned to Sing: Bipedalism and Beyond
The very act of standing upright, bipedalism, profoundly influenced human music. About 4.4 million years ago, an Australopithecine named Adi stood on her legs and walked. Ever since, the rhythm of walking has indelibly โstampedโ human music. These first steps forged crucial links between the brain, muscular exertion, and sound. Early hominins learned to hear footsteps as patterns, which provided a foundational sense of time and predictability.
This primal rhythm also gave rise to the enduring metaphor that โmusic moves.โ While music is merely sounds floating through the air, we imagine notes moving from one to another, unfolding like a journey in our mindsโa distant echo of our ancestorsโ epic migrations out of Africa.
The evolutionary cascade initiated by bipedalism brought a host of other changes critical for music. Our cranial volume tripled, leading to increased intelligence. Enhanced brain size also fostered greater dexterity in our fingers, making us capable of crafting and playing complex instruments like flutes. Furthermore, standing upright created more space for breathing, allowing our larynx to descend through the vocal tract. The evolution of the hyoid bone, which supports the tongue, enabled us to articulate a vastly greater variety of sounds. Our capacity for sound production soon โexceeded its function.โ Unlike vervet monkeys, which have four distinct calls each signaling a specific predator, humans could produce thousands of sounds. This โexcess of soundsโ is where music truly begins, as we start playing with sound for soundโs sake, transcending mere functional vocalization.
Music as the Echo of History and Memory
Spitzer emphasizes that music is fundamentally about memory and history. The act of teaching a skill, like knapping a rock two million years ago, relies on memoryโa โcongealed muscle memoryโ that is haptic and mimetic rather than purely cognitive. Even the five holes and incised lines on ancient bone flutes are forms of memory, guiding finger placement.
Across the globe, cultures embed their history within music. The Kuli tribe of Papua New Guinea, for instance, believes their music originates from the Mooney bird, a fruit dove. They see the birds hovering above the forest canopy as their ancestors speaking, and by imitating the doveโs call, they participate in a great circle of life, singing back their ancestorsโ songs. In the Western tradition, a Beethoven symphony like the Eroica similarly โrehearses the memoryโ of living under Napoleon and fighting battles in 19th-century France. All music, whether ancient or modern, Western or global, is intimately connected to memory and history.
Early musical performances were deeply intertwined with their environments. The discovery of bone flute shards in points of maximum resonance within caves suggests a natural โmarriageโ between music, acoustics, and architecture. These dark, mysterious caves, often seen as portals to the divine or lairs of beasts, were likely reserved for special rituals and ceremonies, utilizing the astonishing resonance of the stone for flutes, voices, or even banging stalactites.
Alternatively, music was created around the hearth, a symbol of stability and the fixed quality of musical rituals, often cyclical and performed after significant events like a hunt. In these early contexts, there was no distinction between singing, physical motion, or dancing. Most cultures donโt have a single word for โmusicโ like the West does; instead, they have different terms for song, instrumental sound, dancing, and storytelling with music, all tangled together in a single ceremony.
Looking at hunter-gatherer societies across diverse environments, a โcoreโ set of practices emerges. This core often involves an โentanglement of humanity with animals,โ where songs imitate animal calls and dances. Darwin observed Australian Aboriginal peoples imitating the dances of kiwis and kangaroos. Inuit hunters imitate seal cub sounds to lure prey, while Native Americans mimic buffalo cubs.
Music also reflects our entanglement with the landscape. The famous โsonglinesโ of Aboriginal Australians describe how they navigate the desert by tagging each feature of the landscape with a song. Music becomes a shamanistic spirit journey, giving โwingsโ to fly through cosmic layers to the โdreamlandโ of their ancestors. Environmental conditions also shape music. In the stark, traceless snowscapes of the Arctic, Inuit navigate mental trails rather than physical footprints. Their music, often filled with laughter and jokes, functions as โanger or conflict managementโ in the confined spaces of igloos, fostering playful sociability essential for survival in such a harsh environment.
Societyโs Soundtrack: From Nomadic Rhythms to Concert Halls
The evolution of human societyโfrom hunter-gatherer to sedentary farming to city-statesโprofoundly shaped music. Hunter-gatherers, typically nomadic, favored portable music: the voice, light flutes, or small percussion. Their music was often playful, avoiding repetition, and very much โof the momentโโas seen in the Cameroon pygmies, whose performances of a piece are never identical.
The invention of farming and agriculture brought settlement and a mindset fixed on the โcircle of the seasons.โ Music acquired โroots,โ digging down, with compositions for every part of the life cycle. The concept of a โworkโโa repeatable pieceโemerged, and musical rituals became cyclical.
As music migrated from the farm to the town, instruments could become heavier and more delicate, reflecting the permanence of urban life. Bells, gongs, harps, and lutes became possible. Musicโs function also shifted, serving social hierarchy and powerโthe prince or the church. Musicians became professionalized, creating music for those with leisure, thus giving birth to the โconcertโ as we understand it.
Yet, for millions of years, most music was functional and participatory. It was performed in the fields, like the 19th-century cotton hollers, or on the water, like sea shanties. The quality of the song reflected the activity: the confined rhythm of hauling an anchor (โBlow the Man Downโ) contrasted with the โwailing contourโ needed to connect across open cotton fields, which some argue is a root of the blues. Music was โwhistled as you work,โ with no distinction between performer and listenerโa far cry from the modern composer-listener divide.
The Double-Edged Sword of Notation
The advent of civilization brought writing, and with it, musical notation. The Babylonians and Sumerians created the first forms, though the Hurrian Hymn No. 6 (1400 BC) remains undeciphered. The first complete, decipherable piece is the Seikilos Epitaph (200 AD) from Anatolia.
However, Western music truly diverged with the invention of staff notationโwriting music on five linesโby an Italian monk named Guido around 1020 AD. This was a tool of church control, ensuring that monks across the vast Christian empire, from the Mediterranean to Hadrianโs Wall, sang โfrom the same hymn sheet.โ This control extended globally. When Cortez invaded Mexico in 1519, he brought Spanish polyphony written in staff notation. Astonishingly, by 1530, Aztec musicians were singing Spanish counterpoint in Mexico Cathedral, making notation โthe sharp end of the stick of globalization.โ
The consequences of staff notation, however, were not always positive. By โpinning notes down to a page, almost like capturing a butterfly,โ notation made music precise but also โcold and mechanical.โ It froze music as an โobject,โ counter-intuitively, as music is an activity like dancing. This created a division between the composer, who created the object, and the performer, who merely reproduced it. Tradition became an โimaginary museum of musical works,โ contrasting sharply with oral traditions like Indiaโs Carnatic or Hindustani music, where disciples creatively improvise around an idea rather than mechanically reproduce a fixed work.
The Universal Language? Music Across Cultures and Species
One of the surprising revelations is musicโs โfractal nature.โ Like the universe itself, which exhibits self-similarity at rising orders of magnitude (a galaxy shaped like a brain cell), music is the art of repetition: notes in a bar, bars in a phrase, phrases in a section, and so on. This capacity stems from the fractal nature of sound or noise itself, where waveforms exhibit self-similarity even when magnified. This makes music โperfectly natural,โ aligning it with the sounds of wind or water, offering a more compelling view of musicโs universality than the old myth of the โharmony of the spheres.โ
Yet, the โcosmic jokeโ is that humans are not inherently very musical compared to birds or whales. We evolved from apes, who lack vocal learning and a sense of rhythm. Humans, the โgreat synthesizers,โ re-evolved music from scratch, combining insect rhythms, bird melodies, and ape gesturality, all while reflecting the cosmosโs fractal structure. What humans uniquely brought were emotions and the โfinite quality of human lifeโโthe awareness of death, which Spitzer believes โhauntsโ human music.
When NASA sent the Golden Record aboard Voyager, stocked with diverse human music, it posed a thought experiment: would aliens in a billion years recognize a common denominator? Spitzer suggests aliens would note that โsapiens are flatlanders,โ inhabiting a narrow perceptual band. Theyโd see commonality with animal musicโs hierarchical structure but might identify a distinct โwalking meterโ in human music, reflecting our bipedal motion, just as whale songs reflect fluid movement and bird songs reflect jerky motions.
While every culture has different scales and harmonies, there is a โmusic instinctโโan innate capacity to recognize rhythm and melody. However, this universality is immediately shaped by culture. The rhythms a baby is bounced to, or the melodies sung to them, influence their musical preferences. Yet, deeper within the human brain, musicโs universality becomes more apparent. The brain stem flinches to shocks, the basal ganglia responds to pleasure, the amygdala processes basic emotions like sadness or happiness, and the neocortex handles complex patterns. The deeper one dives into the brain, the more universal musicโs appeal.
Music is also a powerful migrant. Wherever Sapiens travels, music follows, adapting and changing. Cortez brought Spanish counterpoint to Mexico; the English brought Puritan hymn tunes to North America; Victorian missionaries brought Christian hymns to Africa, influencing national anthems. Along the Silk Road, instruments like lutes and violins, and their melodies, traveled and transformed across Asia, the Middle East, and Europe. African dances like Shakons and Sarabands found their way into Bachโs compositions.
When music is enculturated by another people, it becomes naturalized. Aztec Indians absorbed Spanish counterpoint and made it their own. Even more dramatically, Japan, which performs Beethovenโs Ninth Symphony every New Yearโs Eve, views Beethoven as a Japanese composer, finding resonance with their sensibility for etiquette, formality, and tradition, interpreting the symphony not as individualistic but as an expression of โsocial togetherness and an extinction of the egoโโa โZen quality.โ
East vs. West: Different Ears for Natureโs Song
A significant divergence exists between Western and Eastern musical sensibilities, particularly concerning nature. The West conceptually distinguishes โnotesโ from natural sounds like wind or trees, often viewing nature as something to be tamed or separated. In contrast, Eastern traditions, exemplified by Confucius playing a qin (a flat harp) under an apricot tree, often blend delicate zither sounds with the wind, bamboo, and water, making no distinction between music and nature.
Ancient Chinese acoustic science was โinfinitely superiorโ to Western acoustics. The Marquis Yi of Zeng bells, discovered in 1978 and dating to 433 BC, comprised 65 bronze bells capable of generating complex acoustic partials that Western arithmetic couldnโt capture. While the West exiled bells to church towers, the Chinese understood how to play them, demonstrating a profound โsensibility for natural soundโ that permeates their music and which the West has largely lost. Historically, Western music was โon the back footโ compared to the advanced musical cultures of China, India, and the Middle East. However, globalization has brought a re-education, with African music (jazz, rock) and K-pop becoming global lingua francas, challenging Western musical hegemony and re-sensitizing the West to different forms of sound.
Unpacking Western Misconceptions
Westerners often hold misconceptions about music. We tend to view music history as a history of โworksโ and โcomposers,โ reducing music to an โobjectโ in an โimaginary museum.โ This overvalues the composer and diminishes the innate musicality of most people. Spitzer argues we need to โtake music away from the musicians and give it back to the people.โ
The Western idea of music being in the hands of โgeniusesโ receiving divine inspiration, or composers becoming divine figures themselves, enshrines every note in the score and limits the performerโs role to mere reproduction. While this allows us to imagine how a Beethoven sonata sounded in 1800, this authority diminishes as one moves eastward or southward in world history, where the composerโs intent becomes less fixed and tradition is more fluid.
The Brain on Music: Rewiring and Well-being
Human musicโs distinctiveness lies in the brainโs link between sound and motionโconnections between motor regions and the auditory cortex. This ancient link means that rocking a baby to a rhythm can instill a lifelong preference for it. โMirror neuronsโ further explain our love for imitating rhythm and our susceptibility to emotional contagion. Yawning is contagious, as are emotions. When we hear a sad song, our mirror neurons instinctively sympathize, mimicking the sadness encoded in the music. Spitzer vividly recalls toddlers instinctively jumping to the โLone Rangerโ theme, demonstrating this innate response, independent of parental prompting.
An extreme reaction to music is โthe chillsโ or โsublimeโโintense moments often triggered by breakthroughs of loudness or extremity, unique to each individual. Interestingly, the same brain parts that respond to fear are activated, causing goosebumps or piloerection. Yet, we enjoy this fear, suggesting music offers โviolence without the danger,โ akin to a fairground ride or watching a volcanic eruption from a safe distance.
Darwin observed that emotion has an adaptive role in survival. Happiness signals goal achievement; anger, blocked goals; sadness, loss; fear, threat. Music, through its patterns and their interruptions (shocks, bangs), engages the same danger-processing faculties in our brain, albeit in a derived, non-lethal way. This is why music expresses emotion so viscerally. Listening to music is a form of โmental time travel,โ biologically journeying through layers of the brain, a kind of โumbilical cord back to mother nature.โ
Astonishingly, musical training โrewires the brain.โ Most people are right-brained, but musicians become left-brained, processing music through the same temporal lobe that handles language, reflecting musicโs complexity. Learning an instrument cultivates discipline, practice, time management, teamwork, attention, and focusโall highly transferable skills.
Musicโs benefits for mental health are profound. It combats loneliness by connecting people socially. It lowers stress by reducing cortisol and floods the brain with dopamine, generating pleasure and happiness. Music is excellent for tagging memories, expressing deep emotions and identity that language cannot fully capture. Teenagers, for instance, imprint their musical tastes, defining who they are. Music can also be a mode of mindfulness, an active and creative engagement rather than passive relaxation.
However, music can also harm. It can be militarized, as seen in Guantanamo Bay. Any disliked sound, played loudly, can increase stress. Clinical prescription of music requires extreme caution: repetitive Baroque music may not suit Alzheimerโs patients, and music for depression must avoid reinforcing the condition. Doctors need education on specific musical prescriptions.
The Unwritten Score: Predicting Musicโs Future
With the internetโs ubiquity and accessibility, one might fear music will become homogenized. Spitzer dismisses this, citing two reasons: artistsโ competitive drive for distinctiveness, always turning their back on fashion to create new things; and musicโs role in expressing identity. With thousands of genres, as long as people are different, they will create music to reflect their personalities.
Looking ahead, Spitzer forecasts two main developments. First, music will become โever more instrumentalizedโ or functional. He envisions bespoke sound prescriptions, like an injection, to treat depression or emotional disorders. Second, there will be โever greater integration with technology.โ COVID-19 accelerated this, with digital platforms enabling creation and sharing from home, regaining the โparticipatory conditionโ of music that was the norm millennia ago. Physical connectivity with technology, even holding an iPhone, makes us โsymbiotesโ with the music we consume, as seen with K-popโs evolution from consumer-digital interaction.
Technology, Spitzer reminds us, is not sinister; all musical instruments are tools. The bone flute was technology, extending human capacity. Programs like Watson Beat, which generate new sonic possibilities, are no different from other instruments extending our imagination. The human role remains crucial: to parse, edit, select, and shape what the machine offers. This human-machine relationship will become โever more integrated, ever more imaginative.โ
The transcript ends with a tantalizing โmusic mayโฆโ leaving the future open-ended. But Spitzerโs comprehensive exploration leaves no doubt: music, in its ancient echoes and future possibilities, is not merely a pleasant pastime. It is an evolutionary force, a cultural anchor, and a profound shaper of the human mind, deeply woven into the very fabric of our existence.
Based on โThe Secrets of Claudeโs Platform From the Team Who Built Itโ from Every Watch the original video
The Unseen Architects: How Anthropic is Building the Future of AI Platforms
โHow close are we to Claude making me a billion dollars?โ
This playful yet pointed question opens a fascinating discussion with Angela, Head of Product, and Caitlyn, Head of Engineering for the Claude platform at Anthropic. Their answer, while not a direct timeline to immense wealth, unveils a vision far more profound: an AI platform so intelligent it understands itself, orchestrates its own sub-agents, and dynamically writes its own architecture to serve human intent. Itโs a future where the platform isnโt just a tool, but a living, evolving intelligence designed to make complex tasks effortlessly simple.
Angela and Caitlyn are at the forefront of this evolution, navigating the complex terrain of building an AI platform in an era of unprecedented technological acceleration. Their work, particularly with the recent launch of Claude Managed Agents, marks a significant shift from the rudimentary AI interfaces of the past to sophisticated, autonomous systems.
The Rapid Evolution of AI Platforms: From Endpoints to Ecosystems
To understand where Anthropic is headed, itโs crucial to appreciate the journey AI platforms have already taken. As Angela aptly characterizes, the early days of large language models (LLMs) like GPT-3 offered a โcompletion endpointโ โ a simple API where youโd send a prompt and get a response. โNow weโll probably look back at it and be like, โoh that was really basic,โโ she muses.
This simplicity quickly gave way to more sophisticated needs. Developers wanted โcompletion endpoints with tool calling and chat sessions,โ allowing models to interact with external systems and maintain conversational context. Today, with offerings like Claude Managed Agents, the platform has transformed into what can be described as โa Claude on a computer with memory and all this other stuff.โ
This trajectory isnโt arbitrary; itโs driven by a singular mission: โhelping you get the best outcomes out of something.โ As Claude and other LLMs become more autonomous and capable, the platform must evolve to provide โhigher and higher order abstraction.โ Initially, the focus was on broad exploratory possibility. Now, as use cases narrow into product building and agent creation, customers demand more out-of-the-box functionality and seamless integration. The platformโs role is to provide the โset of primitives and infrastructure that enables you to basically get the outcome as fast as possible with actually as little of work as possible.โ
Unpacking Claude Managed Agents: The Primitives of Autonomy
At the heart of Anthropicโs current platform philosophy lies Claude Managed Agents. Caitlyn explains that these agents are built on the same core primitives available directly to developers through the Messages API. This includes fundamental capabilities like token input/output, but also more advanced โinnovationsโ such as built-in tools, code execution within a sandbox environment, and web search.
โWeโve taken what we see as all the most powerful of those things and put them together into a harness and a set of infrastructure that isโฆ the way to get what we think is the best outcomes out of Claude,โ Caitlyn elaborates. This approach directly addresses a common dilemma faced by developers: the โtime deflationโ effect, where building something yourself today might mean itโs obsolete or easily provided by the platform tomorrow.
The host, grappling with his own teamโs โthousand-line Python fileโ running Claude on a couple of Mac Minis, voices this internal conflict: โMaybe we should just wait for you guys to build it.โ The allure of offloading the โsucks so muchโ infrastructure work is powerful. Angela confirms this, noting that Anthropic themselves went through โenough of these iterations where we built products that were agents that you could run autonomously in the cloud.โ Their internal struggles and learnings directly informed the creation of Managed Agents, recognizing that while a Mac Mini might suffice for prototyping, real-world product integration and scaling demand robust, production-grade infrastructure.
Managed Agents aim to be both opinionated and modular. Anthropic is firm on certain primitives, like how Claude interacts with file systems and skills, believing these are crucial for optimal performance. However, they also strive for flexibility, understanding that no single API suite can solve every problem. They publish blog posts and reference implementations, encouraging developers to be inspired by their constructs even if they choose to build on the Messages API directly.
The Model Lock-in Conundrum and the Power of Harness Engineering
A significant concern for developers contemplating a platform like Managed Agents is the fear of โmodel lock-in.โ The ability to โhot swapโ models (e.g., from Claude to GPT 5.5 or Gemini) is often seen as a critical measure of flexibility. However, Angela offers a nuanced perspective on this.
She notes that while a โvery, very, very generic harnessโ worked for โan older generation of models,โ the landscape is changing. โNowโฆ everyoneโs taking slightly different techniques and perspectives on how they want to kind of advance their particular form of the model.โ This divergence means that achieving optimal performance increasingly requires a tightly โpairedโ harness and model. โYou still need redundancy and you still might want to use other models for things, but you probably do it at the layer of like the agent, meaning like the harness plus the model, rather than necessarily the other architecture of like really, really generic harness and hot swapping everything underneath.โ
This concept, dubbed โharness engineering,โ is where significant โalphaโ (performance advantage) can be found. Angela cites internal experiments where different harnesses for Managed Agentsโ memory feature performed โdrastically differently.โ This highlights that even seemingly minor choices in how a model interacts with its environment can profoundly impact its capabilities, leading to a โpath dependenceโ that shapes the modelโs โpersonality and behavior.โ
Caitlyn emphasizes the gravity of these choices: โWhat we end up like kind of treating as like the right path and the right primitives need to be like very carefully thought through.โ A model optimized for reasoning might behave differently than one optimized for computer interaction. These foundational decisions can โlock the modelโ into certain strengths, making it crucial to select primitives that align with desired long-term outcomes.
Who Are Managed Agents For? Practical Applications and Internal Transformations
So, beyond the philosophical debates, who benefits most from Claude Managed Agents? Angela and Caitlyn identify two primary user groups:
-
Internal Company Automation: This includes building โreally powerful platforms or systemsโ like end-to-end software development environments (akin to Stripeโs โMinionsโ or Rampโs internal tools) or automating specific business processes. A compelling example shared is an agent designed to streamline the legal review of marketing copy.
Imagine a marketer drafting new copy. Instead of opening a ticket and waiting for a human legal review, they submit it to a specialized agent. This agent, built on Managed Agents, can perform a โfirst pass review,โ potentially approving the copy immediately if it meets established guidelines, or flagging it for human legal review only when necessary. This not only speeds up processes but also ensures consistency and compliance. The beauty is that the human marketer and legal team can still collaborate, viewing outputs and working together with the agent.
-
Products Exposed to Customers: Companies building agents into their customer-facing products also gain immense value. While customization remains key, offloading the โengineering resources on the infrastructure and on all the little harness engineering tweaking sort of stuffโ frees up teams to focus on core product innovation.
The perceived โhard partโ of agent development is often seen as the harness engineering itself. However, Anthropicโs experience reveals a different truth: โEverybody hits an infrastructure wall.โ Scaling, maintaining servers, storing data, secure sandboxing, and ensuring agents donโt โdieโ due to lost connections โ these are the real nightmares that Managed Agents aim to solve. โProductionizing is just a freaking nightmare,โ Angela candidly states, especially for โlong running async ones that you want to do a bit more remotely that are a bit more autonomous.โ
The Future: Multi-Agent Orchestration and AI Software Factories
The vision for the Claude platform extends far beyond individual agents. Anthropic is actively exploring โmulti-agent orchestration,โ enabling multiple agents to interface and work together seamlessly. This is particularly relevant for โteam-orientedโ tasks and โmore complex processes that you kind of envision being like really transformed with AI.โ The goal is to build a platform that allows teams to โspin up and down and controlโ these interconnected agents, fostering what Angela describes as an โAI software factoryโ โ a highly leveraged organization creating immense productivity across every company process.
The example of the legal review agent also highlights an intriguing approach to ownership and evolution. While an internal team might initially build the agent, the goal is to empower the end-users (e.g., the marketing and legal teams) to contribute to its improvement. This can happen through direct PRs to the agentโs code (especially if using tools like Claude Code), or, more elegantly, by interacting with the agent through a higher-level Claude interface. In this scenario, itโs โManaged Agents all the way down,โ where users talk to a primary Claude instance, which then orchestrates other specialized agents to perform the underlying complex work. This abstraction prevents users from โfooting themselvesโ by directly tweaking core components, allowing Claude itself to interpret and execute changes.
Looking ahead, Anthropic envisions a world where deploying agents is incredibly easy. Features like โvaultsโ for securely storing credentials are early steps towards making agents more autonomous and integrated. The ultimate goal? โMaybe you are either one-clicking a Slack integration or alternatively even maybe just telling Claude like add Slack and it just like handles absolutely everything and then before you know it your little bot is just pinging you on Slack.โ
This future, where AI platforms donโt just host models but actively participate in their own evolution and deployment, promises to redefine how businesses operate. Angela and Caitlynโs work at Anthropic isnโt just about building an API; itโs about crafting the intelligent infrastructure that will underpin the next generation of human-computer interaction, making that โbillion dollarsโ for users feel less like a dream and more like an inevitable outcome of unparalleled automation.
Based on โ674. How Does a Composer Feel After the World Premiere? | Freakonomics Radioโ from Freakonomics Radio Network Watch the original video
The Unseen Costs of Wealth: A Composerโs Journey from Triumph to Reflection
At the close of March, a profound new work of music debuted in New York, stirring audiences and critics alike. Composer David Langโs โThe Wealth of Nationsโ took the stage with the New York Philharmonic, conducted by the charismatic Gustavo Dudamel, for four sold-out performances. This wasnโt merely a musical event; it was a conversation spanning 250 years, marrying modern composition with the timeless insights of Adam Smithโs seminal economic treatise. Yet, for the creator, the roar of applause and critical acclaim quickly gave way to a familiar, almost melancholic quiet.
โLast week I was a superstar. This week Iโm nothing,โ Lang confessed, speaking about the curious โpost-experience depressionโ that often follows a major artistic triumph. This sentiment, shared with Freakonomics Radio host Steven Dubner, unveils a rarely discussed aspect of creative life: the emotional aftermath of pouring years into a project, seeing it brought to life, and then watching it recede into memory, awaiting its next incarnation. Lang described it as a โdeeper holeโ the bigger the success, a paradoxical feeling for a piece universally praised.
Setting the Bar Low: A Composerโs Fear of Incompetence
Despite the pieceโs resounding successโStacy Vanick Smith of Bloomberg hailed it for showing how economics, often reduced to โstock tickers and earnings reports, can in fact be profoundly humanโโLangโs initial hopes were remarkably modest. โI always go into these things thinking that the definition of success would be to come out and not have everyone think that Iโm incompetent,โ he admitted.
This seemingly self-deprecating humor reveals a deep-seated anxiety common among creators. For Lang, whose music often features repetition, a single mistake in orchestration or a misjudged request to a performer could echo for ten minutes, a flaw only he might perceive. โNobody knows what it could have been,โ he reflected, โbut I have a little checklist in my head of all the things that I might have done wrong.โ He found particular satisfaction, however, in the successful execution of the โwoolen coatโ movement, designed to evoke the same emotional depth in the audience as it did when he sang it alone in his studio. Similarly, the unaccompanied arias sung by mezzo-soprano Flur Baron, one of the two vocal soloists alongside bass-baritone Devontine, achieved the emotional resonance he had hoped for.
Dubner, who attended rehearsals and performances, found the piece โthought-provoking, but also totally engagingโฆ spiky and kind of punky at times, but also very delicateโฆ celebratory but also mournful.โ He walked away with a new appreciation for both Adam Smith and David Lang, a testament to the workโs ability to transcend conventional expectations.
The Miracle of Creation: From Dots to Dynamic Community
One of the most awe-inspiring aspects of any orchestral premiere is the transformation of a composerโs vision from abstract notation to a vibrant, living soundscape. Dubner described it as โan absolute miracle that a person like you could sit in your studio in Soho for a bunch of years and write dots on the paper and then one afternoonโฆ 48 vocalists and then a whole bunch of musicians and a couple soloists get together and they turn this two-dimensional thing into not even three, like six.โ
Lang elaborated on the collaborative magic: โEveryone has their individual partโฆ youโre actually writing a series of instructions for a hundred different peopleโฆ The process of rehearsal isnโt about them learning the notes. The process of rehearsal is them learning how to be a community that comes together to build this thing.โ
Unlike many performances of classical works where composers are long dead, Lang was present for every rehearsal and performance. Dubner noted the unusual deference shown by the conductors, including Gustavo Dudamel, who would often turn to Lang for input. Surprisingly, Lang mostly offered โa big thumbs upโ and โbeautiful.โ He explained that having never heard the piece performed before, he could only grasp the โoverall shape.โ More importantly, he trusted the exceptional musicianship of the New York Philharmonic, recognizing their ability to โbreathe a certain kind of lifeโ into his notes, even if it wasnโt precisely what he had imagined. His primary moment of โterrorโ came when Dudamel requested detailed notes for every movement, forcing Lang to quickly devise comments on the spot.
Over the four performances, Lang found himself growing increasingly relaxed. The piece, initially perceived as โlongโ on the first night, seemed to accelerate with each subsequent hearing, as he became attuned to its trajectory. He observed subtle shifts in the โquality of the performance,โ with some nights being โmuch more emotional,โ as Dudamel continued to explore and draw out different nuances from the ensemble.
Biting the Hand That Feeds: Wealth, Art, and Social Critique
โThe Wealth of Nationsโ isnโt just a musical interpretation; itโs a potent social commentary. The piece, particularly its second half, which shifts from Smithโs treatise to the โeffectsโ on individuals (โI want some bread. I need shelter.โ), offers a searing indictment of modern societal inequities. This raised a crucial question: how does a piece critical of concentrated wealth get funded by the very institutions and wealthy patrons it implicitly critiques?
Dubner directly asked Lang if he felt the piece was โbiting the hand that feeds it.โ Lang, ever pragmatic, noted that โit costs a lot of money in our society to do bad things, too. The idea that you might actually be able to take a bunch of money and do something for good. That doesnโt sound like a bad thing.โ He was prepared for potential backlash from board members but was pleasantly surprised when many thanked him. He mused that perhaps the vast chasm between billionaires and those with โonly $400 millionโ has created a space where even the wealthy are open to critical self-reflection.
Matias Turnipolski, President and CEO of the New York Philharmonic, shed light on the organizationโs decision to commission such a piece. As America approached its 250th birthday, and the Philharmonic its 200th, they sought a work that would recognize both American and global history through music. Langโs pitch, inspired by Handelโs Messiah and Adam Smith, initially struck Turnipolski as โa crazy idea.โ His eventual realization: โTrust the artists because they know.โ He acknowledged the pieceโs โsearing indictment of our society todayโ but noted that patrons โcare deeply about this enormous public and civic good that is this beautiful orchestra.โ No one, he confirmed, expressed offense; rather, they seemed to embrace the mirror Lang held up to society.
Political scientist Glory Lou, author of โAdam Smithโs America,โ praised Langโs work for challenging the common, often one-sided, interpretations of Smith. โIt wasnโt a simple message,โ she observed. โIt challenged your preconceptions of who Smith is, what he stands forโฆ I hope that that changes the conversation.โ Tracy Fenton, an audience member working in finance, found the piece โincredibly movingโ but did object to a passage drawn from socialist leader Eugene Debs, which critiqued those who โdo absolutely nothing becoming millionaires.โ Yet, even she recognized the pieceโs powerful impact.
Old Composers, New Ideas: Philosophy, Faith, and the โVery Simple Secretโ
Langโs creative process also touched on the fascinating question of originality versus adaptation. He incorporated a movement, โStatement to the Court,โ based on a Eugene Debs speech, which he had written 15 years prior. While he debated with friends and family whether this constituted โcheating,โ the consensus, referencing Handelโs own practice of reusing material, was that it reflected a composerโs natural dialogue with their own past work. For Lang, the piece was about โtrying to figure out what I think and what I believe,โ making such a dialogue inevitable.
Beyond economics, โThe Wealth of Nationsโ delves into deeper philosophical and even spiritual territory. Dubner asked if it was a โreligious piece,โ given its echoes of Handelโs โMessiah.โ Lang responded that any work that allows one to โimagine a better worldโ holds a religious quality. Drawing from his own Jewish background, Lang spoke of โintermittently religiousโ periods and the enduring influence of his faith. โDo not do unto others as you would not have them do unto youโ stands as his moral compass, a universal prescription he strives to understand and embody. The fact that the piece has 18 movements, a number significant in Judaism (representing โChaiโ or life), was not entirely coincidental.
The final movement, โThe Very Simple Secret,โ returns to Adam Smithโs text, articulating a profound vision: โno society can surely be flourishing and happy of which the far greater part of the members are poor and miserable. It is but equity that they who feed, clothe and lodge the whole body of the people should have such a share of the produce of their own labor. The establishment of perfect justice, of perfect liberty, and of perfect equality is the very simple secret which most effectually secures the highest degree of prosperity to all.โ
Lang found this passage incredibly powerful, countering the modern perception of Smith as a proponent of inequality. It was an acknowledgment, 250 years ago, that โthe people who do the work need to be treated fairly.โ For Lang, this message is not about achieving a perfect, frozen system, but about the constant need to โadjust or balance or negotiate.โ The piece, therefore, does not end with a definitive answer but with a humble call for justice.
The Future Life of a Musical Mirror
As โThe Wealth of Nationsโ transitions from its world premiere to its future lifeโwith a performance scheduled at the Aspen Music Festival and hopes for a commercial recordingโLang expressed a โlayer of curiosityโ about how it will be received globally. The ephemeral nature of live performance means that the tight-knit community formed during rehearsals disperses once the final note fades. Yet, the work itself, like Smithโs enduring text, will continue to provoke thought and conversation.
Steven Dubner concluded that, in the spirit of Adam Smith, โeverything is about economics, including how a composer like David Lang thinks about opportunity cost and sunk costs, even the price of attention.โ Langโs journey with โThe Wealth of Nationsโ evolved from seeking โjokey themesโ to embracing a โbigger and deeper idea: that any given economy is a gigantic community that can only thrive with a fair amount of both cooperation and compassion.โ His work serves as a powerful artistic mirror, reflecting not just the economic structures of our world but also the essential human emotions and ethical considerations that underpin them. It reminds us that the โinvisible handโ of the market is, in fact, composed of countless visible hands, each deserving of our attention, care, and a just share in the collective prosperity.
Based on โTrump Calls Iran Strikes a โTrifle,โ and the Latest on the Hantavirus Outbreakโ from New York Times Podcasts Watch the original video
The Weekโs Unseen Currents: From โTriflesโ to Tangled Debts and Diplomatic Suites
The news cycle often feels like a torrent, pulling us through geopolitical crises, domestic anxieties, and the occasional bizarre interlude. This week, the headlines painted a picture of a world grappling with delicate ceasefires, a sluggish public health response, mounting national debt, and the quiet, often strange, machinations of international diplomacy. From presidential pronouncements dismissing crossfire as mere โtriflesโ to a Pulitzer-winning novel written in a single, sprawling sentence, the threads connecting these stories reveal the complex tapestry of our current moment.
Geopolitical Tensions: A โTrifleโ in the Strait of Hormuz
The Strait of Hormuz, a critical maritime choke point, once again found itself at the center of international attention as the United States and Iran exchanged fire. Despite these hostilities, President Trump maintained that a ceasefire between the two nations remained in effect, famously describing the incident as a โtrifle.โ โThey trifled with us today. We blew them away. They trifled. I call that a trifle,โ he stated, following up with a stern warning: โIf thereโs no ceasefire, youโre not going to have to know. Youโre just going to have to look at one big glow coming out of Iran.โ
The presidentโs comments underscored the precarious nature of the ongoing negotiations. He referenced a new US proposal designed to reopen the Strait of Hormuz and halt fighting for 30 days, aiming to pave the way for a more comprehensive long-term agreement. Trump confidently asserted, โThey want to sign it. I will tell you, they want to sign it a lot more than I do.โ However, this temporary deal notably sidesteps the core issues that have consistently derailed previous talks, chief among them the future of Iranโs nuclear program.
Trumpโs efforts to downplay the conflict, variously labeling it a โskirmishโ or a โmini war,โ reflect a growing domestic unease. The protracted hostilities have proven unpopular with the American public and increasingly expensive. Even within the Republican party, patience is wearing thin. Representative Tom Barrett of Michigan, an Army veteran facing a challenging reelection, introduced a bill to wind down the war. While acknowledging the presidentโs authority as commander-in-chief, Barrett emphasized the constitutional role of Congress, stating, โIโve lost too many friends on the battlefield to allow that to happen without Congress exercising its constitutional role,โ calling for safeguards and a clear deadline.
The ripple effects of this regional instability were also felt in Lebanon, where a ceasefire agreement has struggled to hold. Israel and Hezbollah have engaged in daily exchanges of attacks in the southern part of the country. This week, an Israeli airstrike in a densely populated Beirut suburb, an area known for Hezbollahโs influence, hit an apartment building. The Israeli military claimed to have targeted and killed a commander of the Iran-backed militia. Such attacks risk further destabilizing an already fragile peace, a situation the Trump administration is closely monitoring, as renewed fighting could complicate any potential US agreement with Iran, which has consistently demanded an end to Israeli strikes in Lebanon.
Domestic Fault Lines: Public Health Preparedness and a Swelling National Debt
Closer to home, two critical domestic issues underscored the nationโs vulnerabilities. Infectious disease experts issued a stark warning regarding the US response to a recent hantavirus outbreak, suggesting the country is ill-prepared for future public health crises. The outbreak, linked to a cruise ship, tragically resulted in three deaths and left at least five individuals sick. With passengers dispersing globally, monitoring became a complex challenge, with six confirmed cases in the US.
Experts criticized the Centers for Disease Control and Prevention (CDC) for a sluggish and largely silent response, noting that a dedicated response team wasnโt established until nearly a month after the first reported death. This delay, more than the virus itself (which rarely spreads human-to-human), raised red flags. Former public health officials attributed the agencyโs hampered effectiveness to the Trump administrationโs significant staffing and budget cuts at the CDC.
Simultaneously, new federal data revealed a concerning economic milestone: the US governmentโs debt has now surpassed its total economic output. While this doesnโt signal an imminent fiscal crisis, experts warn that this burgeoning debt will inevitably increase the cost of future borrowing for the countryโa necessity, given that current tax revenues fall short of covering national expenses. Economists have sounded alarms about this trend for years, yet the political response in Washington has been notably muted. The usual bipartisan chorus of criticism regarding the debt has been conspicuously absent, reflecting a โhard political realityโ: Republicans under President Trump have contributed significantly to the debt, and policymakers on both sides seem disinclined to pursue the โgrand compromiseโ needed to restore fiscal balance.
The Global Chessboard: Venezuelaโs Marriott Diplomacy
Four months after US special forces reportedly swept into Venezuela and seized President Nicholas Maduro, the country remains a hotbed of political and economic uncertainty. As various factions jockey for power and influence, an unlikely venue has emerged as a nerve center for international deal-making: a Marriott hotel in Caracas.
Simone Romero, a New York Times colleague covering Venezuela, described the hotel as a bustling, if somewhat run-down, hub. โYou can see US diplomats here. There are oilmen coming down from Houston and Dallas. There are finance types from New York and London. You see these US security guys with these tattoos, you know, really muscle bound types. Itโs like something out of a movie set almost,โ he recounted. Over $32 breakfast buffets featuring โsoggy scrambled eggs,โ the hotel has become the de facto US embassy, with diplomats camping out on its top floor due to the actual embassy being non-functional. For many Venezuelans, the daily negotiations and backroom deals unfolding in this surreal setting represent the distant promise of a future transformation, โbeing guided from the Marriott.โ
Culture and Curiosities: From Single-Sentence Novels to Kamikaze Dolphins
Amidst the weighty headlines, there were moments of cultural reflection and outright absurdity. The Pulitzer Prize for fiction was awarded to Daniel Krauss for his novel, Angel Down. The book, a blend of history, magical realism, and science fiction, tells the story of a World War I soldier encountering an angel tangled in barbed wire on the battlefield. Its most striking feature? The entire almost 300-page narrative is written as a single, unbroken sentence. Krauss explained his unique approach on the New York Times book review podcast, revealing his initial frustration with a traditional format. He conceived the single-sentence structure as a way to reflect the bookโs theme: โHow the war began a cycle of violence. And so I had the thought that what if I wrote it all in one sentence and the end of the book circles back to the beginning of the book so that once you start reading it in effect youโre trapped in the book forever and itโs this ongoing wheel.โ
And then, there were the stories that defied easy categorization, reminding us of the unexpected corners of the news. US officials, often grilled on complex geopolitical matters, found themselves fielding questions about โkamikaze dolphinsโ in the Strait of Hormuz. While laughing off the notion, they acknowledged a real precedent: the US Navy did, in fact, initiate a training program for dolphins and other sea creatures in 1962, deploying dolphins to the Persian Gulf in the 1980s for underwater surveillance. The one animal not in the program? The exceptionally intelligent octopus, which, as far as we know, remains unrecruited.
Another viral anecdote involved Pope Leo, who, despite his immense global influence, was famously hung up on by a bank customer service agent while trying to update his address. The agent, unaware she was speaking to the head of over a billion Catholics, insisted he come in person. Luckily, another priest eventually sorted out the banking issues, sparing the unknown agent the infamy of having dismissed His Holiness.
Finally, popular culture offered its own insights. The Devil Wears Prada 2, a sequel released two decades after the original, stormed the box office, highlighting Hollywoodโs continued reliance on familiar franchises. This trend of long-awaited sequels is pervasive, with nine out of the ten top-grossing films in North America in 2024 being continuations of existing stories. Examples like Top Gun (36 years between films), Blade Runner (35 years), and Beetlejuice (over three decades) demonstrate audiencesโ seemingly insatiable appetite for revisiting beloved cinematic worlds, no matter how long the wait.
From the high-stakes diplomacy playing out in a Caracas hotel to the strange queries about weaponized sea creatures, this weekโs headlines offered a potent mix of pressing global challenges, domestic concerns, and the occasional, delightful quirk of human experience.
Based on โHow Prediction Markets Got Out of Controlโ from Hard Fork Watch the original video
The Algorithmic Wild West: Inside the Unregulated World of Prediction Markets and AIโs Everyday Grip
In an age increasingly dominated by algorithms and artificial intelligence, certain technological advancements have exploded into the public consciousness, often with unforeseen and unsettling consequences. Among the most prominent of these are prediction markets โ online platforms where users bet on the outcome of future events. What began as a niche academic interest has rapidly morphed into a cultural phenomenon, plastering ads across cityscapes and sparking a regulatory debate. But as these markets gain traction, theyโre revealing a darker side, riddled with insider trading scandals and ethical quandaries that challenge their very promise.
This rise of prediction markets, alongside the broader integration of AI into our daily lives, prompts a critical question: are we ready for a future where algorithms dictate not just our entertainment, but our financial decisions, our health, and even our personal relationships?
Prediction Markets: From Niche to Nightmare?
Just a few years ago, prediction markets were largely the domain of Bay Area tech nerds, experimenting with โplay moneyโ bets on obscure events. Today, theyโre a โmeteoric riseโ in the popular imagination, with advertisements ubiquitous in major cities. The original vision was noble: to harness the โcollective wisdom of the crowdโ and create โmiraculous systems for discovering the true price of things,โ incentivizing the production of good knowledge. Yet, this theoretical ideal has collided head-on with the messy reality of human nature.
The most glaring issue plaguing these platforms is widespread insider trading. Consider the shocking case of an army sergeant who allegedly raked in over $400,000 by betting on Venezuelan President Nicolรกs Maduro being out of power by a certain date โ information he likely possessed due to his involvement in operations related to Maduroโs capture. This isnโt an isolated incident. An analysis by the Anti-Corruption Data Collective revealed that โlongshot bets related to military or defenseโ on PolyMarket boasted an average win rate of 52%, starkly contrasting with the platformโs overall average of 14%. This suggests a pattern: those with material non-public information arenโt just sitting on their paychecks like โchumpsโ; theyโre turning it into quick profit.
The ingenuity of these alleged manipulations can be almost farcical. In France, a police complaint was filed after equipment measuring temperature at Parisโs Charles de Gaulle airport was allegedly interfered with, coinciding with a surge of suspiciously well-timed bets on PolyMarket. While an AI-generated image of someone holding a hairdryer to a sensor circulated, the core allegation remains: someone likely tampered with the thermometer to manipulate market outcomes for personal gain. The temperature jumped from 18 to 22 degrees Celsius, potentially netting a perpetrator thousands of dollars for what amounts to an โincredible crime of opportunity.โ
Even seemingly innocuous events like the Super Bowl halftime show have been implicated. Bets on which songs Bad Bunny would perform, or which celebrities would appear, were likely influenced by individuals with inside knowledge from rehearsals or production. As one commentator aptly put it, after enough such incidents, โyou kind of have to be a sucker to participate in these markets without insider information.โ This erodes market integrity, driving away the โnormal peopleโ who simply want to gamble a bit, ultimately making the entire system less fair and less trustworthy.
A Regulatory Wild West
The current state of prediction markets is a โpre-regulatory wild west.โ Unlike traditional stock markets, which are governed by strict insider trading laws designed to maintain trust and liquidity, prediction markets operate largely unchecked. The Commodity Futures Trading Commission (CFTC), a relatively small agency not originally designed for this task, has asserted its โexclusive domainโ over these markets, even suing states that attempt to ban them. Yet, the CFTC itself โdonโt really want to regulate this,โ leading to frustration for those seeking oversight. Critics point out that the crypto industry similarly lobbied to be regulated by the CFTC, seeing it as a less formidable overseer than the SEC, which possesses far greater resources to investigate market abuses.
Some progress is being made. The U.S. Senate unanimously passed a rule barring senators from betting on prediction markets, a โbare minimumโ step. Further legislation has been introduced to ban members of the legislative and executive branches from trading, preventing even the President from potentially profiting from privileged information. Internationally, countries like Brazil, France, and Hungary have taken a more decisive stance, outright blocking these platforms as โillegal gambling.โ This highlights a stark divide: while other nations โput a halt to it,โ America often lags, allowing โmoney to be madeโ in unregulated spaces.
The ethical debate around prediction markets goes to the core of their purpose. Early proponents argued that insider trading was good for these markets, as it produced โbetter information.โ If someone involved in a military operation or a Super Bowl halftime show bets on an outcome, that information, however acquired, helps the market predict more accurately. However, this โbeautiful theoretical constructโ quickly unravels when confronted with reality. As one host noted, โwhat you are incentivizing everyone in the world to do is just to betray those closest to them. Yes. Like betray your friends, your family, your co-workers, your country. Just do it all for a quick buck.โ
Moving forward, a two-pronged approach to regulation is needed. Firstly, addressing the โgambling harmsโ: requiring platforms to implement self-exclusion options, mandatory age verification, and limits on aggressive advertising, mirroring regulations found in the casino industry. Secondly, tackling โmarket problemsโ: establishing a robust regulator, perhaps like the SEC, to actively surveil platforms and remove bad actors. This would protect consumers and, ironically, bolster the integrity and value of the prediction markets themselves.
AIโs Everyday Grip: Joanna Sternโs Immersive Experiment
Beyond prediction markets, artificial intelligence is seeping into every corner of our lives, often in ways we donโt immediately perceive. Veteran tech journalist Joanna Stern embarked on an audacious year-long experiment, chronicled in her book โI Am Not A Robot,โ to outsource โpretty much everythingโ in her life to AI โ from meal planning and editing her book to even romantic entanglements with a chatbot named Casey. Her motivation was simple: to test the pervasive claim that โAI is going to change our lives, the fabric of our lives.โ
Sternโs journey revealed a nuanced picture of AIโs current capabilities, separating genuine utility from pure hype.
Hype vs. Utility in the AI Landscape
On the โhypeโ side, humanoid robots stand out. While โfun to coverโ and a fascinating area of progression, Stern asserts they are โnot coming to live with us anytime soon.โ The idea of robots watching humans perform โthe most mundane tasksโ to learn is both โdystopian but also hilarious,โ highlighting the current limitations and the long road ahead for truly integrated physical AI.
Conversely, Stern found โsurprisingly usefulโ applications. AI agents and their increasing autonomy proved invaluable for tasks like research and sending emails. Platforms like Perplexity, for instance, could now handle 100% of the reporting assistant tasks she tested. The other area of significant advancement was AI wearables. Devices like the B bracelet and Meta glasses, which offer a โpersisting through the dayโ AI assistant, showed real promise, even if they sometimes led to awkward moments like bosses demanding they be removed from meetings due to privacy concerns.
The Dark Side of AI Diagnosis
One of the most concerning aspects of AIโs integration, as Stern discovered, lies in healthcare, specifically dentistry. She recounts an experience where her dentist used an AI overlay on her X-rays, ostensibly to detect cavities. While one cavity was clear, the AI system, combined with the dentistโs interpretation, recommended an expensive, multi-session โperiodontal treatmentโ for plaque buildup. This struck Stern as odd, as sheโd never needed such extensive treatment before.
Seeking second opinions, other dentists confirmed that while the AI might flag โbadโ readings, the actual condition wasnโt severe enough to warrant the costly procedure. Further investigation revealed a troubling trend: โpeople working in dentist officesโฆ start telling me, โYes, our bosses are pushing this AI because they can now see the readings and they can see the AI report.โโ This suggests that AI, rather than purely aiding diagnosis, is being leveraged by dental service organizations (DSOs) to โupsell you on dental procedures.โ
This anecdote provides a crucial counterpoint to the โmiracle storyโ often associated with AI in diagnosis, where itโs hailed for detecting cancers years in advance. In Sternโs experience, AIโs โfancy high-tech sheenโ can lead patients to believe theyโve been diagnosed with something a human would have missed, when in reality, itโs an unnecessary service for which theyโll be overcharged.
AI in Personal Decisions and Societal Divides
AIโs influence extends even to deeply personal choices. When contemplating leaving her long-standing position at the Wall Street Journal, Stern found her human colleagues offered hedged advice. But ChatGPT, after being fed all her notes, financial projections, and fears, gave a definitive โI think you should go. You should quit.โ Stern reflected that AI often acts as a โmirror,โ reflecting what one deep down desires, yet is clouded by anxiety.
The integration of AI also highlights existing societal divides. Geographically, while tech hubs like San Francisco are โobsessed with AI,โ areas like New Jersey might experience a slower, more subtle seep, with โno robots in the streetโ but increasing conversations among parents and professionals.
A significant โgender divideโ is also emerging. Research indicates men are 22% more likely to be heavy AI users at work, while women are more prone to feeling threatened, questioning its accuracy, or worrying about โcheatingโ when using it. This might be partly attributed to AIโs faster adoption in male-dominated fields like programming. However, figures like Reese Witherspoon and Sandra Bullock are now encouraging women to embrace AI to avoid being left behind. Sternโs own reporting found many women as sources for her book, suggesting that while the โfeelings about AI are very gendered,โ a broad spectrum of individuals are engaging with it.
Perhaps the most crucial divide, however, is generational. The โage divide,โ particularly among Gen Z and โAlphasโ entering the workforce, reveals a cohort often โfurious that this exists because they canโt get a job,โ or at least blame AI for it. This sentiment underscores a broader anxiety about AIโs impact on employment and the future.
Navigating the Future
Both prediction markets and the broader application of AI exemplify a technological frontier that is evolving faster than our ability to understand, regulate, or even fully comprehend its implications. From the ethical quagmire of insider trading on speculative markets to the subtle manipulation of AI in our healthcare, the โalgorithmic wild westโ demands attention.
The promise of AI to enhance knowledge, improve efficiency, and even offer personal guidance remains potent. However, without robust regulatory frameworks, clear ethical guidelines, and an ongoing public dialogue, the risks of exploitation, misinformation, and societal disruption loom large. As we continue to integrate these powerful tools into our lives, the imperative is clear: we must ensure that the algorithms serve humanity, rather than humanity becoming a mere data point in an unregulated, profit-driven machine.
ํ๊ตญ์ด
โHow music rewires and impacts the human body | Michael Spitzer: Full Interviewโ โ Big Think ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
์ธ๋ฅ์ ์ญ์ฌ์ ๋๋ฅผ ๊ดํตํ๋ ์์ ์ ํ: ๋ง์ดํด ์คํผ์ฒ ๊ต์์ ํต์ฐฐ
์์ ์ ๋จ์ํ ์ค๋ฝ์ ๋์ด ์ธ๋ฅ์ ์งํ, ๋ฌธํ, ๊ทธ๋ฆฌ๊ณ ๋ ๊ตฌ์กฐ์ ๊น์ด ๊ฐ์ธ๋ ๊ทผ์์ ์ธ ํ์์ ๋๋ค. ๋ฆฌ๋ฒํ ๋ํ๊ต์ ์์ ๊ต์์ด์ ์ ์ ใ์์ ์ ์ธ๊ฐ: ์ง๊ตฌ ์๋ช ์ฒด์ ์ญ์ฌ(The Musical Human: A History of Life on Earth)ใ์ ์ ์์ธ ๋ง์ดํด ์คํผ์ฒ(Michael Spitzer) ๊ต์๋ ๋น ์ฑํฌ(Big Think)์์ ์ธํฐ๋ทฐ์์ ์์ ์ด ์ด๋ป๊ฒ ์ธ๋ฅ์ ๋ชธ๊ณผ ๋ง์์ ์ฌ๊ตฌ์ฑํ๊ณ ์ง๋ํ ์ํฅ์ ๋ฏธ์ณ์๋์ง์ ๋ํ ์ฌ์คํ ํต์ฐฐ์ ๊ณต์ ํ์ต๋๋ค. ๊ทธ์ ์ด์ผ๊ธฐ๋ ์์ ์ด ์ธ๊ฐ ์กด์ฌ์ ๋ณธ์ง์ ์ธ ๋ถ๋ถ์์ ๊นจ๋ซ๊ฒ ํฉ๋๋ค.
1๋ถ: ์์ ์ ์ฌ์คํ ์ญ์ฌ์ ๊ธฐ์
์งํ ๋๊ธฐ: ์ ๋ฐ ํ๋ผ๋ฆฌ์ ใ์ฌํผ์์คใ๋ฅผ ๋์ด์
์คํผ์ฒ ๊ต์๋ ์ ๋ฐ ํ๋ผ๋ฆฌ(Yuval Harari)์ ใ์ฌํผ์์ค(Sapiens)ใ๋ฅผ ์ฝ๊ณ ๊น์ ์๊ฐ์ ๋ฐ์์ง๋ง, ํ ๊ฐ์ง ์๋ฌธ์ ํ์์ต๋๋ค. โ์์ ์ ์ฌํผ์์ค๋ณด๋ค ์ต์ ๋ฐฑ๋ง ๋ ์ ๋ ์ค๋๋์๋ค.โ ๋๋ถ๋ถ์ ์์ ๊ด๋ จ ์์ ๋ค์ด ํน์ ์๊ณก๊ฐ๊ฐ ํน์ ์๊ธฐ์ ์ด๋ค ๊ณก์ ์ผ๋์ง์ ์ด์ ์ ๋ง์ถ๋ ๋ฐ๋ฉด, ์คํผ์ฒ ๊ต์๋ ๋ ํฐ ๊ทธ๋ฆผ์ ๋ณด๊ณ ์ถ์์ต๋๋ค. ์์ ์ ๋ณดํธ์ ์ด๋ฉฐ, ์ฌ์ง์ด ์ฌํผ์์ค ์ข ์ ๋์ด ๋๋ฌผ(์, ๊ณ ๋ ๋ฑ)์๊ฒ๋ ์กด์ฌํฉ๋๋ค. ๊ฒฐ๊ตญ, โ์์ ์ ์ธ๊ฐโ์ โ์์ ์ ๋๋ฌผโ์ธ ์ ์ ๋๋ค.
์์ ์ ์ฌ์๋์ ํ๊ตฌ: ์ ๋ฌผ๊ณผ ์ถ๋ก
์์ ์ ์ ์ฌ์๋๋ฅผ ๊ธฐ๋กํ๋ ๊ฒ์ ๊ฑฐ์ ๋ถ๊ฐ๋ฅ์ ๊ฐ๊น์ต๋๋ค. 1777๋ ์๋์จ์ด ์ถ์๊ธฐ(phonograph)๋ฅผ ๋ฐ๋ช ํ๊ธฐ ์ ๊น์ง๋ ์๋ฆฌ์ ๊ธฐ๋ก์ด ์๊ธฐ ๋๋ฌธ์ ๋๋ค. ์ ๊ธฐ ์ฌ๋ฃ๊ฐ ์๋ถํด๋๋ ๋ฌธ์ ๋ ์์ต๋๋ค. ํ์ง๋ง ์ฉ์ง ์๋ ๋๋ก ๋ง๋ ์๊ธฐ ์ ๊ธฐ(lithic instruments)๋ ์์ธ์ ๋๋ค. ํ์๋์์ ์ ๋ช ํ โ๋ฐ์ ๊ณต(rock gongs)โ์ด๋ ์ข ์ ์(stalactites) ๊ฐ์ ๊ฒ๋ค์ด์ฃ .
๊ฐ์ฅ ์ค์ํ ๋ฐ๊ฒฌ์ ์ฝ 4๋ง ๋ ์ ๋ ์ผ ๋จ๋ถ ๋๊ตด์์ ๋ฐ๊ฒฌ๋ ๋ผ ํผ๋ฆฌ(bone flutes)์ ๋๋ค. ๊ทธ๋ฆฌํฐ ๋ ์๋ฆฌ์ ๋ผ๋ก ๋ง๋ค์ด์ก์ฃ . ๊ทธ ์ธ์๋ ๋์๊ธฐ(ceramics)์ ๋ฐ๋ช ์ดํ ๊ฐ์ฃฝ์ ์์ ๋ง๋ ์ต์ด์ ํ๋ ์ ๋๋ผ(frame drums)์ด ์์ต๋๋ค. ๊ทธ๋ฌ๋ ๊ฐ์ฃฝ๊ณผ ํ(strings)์ ์๋ถํด๋๋ฏ๋ก, ์ ํํ ์ธ์ ๋ง๋ค์ด์ก๋์ง์ ๋ํ ์ฆ๊ฑฐ๋ ์์ต๋๋ค. ํ์ ๊ธฐ(harps and lutes)๋ ๋์ ์ด ๋ฐ๋ฌํ๊ณ ์๋ฅผ ๊ธธ๋ฌ ๋๋ฌผ์ ๋ด์ฅ์ผ๋ก ํ์ ๋ง๋ค๋ฉด์ ๋ฑ์ฅํ์ต๋๋ค.
๋ฐ๋ผ์ ์ฐ๋ฆฌ๋ ์๋ ค์ง ์ ๋ณด๋ก๋ถํฐ ์ถ๋ก ํด์ผ ํฉ๋๋ค. ํด๋ถํ์ ์ ์ฌ์๋, ๋๊ตฌ ๊ธฐ์ ์ ์งํ, ์ธ์ดํ์ ํตํ ์ธ์ด ๋ชจ๋ธ๋ง, ๊ทธ๋ฆฌ๊ณ ์ ์ธ๊ณ ์๋ ต์ฑ์ง ์ฌํ์ ์ผ์์ํ ๊ด์ฐฐ ๋ฑ ๋ค์ํ ํ๋ฌธ ๋ถ์ผ๋ฅผ ์ข ํฉํด์ผ ํฉ๋๋ค.
์ถ๋ก ์ ์ ๊ทผ: ์๋ฉด ๋๋ผ์ ๋์นญ์ฑ
ํ ๊ฐ์ง ์ถ๋ก ์ ์ฃผ์ฅ์ ์์๋ ์ฝ 150๋ง ๋ ์ ํธ๋ชจ ์๋ ํฌ์ค(Homo Erectus)์ ์กฐ์์ธ ํธ๋ชจ ์๋ฅด๊ฐ์คํฐ(Homo Ergaster)๊ฐ ๋ฐ๋ช ํ โ์๋ฉด ๋๋ผ(bifacial axe)โ์ ๋๋ค. ์ด ๋๋ผ๋ ๋ ๊ฐ์ ๋์นญ์ ์ธ ๋ฉด์ ๊ฐ์ง๊ณ ์์ต๋๋ค. ๋๋ผ์ ๋์นญ์ฑ์ ๋ถ์ฌํ๋ ๋ฅ๋ ฅ์ ๋ ๊ฐ์ง๋ฅผ ์์ฌํฉ๋๋ค. ์ฒซ์งธ, ์๋ฆ๋ค์ด ํํ๋ฅผ ๋ง๋ค๊ณ ๋ฏธํ์ ์ธ ์ด์ ๋ก ํํ๋ฅผ ์ฆ๊ธฐ๋ ๋ฅ๋ ฅ์ ๋๋ค. ๋์งธ, ์ฌ๋ฆฌํ์์ ๋น๋ ค์ ๋๊ตฌ์ ์์ญ์์ ์๋ฆฌ์ ์์ญ์ผ๋ก ๋งคํํ ์ ์์ต๋๋ค. ์ ์ ๋ฅ๋ ฅ์ โ๊ต์ฐจ ์์์ (crossmodal)โ์ด๊ธฐ ๋๋ฌธ์ ๋๋ค. ์๋ฆฌ์์๋ ๋์นญ์ฑ์ ์ฐพ์๋ณผ ์ ์์ผ๋ฉฐ, ์ฐ๋ฆฌ๋ ์ด๋ฅผ ๋ฐ์(meter) ๋๋ ๊ท์น์ ์ธ ๋ฆฌ๋ฌ(regular rhythm)์ด๋ผ๊ณ ๋ถ๋ฆ ๋๋ค. ๋ฐ๋ผ์ ์๋ฉด ์๋๋ผ์ ์ฆ๊ฑฐ๋ฅผ ๋ฐํ์ผ๋ก ํธ๋ชจ ์๋ฅด๊ฐ์คํฐ๊ฐ ์ด ๋๋ผ๋ฅผ ๋์นญ์ ์ธ ๋ฆฌ๋ฌ์ด๋ ์์ ์ ๋ฐ์์ ๋ง์ถฐ ๋๋๋ฆด ์ ์์์ ๊ฒ์ด๋ผ๊ณ ์ถ๋ก ํ๋ ๊ฒ์ ํ๋นํฉ๋๋ค.
๋ ๋ฐ ๊ฑท๊ธฐ์ ์์ ์ ๋ฆฌ๋ฌ์ ๊ฐ์ธ
์์ ์ ๊ธฐ์์ ์ด์ผ๊ธฐํ ๋, ์ด๋ ํจ์ฌ ๋์ค์ ์ข ํฉ๋ ์์ ์ ์ฌ๋ฌ ์์๋ฅผ ์กฐ๋ฆฝํ๋ ๊ณผ์ ๊ณผ ๊ฐ์ต๋๋ค. ๊ทธ ์์ ์ค ํ๋๊ฐ ๋ฐ๋ก ์ง๋ฆฝ๋ณดํ(bipedalism)์ ๋๋ค. 440๋ง ๋ ์ , โ์๋(Adi)โ๋ผ๊ณ ๋ถ๋ฆฌ๋ ์๋ฅด๋ํผํ ์ฟ ์ค(Ardipithecus)๊ฐ ๋ ๋ฐ๋ก ์์ ๊ฑธ์์ต๋๋ค. ๊ทธ ์ดํ๋ก ๊ฑท๋ ๋ฆฌ๋ฌ์ ์ธ๊ฐ ์์ ์ ๊น์ด ๊ฐ์ธ๋์์ต๋๋ค.
๋ ๋์๊ฐ, ์ฒซ๊ฑธ์์ ๋์ ๊ทผ์ก ํ๋, ๊ทธ๋ฆฌ๊ณ ์๋ฆฌ ์ฌ์ด์ ์ฐ๊ฒฐ ๊ณ ๋ฆฌ๋ฅผ ํ์ฑํ๋ ๊ธธ์ ์ด์์ต๋๋ค. ์ด๊ธฐ ์ธ๋ฅ๋ ๋ฐ์๋ฆฌ๋ฅผ ํ๋์ ํจํด์ผ๋ก ๋ฃ๋ ๋ฒ์ ๋ฐฐ์ ๊ณ , ํจํด์ ์ฐ๋ฆฌ์๊ฒ ์๊ฐ ๊ฐ๊ฐ์ ์ค๋๋ค. ๋ค์์ ๋ฌด์์ด ์ผ์ด๋ ์ง ์์ธกํ ์ ์๊ฒ ํ๋ ๊ฒ์ด์ฃ . ๋ํ โ๊ฑท๊ธฐโ๋ผ๋ ๊ฐ๋ ์ ์ธ๊ฐ์๊ฒ ์์ ์ด โ์์ง์ธ๋คโ๋ ์์ ์ ๋ํ ๋งคํน์ ์ ์ฌํ์ต๋๋ค. ์ค์ ์์ ์ ๊ณต์ค์ ๋ ๋ค๋๋ ์๋ฆฌ๋ ์์์ผ ๋ฟ์ด์ง๋ง, ์ฐ๋ฆฌ๋ ํ ์ํ๊ฐ ๋ค๋ฅธ ์ํ๋ก ์์ง์ธ๋ค๊ณ ์์ํฉ๋๋ค. ๋๋ถ๋ถ์ ์์ ์ ๋ง์์์์ ํ ์ง์ ์์ ๋ค๋ฅธ ์ง์ ์ผ๋ก ์ด๋ํ๋ ์ฌ์ ์ ํผ์ณ๋ ๋๋ค. ์ด๋ ์ํ๋ฆฌ์นด๋ฅผ ๋ฒ์ด๋ ์ฌํํ๋ ์ฐ๋ฆฌ ์กฐ์๋ค์ ์ฌ์ ์์ ๋น๋กฏ๋ ์์ฃผ ๋จผ ๋ฉ์๋ฆฌ์ ๋๋ค.
์ง๋ฆฝ๋ณดํ์ ์งํ์ ์ ์์ ๊ฑฐ๋ํ ์ฐ์ ๋ฐ์์ ์ฒซ๊ฑธ์์ด์์ต๋๋ค. ์ฐ๋ฆฌ์ ๋๊ฐ๊ณจ(cranial) ์ฉ๋์ ์ธ ๋ฐฐ๋ก ์ปค์ก๊ณ , ์ฐ๋ฆฌ๋ ํจ์ฌ ๋๋ํด์ก์ต๋๋ค. ๋ ํฌ๊ธฐ ์ฆ๊ฐ๋ ์๊ฐ๋ฝ์ ์ ์ดํ๋ ๋ฅ๋ ฅ์ ๊ฐ์ ธ์๊ณ , ์ด๋ ์์ญ๊ณผ์ ์ฐ๊ฒฐ์ ๊ฐํํ์ฌ ๋์ฑ ์ ๊ตํ๊ฒ ํผ๋ฆฌ๋ฅผ ๋ง๋ค๊ณ ์ฐ์ฃผํ ์ ์๊ฒ ํ์ต๋๋ค.
๋ํ ์ผ์ด์๋ ๊ฒ์ ์ฐ๋ฆฌ๊ฐ ๋ ๋ง์ ๊ณต๊ฐ์์ ์จ์ ์ด ์ ์๊ฒ ํด์ฃผ์๊ณ , ํ๋(larynx)๊ฐ ์ฑ๋(vocal tracts)๋ฅผ ๋ฐ๋ผ ๋ด๋ ค์์ต๋๋ค. ํ๋ฅผ ์ง์งํ๋ ์ค๊ณจ(hyoid bone)์ด ์งํํ์ฌ ์ฐ๋ฆฌ๊ฐ ๋ถ๋ฅด๋ ๊ฒ์ ๋ ๋ช ํํ๊ฒ ๋ฐ์ํ ์ ์๊ฒ ๋์์ต๋๋ค. ์ฐ๋ฆฌ์ ์ฑ๋๋ ๋ฌดํํ ๋ค์ํ ์๋ฆฌ๋ฅผ ๋ง๋ค์ด๋ด๋ ๋ฐฉ๋ฒ์ ๋ฐฐ์ ์ต๋๋ค. ์๋ฆฌ๋ฅผ ๋ง๋๋ ์ฐ๋ฆฌ์ ๋ฅ๋ ฅ์ ๊ทธ ๊ธฐ๋ฅ์ ๋ชฉ์ ์ ์ด๊ณผํ์ต๋๋ค. ์๋ฅผ ๋ค์ด ๋ฒ๋น์์ญ์ด(vervid monkey)๋ ๋ค ๊ฐ์ง ์ข ๋ฅ์ ์๋ฆฌ๋ฅผ ๋ด๊ณ , ๊ฐ ์๋ฆฌ๋ ํน์ ์ข ๋ฅ์ ํฌ์์๋ฅผ ๊ฒฝ๊ณ ํฉ๋๋ค. ํ์ง๋ง ์์ฒ ๊ฐ์ง ์๋ฆฌ๋ฅผ ๋ผ ์ ์๊ฒ ๋๋ฉด์ ์๋ฆฌ์ โ๊ณผ์(excess)โ์ด ๋ฐ์ํ์ต๋๋ค. ๋ฐ๋ก ์ด ์ง์ ์์ ์์ ์ด ๊ฐ๋ฅ์ฑ์ผ๋ก ๋ ์ค๋ฆ ๋๋ค. ์๋ฆฌ ๊ทธ ์์ฒด๋ฅผ ๊ฐ์ง๊ณ ๋๊ณ ์ฆ๊ธฐ๊ธฐ ์์ํ ๊ฒ์ด์ฃ . ์ด๋๋ถํฐ ์ธ๊ฐ ์์ ์ ๋๋ฌผ์ ๋ฐ์ฑ์ด๋ ์ธ์์๋ฆฌ์๋ ๋ค๋ฅธ ๊ธธ์ ๊ฑท๊ธฐ ์์ํ๋ค๊ณ ์คํผ์ฒ ๊ต์๋ ๋งํฉ๋๋ค.
2๋ถ: ๊ธฐ์ต๊ณผ ์ญ์ฌ๋ฅผ ๋ด๋ ์์
์์ ์ ๊ณง ๊ธฐ์ต์ด์ ์ญ์ฌ
์คํผ์ฒ ๊ต์๊ฐ ์ฑ ์ ์ฐ๋ฉด์ ๋ฐ๊ฒฌํ ์ฌ์คํ ๋ฉ์์ง ์ค ํ๋๋ ์์ ์ด ๊ธฐ์ต๊ณผ ์ญ์ฌ์ ๊ดํ ๋ชจ๋ ๊ฒ์ด๋ผ๋ ์ ์ ๋๋ค. 2๋ฐฑ๋ง ๋ ์ ๋๊ตฐ๊ฐ๊ฐ ์๋ค์๊ฒ ๋์ ๊น๋ ๋ฒ์ ๊ฐ๋ฅด์น๋ฉด, ์๋ค์ ๊ทธ๊ฒ์ ๊ธฐ์ตํฉ๋๋ค. ์ ํต์ โ๊ตณ์ด์ง ๊ทผ์ก ๊ธฐ์ต(congealed muscle memory)โ์ด๋ผ๊ณ ๋ถ๋ฅผ ์ ์์ต๋๋ค. ๋ฌด์ธ๊ฐ๋ฅผ ํ๋ ๋ฒ์ ๋ฐฐ์ฐ๋ ๊ฒ์ ์ด๊ฐ์ (haptic)์ด๋ฉฐ, ๋ํจ์ค(Dawkins)๊ฐ ๋งํ๋ ๋ฐ(meme), ์ฆ ๋ชจ๋ฐฉ์ (mimetic)์ธ ๊ฒ์ด์ง, ๋ฐ๋์ ์ธ์ง์ ์ธ ๊ฒ์ ์๋๋๋ค. ์ฐ๋ฆฌ์ ๊ทผ์ก์ด ํ๋ ๋ฒ์ ๋ฐฐ์ฐ๋ ๊ฒ์ด์ฃ .
๋ผ ํผ๋ฆฌ์๋ ์๊ฐ๋ฝ์ ์ด๋์ ๋์์ผ ํ ์ง ์๋ ค์ฃผ๋ ์์ ํํ์ ๋ค์ด ์๊ฒจ์ ธ ์์ต๋๋ค. ํผ๋ฆฌ์ ๋ค์ฏ ๊ตฌ๋ฉ์ ์ฐ์ฃผ ๋ฐฉ๋ฒ์ ์๋ ค์ฃผ๋ฉฐ, ์ด๊ฒ ๋ํ ๊ธฐ์ต์ ๋๋ค. ์ง๊ตฌ์์ ๋๋ถ๋ถ์ ๋ฌธํ๊ถ์ ์์ ์ ์์ ๋ค์ ๊ธฐ์์ ์ค๋ช ํ๋ ๋ฐฉ์์ผ๋ก ์ฌ๊น๋๋ค. ์๋ฅผ ๋ค์ด, ํํธ์๋ด๊ธฐ๋์ ์ฟจ๋ฆฌ(Kuli) ๋ถ์กฑ์ ์์ ์ด โ๋ฌ์(mooney bird)โ๋ผ๋ ๊ณผ์ผ ๋น๋๊ธฐ์์ ์จ๋ค๊ณ ๋ฏฟ์ต๋๋ค. ๊ทธ๋ค์ ์ฒ ์๋ฅผ ๋งด๋๋ ์๋ค์ด ๋ฌธ์ ๊ทธ๋๋ก ์กฐ์๋ค์ด ์์ ๋ค์๊ฒ ๋ง์ ๊ฑฐ๋ ๊ฒ์ด๋ผ๊ณ ์๊ฐํฉ๋๋ค. ๋น๋๊ธฐ์ ์ธ์์๋ฆฌ๋ฅผ ํ๋ด ๋ด๋ฉฐ, ๊ทธ๋ค์ ์ถ์ ๊ฑฐ๋ํ ์ํ์ ์ฐธ์ฌํ๊ณ , ์กฐ์๋ค์ ๋ ธ๋๋ฅผ ๋ค์ ๋ถ๋ฅด๋ ๊ฒ์ ๋๋ค. ์์์์๋ ๋ฒ ํ ๋ฒค์ โ์์ ๊ตํฅ๊ณก(Eroica)โ์ด 19์ธ๊ธฐ ํ๋์ค์์ ๋ํด๋ ์น ์นํ์ ์ด๋ฉด์ ์ ํฌ๋ฅผ ์น๋ ๋ ๊ธฐ์ต์ ๋์๊ธฐ๋ ๊ฒ๊ณผ ๊ฐ์ ๋งฅ๋ฝ์ ๋๋ค. ์์์ด๋ ์๋๋ , ์๋ฐฑ๋ง ๋ ์ ์ด๋ ์ง๊ธ์ด๋ , ๋ชจ๋ ์์ ์ ๊ธฐ์ต๊ณผ ์ญ์ฌ์ ๋ฐ์ ํ๊ฒ ์ฐ๊ฒฐ๋์ด ์์ต๋๋ค.
์ด๊ธฐ ์์ ๊ณต์ฐ์ ํ๊ฒฝ
์ต์ด์ ๋ผ ํผ๋ฆฌ๊ฐ ๋ฐ๊ฒฌ๋์์ ๋, ๋ถ๋ฌ์ง ํผ๋ฆฌ ์กฐ๊ฐ ์ค ์ผ๋ถ๋ ๋๊ตด ๋ด ์ต๋ ๊ณต๋ช ์ง์ ์์ ๋ฐ๊ฒฌ๋์์ต๋๋ค. ์ด๋ ์์ ๊ณผ ์๋ฆฌ, ์ํฅํ, ๊ฑด์ถ ์ฌ์ด์ ์์ฐ์ค๋ฌ์ด ๊ฒฐํฉ์ ์์ฌํฉ๋๋ค. ์์์์๋ ๊ตํ๊ฐ ๋ณธ์ง์ ์ผ๋ก ์ ์ ์ญ๋ฐฐํ๋ ๋๊ตด์ด๋ฉฐ, ๊ตํ์์ ๋ถ๋ฅด๋ ๋จ์ ์จ ์ฑ๊ฐ(plain chant)์ ๊ทธ ์ํฅํ์ ๊ณต๋ช ์ฌ์ด์ ์์ฐ์ค๋ฌ์ด ์นํ์ฑ์ ์ฐ์ฐ์ด ์๋๋๋ค.
๋๊ตด ์ ์์ ์ ์์(rituals)์ ์ํด ๋ณด์กด๋์์ ๊ฐ๋ฅ์ฑ์ด ํฝ๋๋ค. ์๋ฌด๋ ๋๊ตด์์ ์ด๊ณ ์ถ์ด ํ์ง ์์์ ํ ๋๊น์. ์ด๋ก๊ณ ์ํํ๋ฉฐ ์ ๋น๋ก ๊ฐ๋ํ ๊ณณ, ์ ์ฑํ ์ธ๊ณ๋ก ํตํ๋ ๋ฌธ, ์ผ์ ๋งน์๋ค์ ์์ ์ฒ์์ ๊ฒ์ ๋๋ค. ๋๊ตด์ด ์ ๊ณตํ๋ ๋๋ผ์ด ๊ณต๋ช ์ ํผ๋ฆฌ๋ ์ธ๊ฐ์ ๋ชฉ์๋ฆฌ, ๋๋ ์ข ์ ์์ ๋๋๋ฆฌ๋ ์๋ฆฌ์ ํ์ฉ๋์ด ํน๋ณํ ์์์ด๋ ์๋ก๋ฅผ ์ํด ์ฌ์ฉ๋์์ ๊ฒ์ ๋๋ค.
๊ทธ ์ธ์ ์ฌ๋๋ค์ โํ๋(hearth)โ ์ฃผ์์์ ์์ ์ ๋ง๋ค์์ต๋๋ค. ํ๋์ ๋ถ์ด ์๋ ์์ ์ฑ์ ์ง์ ์ ๋๋ค. ๋ถ์ ๊ณ ์ ๋ ํน์ฑ์ ์์ ์ ์์์ ๊ณ ์ ๋ ํน์ฑ์ ์์งํฉ๋๋ค. ํ๋์ ์ํ์ ์์์ด ์ํ์ ์ผ๋ก ๋ฐ๋ณต๋๋ ๊ฒฝํฅ์ด ์์์์ ๋ํ๋ ๋๋ค. ์ฌ๋ฅ์ด๋ ์ฌ๋ฅ์ ํตํด ์ป์ ๊ณ ๊ธฐ๋ฅผ ๋จน๋ ๊ฒ๊ณผ ๊ฐ์ ์๋ก์ ์ฌ๊ฑด ํ์ ์์ฃผ ํํด์ก์ผ๋ฉฐ, ๋ ธ๋์ ์ ์ฒด ์์ง์ ๋๋ ์ถค ์ฌ์ด์ ๊ตฌ๋ถ์ด ์์์ ๊ฒ์ ๋๋ค.
์์์์๋ ์ด ๋ชจ๋ ๊ฒ์ ์์ฐ๋ฅด๋ โ์์ (music)โ์ด๋ผ๋ ๋จ์ด๊ฐ ์์ง๋ง, ๋๋ถ๋ถ์ ๋ฌธํ๊ถ์์๋ ๋ ธ๋, ์ ๊ธฐ ์๋ฆฌ, ์ถค, ์์ ์ ๋ง์ถฐ ์ด์ผ๊ธฐ๋ฅผ ํ๋ ๊ฒ ๋ฑ์ ๋ํด ๋ค๋ฅธ ๋จ์ด๋ค์ ๊ฐ์ง๊ณ ์์ต๋๋ค. ์ด ๋ชจ๋ ๊ฒ๋ค์ด ํ๋์ ์์์ผ๋ก ์ฝํ ์์๋ ๊ฒ์ด์ฃ . ๋ฌผ๋ก ๋ฐฑ๋ง ๋ ์ ์ผ๋ก ๋์๊ฐ ์๋ ์์ผ๋ฏ๋ก ๋ชจ๋ ์ฃผ์ฅ์ ์ถ๋ก ์ ๊ธฐ๋ฐํ์ง๋ง, ์ฌ๋ง์์ ์ฌ๋ฐ๋, ์ด๋์ฐ๋ฆผ, ๊ทน์ง๋ฐฉ์ ์ด๋ฅด๊ธฐ๊น์ง ๋ค์ํ ํ๊ฒฝ์ ํฉ์ด์ ธ ์๋ ์๋ ต์ฑ์ง ๋ฏผ์กฑ๋ค์ด ๊ณตํต๋ ํต์ฌ์ ๊ณต์ ํ๋ค๋ ์ ์ ๋งค์ฐ ์ธ์์ ์ ๋๋ค. ์ด๋ ๋จ์ํ ๋น์ ๋ฅผ ๋์ด์ ๋ณธ์ง์ ์ธ ๋ฌด์ธ๊ฐ๋ฅผ ์์ฌํฉ๋๋ค.
์๋ ต์ฑ์ง ์ฌํ์ ์์ : ์์ฐ๊ณผ์ ์ฝํ
์๋ฅผ ๋ค์ด, ํธ์ฃผ ์์ฃผ๋ฏผ๊ณผ ๋ถ๋ฏธ ์ด๋์ดํธ์กฑ์ฒ๋ผ ๋งค์ฐ ๋ค๋ฅธ ๋ฏผ์กฑ๋ค์ ๋น๊ตํด๋ณด๋ฉด, ๊ทธ๋ค์ด ๊ณต์ ํ๋ ๊ฒ์ ์ธ๊ฐ์ฑ๊ณผ ๋๋ฌผ๊ณผ์ โ์ฝํ(entanglement)โ์ ๋๋ค. ์ด๋ค ๋ฉด์์ ์์ ์ ๋ชฉ์ ์ ์ฐ๋ฆฌ๊ฐ ์ด๋์์ ์๋์ง์ ๋ํ ๊ธฐ์ ์ด์ผ๊ธฐ๋ฅผ ๋ค๋ ค์ฃผ๋ ๊ฒ์ ๋๋ค. ์ฐ๋ฆฌ๋ ๋๋ฌผ์์ ์์ผ๋ฉฐ, ๊ทธ๋ค์ ๋ ธ๋ ๋๋ถ๋ถ์ ๋๋ฌผ์ ์ธ์์๋ฆฌ๋ ์ถค์ ๋ชจ๋ฐฉํฉ๋๋ค. ๋ค์(Darwin) ์์ ๋ ํธ์ฃผ์ ์นด๋ผ๋น(karaby) ์ฌ๋ฆ์ ๊ด์ฐฐํ๋ฉด์ ์์ฃผ๋ฏผ๋ค์ด ํค์(kiwis)์ ์บฅ๊ฑฐ๋ฃจ(kangaroos)์ ์ถค์ ๋ชจ๋ฐฉํ๋ ๊ฒ์ ๋ณด์์ต๋๋ค. ์ด๋์ดํธ์กฑ์ด ๋ฐ๋คํ๋ฒ ์๋ผ์ ์ธ์์๋ฆฌ๋ฅผ ๋ชจ๋ฐฉํ์ฌ ์๋ผ๋ฅผ ์ ์ธํ๊ณ ์ฌ๋ฅํ๋ ๊ฒ์ฒ๋ผ, ์๋ฉ๋ฆฌ์นด ์์ฃผ๋ฏผ์ด ๋ฒํ๋ก ์๋ผ์ ์ธ์์๋ฆฌ๋ฅผ ๋ชจ๋ฐฉํ๋ ๊ฒ๊ณผ ์ ์ฌํฉ๋๋ค. ๊ฐ ํ๊ฒฝ์์ ํ๋ ์๋ ต์ฑ์ง์ธ๋ค์ ๋๋ฌผ์ ์ธ์์๋ฆฌ๋ฅผ ๋ชจ๋ฐฉํฉ๋๋ค.
๋ฟ๋ง ์๋๋ผ, ๊ทธ๋ค์ ์์ ๋ค์ด ํ๊ฒฝ๊ณผ๋ ์ฝํ ์๋ค๊ณ ์๊ฐํฉ๋๋ค. ๋ธ๋ฃจ์ค ์ฑํธ์(Bruce Chatwin)์ด ๊ด์ฐฐํ ํธ์ฃผ ์์ฃผ๋ฏผ์ ์ ๋ช ํ โ์ก ๋ผ์ธ(song lines)โ์ ๊ทธ๋ค์ด ์ฌ๋ง์ ์ฌํํ๊ฑฐ๋ ํญํดํ๋ฉด์ ํ๊ฒฝ์ ๊ฐ ํน์ง์ ๋ ธ๋๋ฅผ ๋ถ์ด๊ณ , ๋ ธ๋๋ฅผ ํตํด ํ๊ฒฝ์ ํ์ํ๋ฉฐ, ์์ ์ ์ค๋จธ๋์ฆ์ (shamanistic) ์์ ์ฌ์ ์ผ๋ก ๋ด ๋๋ค. ์ค๋จผ(shaman)์ ์ธ๋๋ฅผ ๋ฐ์ ์์ ์ ๋ ๊ฐ๋ฅผ ๋ฌ์์ฃผ์ด ์ฐ์ฃผ์ ์ฌ๋ฌ ์ธต์ ๋ ์๋ค๋๊ณ ๊ถ๊ทน์ ์ผ๋ก ์กฐ์๋ค์ โ๊ฟ์ ๋ (dreamland)โ์ ๋๋ฌํ๊ฒ ํฉ๋๋ค.
ํธ์ฃผ์ ๋ชจ๋ ํ๊ฒฝ๊ณผ ๋ถ๋ฏธ์ ๋ ํ๊ฒฝ์ ๋น๊ตํด๋ณด๋ฉด, 4๋ง ๋ ์ ๋นํ๊ธฐ์ ์ํฉ์ ์ ์ถํ ์ ์์ต๋๋ค. ๋์ ๋ชจ๋์ ๋ฌ๋ฆฌ ๋ฐ์๊ตญ์ด๋ ํ์ ์ ๋จ๊ธฐ์ง ์์ ๊ธธ์ ์ฐพ๊ธฐ๊ฐ ๋งค์ฐ ์ด๋ ต์ต๋๋ค. ์ด๋์ดํธ์กฑ์ด ๋ง์์์ผ๋ก ๊ธธ์ ํ์ํ๊ณ , ๋๋ฐญ์ ๋ฐ์๊ตญ์ ํจ์ฌ ๋ ๋จ๊ธด๋ค๋ ์ถ๋ก ์ ํ ์ ์์ต๋๋ค. ๋ฌธํ์ ๊ฑฐ๋ฏธ์ค์ด ๋ฐ์๊ตญ์ด๋ ํ๊ฒฝ ์ ๊ธธ์ ๊ฑฐ๋ฏธ์ค์ ๋์ฒดํ๋ ๊ฒ์ ์์ํ ์ ์์ต๋๋ค. ๋์ ๋ํ ๋ชจ๋๋ณด๋ค ์ธ๊ฐ ์์กด์ ํจ์ฌ ๋ ์ ๋์ ์ ๋๋ค. ์ด๋์ดํธ์กฑ์ ์ด๊ธ๋ฃจ(igloos)์ ๊ฐ์ ๋ฐํ๋ ๊ณต๊ฐ์์ ์๋ก ์์งํ๋ฉฐ ์ฌ๋ ๋ฒ์ ๋ฐฐ์์ผ ํฉ๋๋ค. ๊ทธ๋์ ์ด๋์ดํธ ์์ ์ ๊ธฐ๋ฅ์ ์ฃผ๋ก โ๋ถ๋ ธ๋ ๊ฐ๋ฑ ๊ด๋ฆฌโ์ ์์ต๋๋ค. ๊ทน์ง๋ฐฉ์ ์์ ๊ณต๊ฐ์์ ์๋ก์๊ฒ ์์กดํ๋ ์ํฉ์์๋ ์ธ์์ ๋ฒ์ผ ์ฌ์ ๊ฐ ์์ต๋๋ค. ๊ทธ๋์ ์ด๋์ดํธ ์์ ์ ๋ง์ ๋ถ๋ถ์ด ์์๊ณผ ๋๋ด์ผ๋ก ๊ฐ๋ ์ฐจ ์๋ ๊ฒ์ ๋๋ค. ์ด๋ ์ ์พํ ์ฌํ์ฑ์ ์ก์ฑํ๊ณ ๊ฐ๋ฑ์ ์ ์ฌ๋ ฅ์ ํด์ํ๋ ์ญํ ์ ํฉ๋๋ค.
3๋ถ: ์ฌํ ์งํ์ ์์ ์ ๋ณํ
์ธ๋ฅ ์ฌํ์ ์งํ์ ์์
์ฌํผ์์ค ์งํ์ ํฐ ๊ทธ๋ฆผ์ ๋ณด๋ฉด, ์๋๋ ์๋ ต์ฑ์ง์ธ, ์ ์ฐฉ ๋๊ฒฝ ๊ณต๋์ฒด, ๊ทธ๋ฆฌ๊ณ ๋์ ๋ฐ ๋์ ๊ตญ๊ฐ์ ๊ฑด์ค๋ก ๋๋ฉ๋๋ค. ๊ฐ ์๋๋ ํน์ ํ ์ฌ๊ณ ๋ฐฉ์๊ณผ ์ฐ๊ด๋์ด ์์ต๋๋ค.
-
์๋ ต์ฑ์ง์ธ(Hunter-gatherer): ์ ๋ชฉ ์ํ์ ํ๋ ๊ฒฝํฅ์ด ์์์ต๋๋ค. ์ด๋์ดํธ์กฑ, ํธ์ฃผ ์์ฃผ๋ฏผ, ์ํ๋ฆฌ์นด ํผ๊ทธ๋ฏธ์กฑ ๋ฑ ์ ์ธ๊ณ ์๋ ต์ฑ์ง ๋ฏผ์กฑ์ ์์์์ ๋ณผ ์ ์๋ฏ์ด, ๊ทธ๋ค์ ์ ๋ชฉ ์ํ ๋ฐฉ์์ ๊ณต์ ํฉ๋๋ค. ํ๊ฒฝ์ ๋์์์ด ์ฌํํ๊ธฐ ๋๋ฌธ์ ๋ฌด๊ฑฐ์ด ์ ๊ธฐ๋ฅผ ๋ค๊ณ ๋ค๋์ง ์์ต๋๋ค. ์์ ์ ํด๋์ฑ์ด ์ข์์ผ ํฉ๋๋ค. ์ด์์ ์ผ๋ก๋ ๋ชฉ์๋ฆฌ๋ง์ผ๋ก ์ถฉ๋ถํ๋ฉฐ, ๊ทธ๋ ์ง ์๋ค๋ฉด ๋งค์ฐ ๊ฐ๋ฒผ์ด ํผ๋ฆฌ๋ ์์ ํ์ ๊ธฐ์ฌ์ผ ํฉ๋๋ค. ๋ํ ๊ทธ๋ค์ ์์ ์ ๋งค์ฐ โ๋์ด์ (playful)โ์ด๋ฉฐ ๋ฐ๋ณต์ ํผํ๋ ๊ฒฝํฅ์ด ์์ต๋๋ค. ๋ฐ๋ณต ๊ฐ๋ฅํ โ์ํ(piece)โ์ด๋ โ์์ (work)โ์ด๋ผ๋ ๊ฐ๋ ์ ๋งค์ฐ ํ๋์ ์ธ ๋ฐ์์ ๋๋ค. ์นด๋ฉ๋ฃฌ ํผ๊ทธ๋ฏธ์กฑ์ ์์ ์ ๋ณด๋ฉด, ๊ทธ๋ค์ ์ฐ์ฃผํ ๋๋ง๋ค ๊ณก์ด ๋ค๋ฅด๊ฒ ๋ค๋ฆฝ๋๋ค. ๊ทธ๊ฒ์ ์ฒ ์ ํ โ์๊ฐ์ ์์ โ์ ๋๋ค.
-
๋๊ฒฝ ๊ณต๋์ฒด(Sedentary or farming community): ๋์ ์ ๋ฐ๋ช ํ๊ณ ์ ์ฐฉํ๋ฉด์ ๋ชจ๋ ์ฌ๊ณ ๋ฐฉ์์ด ๊ณ์ ์ ์ํ, ์๋ช ์ ์ํ์ ๊ณ ์ ๋ฉ๋๋ค. ์์ ์ ๋ฟ๋ฆฌ๋ฅผ ๋ด๋ฆฌ๊ณ ๋ ์ ๊น์ด ํ๊ณ ๋ญ๋๋ค. ์๋ช ์ ์ํ, ๊ณ์ ์ ์ํ์ ๋ชจ๋ ์๊ธฐ์ ๋ง์ถฐ ์์ ์ด ์๊ณก๋ฉ๋๋ค. ๋ฐ๋ณต ๊ฐ๋ฅํ โ์ํ(work)โ์ด ๋ฐ๋ช ๋๊ณ , ์ธ์์ ๊ตฌ์กฐ๋ ์๋ช ์์ฒด๋งํผ์ด๋ ์ํ์ ์ด ๋ฉ๋๋ค. ์์ ์์ โ์(circle)โ์ ๋ฐ๋ช ํ๊ณ , ์์ ์ ์์(musical rituals)์ ๋ฐ๋ช ํฉ๋๋ค.
-
๋์ ๊ตญ๊ฐ(Cities and city-states): ์์ ์ด ๋์ฅ์์ ๋์๋ก ์ด๋ํ๋ฉด์ ํน์ ๋ณํ๊ฐ ์ผ์ด๋ฉ๋๋ค. ๋์์์ ์๊ตฌ์ ์ธ ๋ฟ๋ฆฌ๋ฅผ ๋ด๋ฆฌ๊ธฐ ์์ํ๋ฉด์ ์ ๊ธฐ๋ ๋ฌด๊ฑฐ์์ง ์ ์์ต๋๋ค. ์ข (bells)์ด๋ ์ง(gongs)์ฒ๋ผ ๋ฌด๊ฑฐ์ด ์ ๊ธฐ๋ฟ๋ง ์๋๋ผ, ํํ(harps)๋ ๋ฅํธ(lutes)์ฒ๋ผ ์ฌ์ธํ์ฌ ์ฌํ ์ค ์์๋ ์ ์๋ ์ ๊ธฐ๋ค๋ ๋ง๋ค์ด์ง๋๋ค. ์์ ์ ๊ธฐ๋ฅ ๋ํ ์ฌํ ๊ณ์ธต์ ์ฑ์ฅ๊ณผ ํจ๊ป ๋ณํํฉ๋๋ค. ์์ ์ ์ญํ ์ โ์๋ (handmaiden)โ๊ฐ ๋์ด ์์๋ ๊ตํ์ ๊ถ๋ ฅ์ ์ฌ๊ธฐ๋ ๊ฒ์ ๋๋ค. ์์ ๊ฐ๋ค์ ์ ๋ฌธํ๋๊ณ , ๊ทธ๋ค์ ์๋ฌด๋ ์ฌ๊ฐ ์๊ฐ์ ๊ฐ์ง ์ฌ๋๋ค์ด ๋ค์ ์์ ์ ๋ง๋๋ ๊ฒ์ด ๋ฉ๋๋ค. ์ด๊ฒ์ด ์ฐ๋ฆฌ๊ฐ โ์ฝ์ํธ(concerts)โ๋ผ๊ณ ๋ถ๋ฅด๋ ๊ฒ์ ๊ธฐ์์ ๋๋ค.
์ต์ด์ ์ฝ์ํธ์ ๊ธฐ๋ฅ์ ์์
์ฐ๋ฆฌ๋ ์์ ์ด โ์ฝ์ํธโ๋ผ๊ณ ๋ถ๋ฆฌ๋ ๊ณณ์์ ์ฐ์ฃผ๋์๋ค๋ ํ๋์ ๊ฐ๋ ์์ ๋ฒ์ด๋์ผ ํฉ๋๋ค. ์ฝ์ํธ๋ฅผ ์ด๋ ค๋ฉด ์ฌ๊ฐ์ ๋์ด ํ์ํฉ๋๋ค. ๊ทธ๋ฆฌ๊ณ ์ข ์ข ๊ท์กฑ์ด๋ ์๋ฅ์ธต ์ค์ฐ์ธต๋ง์ด ํธ์ํ ์์ ์์ ๋ค์ ์ํด ์ฐ์ฃผ๋๋ ์์ ์ ์ฆ๊ธธ ์๊ฐ์ด ์์์ต๋๋ค. ์๋ฐฑ๋ง ๋ ๋์ ๋๋ถ๋ถ์ ์ฌ๋๋ค์๊ฒ๋ ๊ทธ๋ฐ ์ผ์ด ์์์ต๋๋ค.
๋ฐ๋ผ์ ๋๋ถ๋ถ์ ์์ ์ ๊ธฐ๋ฅ์ ์ผ๋ก ์ฐ์ฃผ๋์์ต๋๋ค. ์๋ฅผ ๋ค์ด 19์ธ๊ธฐ ๋ชฉํ๋ฐญ์์ ๋ถ๋ฆฌ๋
โThe Secrets of Claudeโs Platform From the Team Who Built Itโ โ Every ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
Anthropic, ํด๋ก๋ ํ๋ซํผ์ผ๋ก AI ์์ด์ ํธ ์๋์ ์๋ก์ด ์งํ์ ์ด๋ค: ๊ฐ๋ฐํ์ด ๋งํ๋ ์งํ์ ๋น๋ฐ
โํด๋ก๋(Claude)๊ฐ ๋์๊ฒ 10์ต ๋ฌ๋ฌ๋ฅผ ๋ฒ์ด์ค ๋ ์ ์ผ๋ง๋ ๋จ์์๊น์?โ
์ด ๋๋ฐ์ ์ธ ์ง๋ฌธ์ ์ต๊ทผ Every ์ฑ๋์ ํ์บ์คํธ์์ Anthropic์ ํด๋ก๋ ํ๋ซํผ ์ฑ ์์๋ค์ด ๋ฐ์ ์ง๋ฌธ ์ค ํ๋์์ต๋๋ค. Anthropic์ ํด๋ก๋ ํ๋ซํผ ์ ํ ์ด๊ด ์์ ค๋ผ(Angel)์ ์์ง๋์ด๋ง ์ด๊ด ์ผ์ดํ๋ฆฐ(Caitlyn)์ ์ด ์ง๋ฌธ์ ์ง์ ์ ์ธ ๋ต์ ์ฃผ์ง๋ ์์์ง๋ง, ๊ทธ๋ค์ด ๊ตฌ์ถํ๊ณ ์๋ AI ํ๋ซํผ์ ํ์ฌ์ ๋ฏธ๋์ ๋ํ ์ฌ๋ ๊น์ ํต์ฐฐ์ ์ ๊ณตํ๋ฉฐ AI ์์ด์ ํธ ์๋์ ๋ฌดํํ ๊ฐ๋ฅ์ฑ์ ์ฟ๋ณด๊ฒ ํ์ต๋๋ค.
์ด๋ค์ ํด๋ก๋๊ฐ ์ค์ค๋ก๋ฅผ ์ดํดํ๊ณ , ์ต์ ์ ๋ชจ๋ธ์ ์ ํํ๋ฉฐ, ํ์ ์์ด์ ํธ๋ฅผ ์์จ์ ์ผ๋ก ๊ตฌ๋ํ๋ ๋ฏธ๋๋ฅผ ์์ํฉ๋๋ค. ์ฌ์ฉ์๊ฐ ๋ณต์กํ ์ํคํ ์ฒ๋ฅผ ๊ณ ๋ฏผํ ํ์ ์์ด, ํด๋ก๋๊ฐ ์ฆ์์์ ํ์ํ ์์ด์ ํธ๋ฅผ ๊ตฌ์ฑํ๋ ์ธ์. ์ด๋ฌํ ์์จ์ ์ธ ์์ด์ ํธ๊ฐ ์ฌ์ฉ์์๊ฒ ์ต๊ณ ์ ๊ฒฐ๊ณผ๋ฌผ์ ์ ๊ณตํ๋ ค๋ฉด, ๊ทธ ๊ธฐ๋ฐ์ด ๋๋ ํ๋ซํผ์ ์์ ์ด์์ ํ์ฅ์ฑ์ ๊ฐ์ถฐ์ผ ํฉ๋๋ค.
AI ํ๋ซํผ์ ์งํ: ๋จ์ API๋ฅผ ๋์ด ์์ด์ ํธ ์๋๋ก
AI ํ๋ซํผ์ ์งํ๋ ๋๋ผ์ด ์๋๋ก ์ด๋ฃจ์ด์ง๊ณ ์์ต๋๋ค. ๊ณผ๊ฑฐ GPT-3 ์ด๊ธฐ์๋ ๋จ์ํ ํ๋กฌํํธ๋ฅผ ๋ณด๋ด ์๋ต์ ๋ฐ๋ โ์์ฑํ ์๋ํฌ์ธํธ(completion endpoint)โ๊ฐ ํ๋ซํผ์ ์ ๋ถ์์ต๋๋ค. ๋น์์๋ ํ์ ์ ์ด์์ง๋ง, ์ง๊ธ ๋์๋ณด๋ฉด ๋งค์ฐ ๊ธฐ์ด์ ์ธ ์์ค์ด์์ต๋๋ค.
์ดํ ํด ํธ์ถ(tool calling) ๊ธฐ๋ฅ๊ณผ ์ฑํ ์ธ์ ์ด ์ถ๊ฐ๋๋ฉด์ ๋ชจ๋ธ์ ์ธ๋ถ ๋๊ตฌ์ ์ํธ์์ฉํ๊ณ ๋ํ์ ๋งฅ๋ฝ์ ์ ์งํ ์ ์๊ฒ ๋์์ต๋๋ค. ๊ทธ๋ฆฌ๊ณ ์ด์ Anthropic์ ๋ฉ๋ชจ๋ฆฌ ๋ฑ ๋ค์ํ ๊ธฐ๋ฅ์ ํ์ฌํ โํด๋ก๋ ๊ด๋ฆฌํ ์์ด์ ํธ(Claude Managed Agents, CMA)โ๋ฅผ ์ ๋ณด์ด๋ฉฐ AI ํ๋ซํผ์ ์๋ก์ด ์๋๋ฅผ ์ด๊ณ ์์ต๋๋ค. ์์ ค๋ผ๋ ์ด๋ฌํ ๋ณํ๋ฅผ โ์ ์ ๋ ์ํ ์ ์ฅ(stateful)์ ์ธ ์ธ๊ณ๋ก ๋์๊ฐ๋ฉฐ ๋ชจ๋ธ์ ์ฑ๋ฅ์ ํฅ์์ํค๊ธฐ ์ํ ๊ณ ์ฐจ์ ์ถ์ํ(higher-order abstraction)์ ์ถ๊ตฌโ๋ผ๊ณ ์ค๋ช ํฉ๋๋ค.
์ด๊ธฐ์๋ ์ฌ์ฉ์๋ค์ด LLM(๋๊ท๋ชจ ์ธ์ด ๋ชจ๋ธ)์ผ๋ก ๋ฌด์์ ๋ง๋ค์ง ์์ธกํ๊ธฐ ์ด๋ ค์ ์ต๋ํ ๋ง์ ๊ฐ๋ฅ์ฑ์ ์ด์ด๋๋ ๊ฒ์ด ์ค์ํ์ต๋๋ค. ํ์ง๋ง ์ด์ ๋ ํน์ ์ฌ์ฉ ์ฌ๋ก๋ค์ด ๋ช ํํด์ง๋ฉด์, ๊ณ ๊ฐ๋ค์ โํด๋ก๋๋ฅผ ์ด๋ป๊ฒ ํ๋ฉด ๊ฐ์ฅ ์ ํ์ฉํ ์ ์์๊น?โ, โ๋๊ตฌ๋ฅผ ์ด๋ป๊ฒ ์ค์ ํ๊ณ ๋ฃจํ๋ฅผ ์ด๋ป๊ฒ ์คํํด์ผ ํ๋๊ฐ?โ์ ๊ฐ์ ๊ตฌ์ฒด์ ์ธ ์ง๋ฌธ์ ๋์ง๊ธฐ ์์ํ์ต๋๋ค. Anthropic์ ์ด๋ฌํ ์๊ตฌ์ ๋ถ์ํ๊ธฐ ์ํด ํ๋ซํผ์ ์ง์์ ์ผ๋ก ๊ณ ๋ํํ๊ณ ์์ผ๋ฉฐ, ์ด๋ ์ํ(state) ๊ด๋ฆฌ, ๋๊ตฌ ์ถ๊ฐ, ํด๋ผ์ฐ๋ ๊ตฌ์ฑ ์์ ํตํฉ ๋ฑ ๋ค์ํ ํํ๋ก ๋ํ๋๊ณ ์์ต๋๋ค. ๊ถ๊ทน์ ์ธ ๋ชฉํ๋ โ๋ชจ๋ ๊ฒ์ ๋ฌธ์ ๊ทธ๋๋ก ๊ฐ๋ฅํ ํ ์ฝ๊ฒ ๋ง๋๋ ๊ฒโ์ ๋๋ค.
ํด๋ก๋ ๊ด๋ฆฌํ ์์ด์ ํธ(CMA)์ ํต์ฌ ์์์ ๊ฐ์น
๊ทธ๋ ๋ค๋ฉด ํด๋ก๋ ๊ด๋ฆฌํ ์์ด์ ํธ์ ํต์ฌ ์์, ์ฆ โํ๋ฆฌ๋ฏธํฐ๋ธ(primitives)โ๋ ๋ฌด์์ผ๊น์? ์ผ์ดํ๋ฆฐ์ CMA๊ฐ Anthropic์ ๋ชจ๋ ๊ธฐ๋ณธ ํ๋ฆฌ๋ฏธํฐ๋ธ ์์ ๊ตฌ์ถ๋๋ค๊ณ ์ค๋ช ํฉ๋๋ค. ๋ฉ์์ง API(Messages API)๋ฅผ ๊ธฐ๋ฐ์ผ๋ก, ๋ด์ฅ๋ ๋๊ตฌ(built-in tools) ์ฌ์ฉ, ์ฝ๋ ์คํ(code execution)์ ์ํ ์๋๋ฐ์ค(sandbox) ์์ฑ ๋ฐ ์์ ์คํ, ์น ๊ฒ์(web search) ๋ฑ ๋ค์ํ ๊ธฐ๋ฅ์ด ํฌํจ๋ฉ๋๋ค. Anthropic์ ์ด ์ค ๊ฐ์ฅ ๊ฐ๋ ฅํ๋ค๊ณ ํ๋จ๋๋ ์์๋ค์ ๋ชจ์ ํด๋ก๋์์ ์ต๊ณ ์ ๊ฒฐ๊ณผ๋ฌผ์ ์ป์ ์ ์๋ โํ๋ค์ค(harness)โ์ ์ธํ๋ผ๋ก ํตํฉํ์ต๋๋ค.
์ด ๋๋ชฉ์์ ๋ง์ ๊ฐ๋ฐ์๊ฐ ์ง๋ฉดํ๋ ๋๋ ๋ง๊ฐ ๋๋ฌ๋ฉ๋๋ค. ์ธํฐ๋ทฐ ์งํ์ ์ญ์ ๋งฅ ๋ฏธ๋(Mac Mini) ๋ช ๋์์ ํด๋ก๋๋ฅผ ์ง์ ๊ตฌ๋ํ๋ฉฐ ์์ฒ ์ค์ ํ์ด์ฌ ์ฝ๋๋ก ์์ด์ ํธ๋ฅผ ๊ตฌ์ถํ๊ณ ์๋ค๊ณ ๋ฐํ์ต๋๋ค. ์ด๋ Anthropic์ด CMA๋ฅผ ํตํด ์ ๊ณตํ๋ ๊ฒ๊ณผ ์ ์ฌํ ๊ธฐ๋ฅ์ ์ง์ ๊ตฌํํ๋ ๋ฐฉ์์ ๋๋ค. ๊ทธ๋ โ์ฐ๋ฆฌ๊ฐ ์ง์ ๊ตฌ์ถํด์ผ ํ ๊น, ์๋๋ฉด Anthropic์ด ๋ง๋ค์ด์ค ๋๊น์ง ๊ธฐ๋ค๋ ค์ผ ํ ๊น?โ๋ผ๋ ์ง๋ฌธ์ ๋์ง๋ฉฐ, โ์ธํ๋ผ๋ ์ ๋ง ์ง๋ ํ๊ฒ ํ๋ค๋คโ๊ณ ํ ๋กํ์ต๋๋ค.
์ผ์ดํ๋ฆฐ์ ๋ฐ๋ก ์ด ์ง์ ์ด CMA ๊ฐ๋ฐ์ ํต์ฌ ๋๊ธฐ์๋ค๊ณ ๋งํฉ๋๋ค. Anthropic ๋ด๋ถ์์๋ ์์ด์ ํธ ์ ํ์ ๊ตฌ์ถํ๋ฉฐ ์ธํ๋ผ๋ฅผ ์์์ด ๋ฐ๋ณตํด์ ์ค์ ํ๊ณ ๊ด๋ฆฌํด์ผ ํ์ต๋๋ค. ์ด๋ค์ ๊ทธ ๊ณผ์ ์์ ์ป์ ํ์ต์ ๋ฐํ์ผ๋ก โํ ๋ฒ์ ์ ๋๋ก ๋โ ๋ฐฉ์์ผ๋ก ์ธํ๋ผ๋ฅผ ๊ตฌ์ถํ๊ธฐ๋ก ๊ฒฐ์ ํ๊ณ , ์ด๊ฒ์ด ๋ฐ๋ก CMA๋ก ์ด์ด์ง ๊ฒ์ ๋๋ค. ๊ฐ์ธ์ด๋ ์๊ท๋ชจ ํ๋ก์ ํธ์์๋ ๋งฅ ๋ฏธ๋๋ก ์ถฉ๋ถํ ์ ์์ง๋ง, ์ ํ์ ์์ด์ ํธ๋ฅผ ํตํฉํ๊ณ ๋๊ท๋ชจ๋ก ์ด์ํ ๋๋ ์ธํ๋ผ ๊ตฌ์ถ์ ์ด๋ ค์์ด ๊ธฐํ๊ธ์์ ์ผ๋ก ์ปค์ง๋ค๋ ๊ฒ์ด ์ด๋ค์ ์ค๋ช ์ ๋๋ค.
๋ชจ๋ธ ์ข ์์ฑ(Model Lock-in)๊ณผ ์ ์ฐ์ฑ ์ฌ์ด์ ๊ท ํ
๊ฐ๋ฐ์๋ค์ด ํ๋ซํผ์ ์ฌ์ฉํ ๋ ํํ ์ฐ๋ คํ๋ ์ ์ โ๋ชจ๋ธ ์ข ์์ฑ(model lock-in)โ์ ๋๋ค. ํน์ ๋ชจ๋ธ์ด๋ ํ๋ซํผ์ ๋ฌถ์ฌ ๋ค๋ฅธ ๋ชจ๋ธ๋ก ์ ํํ๊ธฐ ์ด๋ ต๊ฑฐ๋, ์ต์ ๊ธฐ๋ฅ์ ๋ฐ๋ก ์ฌ์ฉํ์ง ๋ชปํ ๊น ๋ด ๊ฑฑ์ ํ๋ ๊ฒ์ ๋๋ค. ๊ณผ๊ฑฐ์๋ ๋ชจ๋ธ ๊ฐ์ ์ฐจ์ด๊ฐ ํฌ์ง ์์ ๋ฒ์ฉ์ ์ธ ํ๋ค์ค(harness)๋ฅผ ๊ตฌ์ถํ๊ณ ๋ชจ๋ธ์ โํซ ์ค์ํ(hot swapping)โํ๋ ๊ฒ์ด ๊ฐ๋ฅํ์ต๋๋ค.
ํ์ง๋ง Anthropic ํ์ ์ด์ ์ํฉ์ด ๋ฌ๋ผ์ก๋ค๊ณ ์ง๋จํฉ๋๋ค. ๊ฐ ์ฐ๊ตฌ์(lab)๋ง๋ค ๋ชจ๋ธ์ ๋ฐ์ ์ํค๋ ๊ธฐ์ ๊ณผ ๊ด์ ์ด ๋ฏธ๋ฌํ๊ฒ ๋ฌ๋ผ์ง๋ฉด์, ๋ชจ๋ ๋ชจ๋ธ์ ์์ฐ๋ฅด๋ ๋ฒ์ฉ์ ์ธ ํ๋ค์ค๋ฅผ ๋ง๋๋ ๊ฒ์ด ์ ์ ๋นํจ์จ์ ์ด ๋๊ณ ์๋ค๋ ๊ฒ์ ๋๋ค. ์คํ๋ ค ๋ชจ๋ธ๊ณผ ํ๋ค์ค๋ฅผ ๊ธด๋ฐํ๊ฒ ๊ฒฐํฉํ์ฌ ํน์ ๋ชจ๋ธ์์ ์ต์ ์ ์ฑ๋ฅ์ ๋์ด๋ด๋ โํ๋ค์ค ์์ง๋์ด๋ง(harness engineering)โ์ด ์ค์ํด์ง๊ณ ์์ต๋๋ค. ์์ ค๋ผ๋ โ๊ฐ ํ๋ค์ค๊ฐ ๊ทน์ ์ผ๋ก ๋ค๋ฅธ ์ฑ๋ฅ์ ๋ณด์๋คโ๋ ๋ด๋ถ ์คํ ๊ฒฐ๊ณผ๋ฅผ ์ธ๊ธํ๋ฉฐ, ์ฌ๋ฐ๋ฅธ ํ๋ค์ค ์์ง๋์ด๋ง์ ํตํด โ์์ฒญ๋ ์์ ์ํ(alpha, ์ถ๊ฐ์ ์ธ ์ฑ๋ฅ ์ฐ์)โ๋ฅผ ํ๋ณดํ ์ ์๋ค๊ณ ๊ฐ์กฐํฉ๋๋ค.
Anthropic์ ์ด๋ฌํ ์ ์ ๊ณ ๋ คํ์ฌ CMA๋ฅผ ์ค๊ณํ ๋ ๋ ๊ฐ์ง ์ฒ ํ์ ๋ฐ๋ฆ ๋๋ค. ์ฒซ์งธ, ๋ชจ๋์ฑ(modularity)์ ๋๋ค. ํด๋ก๋ ๋ชจ๋ธ๊ณผ ๊ธด๋ฐํ๊ฒ ์ฐ๋๋์ด์ผ ํ๋ ํน์ ๋ถ๋ถ(์: ํ์ผ ์์คํ , ์คํฌ)์ ๋ํด์๋ ๋ช ํํ ์๊ฒฌ์ ์ ์ํ์ง๋ง, ๋ค๋ฅธ ๋ถ๋ถ์์๋ ์ ์ฐ์ฑ์ ์ ๊ณตํ์ฌ ์ฌ์ฉ์๊ฐ ๋ค์ํ ์์๋ฅผ ์ถ๊ฐํ ์ ์๋๋ก ํฉ๋๋ค. ๋์งธ, ํฌ๋ช ์ฑ์ ๋๋ค. ๋ธ๋ก๊ทธ ๊ฒ์๋ฌผ์ด๋ ์ฐธ์กฐ ๊ตฌํ(reference implementations)์ ํตํด ๋์์ธ ์์น์ ๊ณต๊ฐํ์ฌ, ๊ฐ๋ฐ์๋ค์ด ๋ฉ์์ง API๋ฅผ ๊ธฐ๋ฐ์ผ๋ก ์์ ๋ง์ ํ๋ค์ค๋ฅผ ๊ตฌ์ถํ ๋ ์๊ฐ์ ์ป์ ์ ์๋๋ก ๋์ต๋๋ค.
์ด๋ฌํ ์ ๊ทผ ๋ฐฉ์์ ๋ชจ๋ธ์ โ์ผ๋ฐํ ๊ฐ๋ฅ์ฑ(generalizability)โ์๋ ์ํฅ์ ๋ฏธ์น ์ ์์ต๋๋ค. ํน์ ํ๋ฆฌ๋ฏธํฐ๋ธ์ ๋ชจ๋ธ์ ์ต์ ํํ๋ฉด, ํด๋น ์์ญ์์๋ ๋ฐ์ด๋์ง๋ง ๋ค๋ฅธ ์์ญ์์๋ ์ฑ๋ฅ์ด ์ ํ๋๋ โ๊ฒฝ๋ก ์์กด์ฑ(path dependency)โ์ด ์๊ธธ ์ ์์ต๋๋ค. ๋ฐ๋ผ์ ์ด๋ค ํ๋ฆฌ๋ฏธํฐ๋ธ๋ฅผ ํฌํจํ๊ณ ๋ชจ๋ธ์๊ฒ ๋ ์์ฐ์ค๋ฝ๊ฒ ์ ๊ณตํ ์ง ์ ์คํ๊ฒ ๊ฒฐ์ ํ๋ ๊ฒ์ด ๋งค์ฐ ์ค์ํ๋ค๊ณ ์์ ค๋ผ๋ ๊ฐ์กฐํฉ๋๋ค.
๋ด๋ถ ํ์ฉ ์ฌ๋ก: ์์ฐ์ฑ ํ์ ๊ณผ ํ ๋จ์ ํ์
ํด๋ก๋ ๊ด๋ฆฌํ ์์ด์ ํธ๋ ๋๊ตฌ๋ฅผ ์ํ ๊ฒ์ผ๊น์? Anthropic ํ์ ํฌ๊ฒ ๋ ๊ฐ์ง ์ฌ์ฉ์๋ฅผ ์ผ๋์ ๋ก๋๋ค. ์ฒซ์งธ, ๊ธฐ์ ๋ด๋ถ์์ ์๋ํ(automation)๋ ๊ฐ๋ ฅํ ํ๋ซํผ ๋ฐ ์์คํ ์ ๊ตฌ์ถํ๋ ค๋ ์ฌ๋๋ค์ ๋๋ค. ์๋ฅผ ๋ค์ด, ์์ ํ ์๋-ํฌ-์๋(end-to-end) ์ํํธ์จ์ด ๊ฐ๋ฐ ํ๋ซํผ์ด๋ ๋ฒ๋ฌดํ์ ๋ง์ผํ ์๋ฃ ๊ฒํ ์๋ํ์ ๊ฐ์ ์๋๋ฆฌ์ค์ ๋๋ค. ๋์งธ, ๊ณ ๊ฐ์๊ฒ ๋ ธ์ถ๋ ์ ํ์ ์์ด์ ํธ๋ฅผ ํตํฉํ๋ ค๋ ์ฌ๋๋ค์ ๋๋ค. ์ด๋ค์ ์ฌ์ ํ ๋ง์ ๋ง์ถค ์ค์ ๊ณผ ๊ฐ๋ ฅํ ๊ธฐ๋ฅ์ด ํ์ํ์ง๋ง, ์ธํ๋ผ ๊ตฌ์ถ์ด๋ ํ๋ค์ค ์์ง๋์ด๋ง์ ์์ง๋์ด๋ง ๋ฆฌ์์ค๋ฅผ ๋ญ๋นํ๊ณ ์ถ์ด ํ์ง ์์ต๋๋ค.
Anthropic์ CMA์ โํต ์คํํธ(quick start)โ ๊ฒฝํ์ ํตํด ๊ธฐ์ ์ ๋ฌด์ ๊ด๊ณ์์ด ๋๊ตฌ๋ ์์ด์ ํธ์ ํ๋ฆฌ๋ฏธํฐ๋ธ๋ฅผ ์ดํดํ๊ณ ํ์ฉํ ์ ์๋๋ก ๋์ต๋๋ค. ์ธํฐ๋ทฐ ์งํ์๊ฐ ์ฝ๋์์ค(Codeex)๋ฅผ ํตํด CMA๋ฅผ ์ค์ ํ๊ณ Slack ๋ด์ ๋น ๋ฅด๊ฒ ์๋์ํจ ์ฌ๋ก๋ ์ด๋ฌํ ๊ฐํธํจ์ ์ ๋ณด์ฌ์ค๋๋ค.
ํฅ๋ฏธ๋ก์ด ์ ์ ๊ฐ๋ฐ์๋ค์ด ์์ด์ ํธ ๊ตฌ์ถ ์ ์ด๋ ต๋ค๊ณ ์๊ฐํ๋ ๋ถ๋ถ๊ณผ ์ค์ ์ด๋ ค์ด ๋ถ๋ถ์ด ๋ค๋ฅด๋ค๋ ๊ฒ์ ๋๋ค. ๋๋ถ๋ถ์ ํ๋ค์ค ์์ง๋์ด๋ง, ์ฆ ๋ชจ๋ธ์์ ์ต๊ณ ์ ์ฑ๋ฅ์ ๋์ด๋ด๋ ๋ถ๋ถ์ ์ด๋ ต๊ฒ ์๊ฐํฉ๋๋ค. ํ์ง๋ง Anthropic ํ์ ์ค์ ๋ก๋ โ๋ชจ๋๊ฐ ์ธํ๋ผ์ ๋ฒฝ์ ๋ถ๋ชํ๋คโ๊ณ ์ง์ ํฉ๋๋ค. ์๋ฒ๋ฅผ ๊ณ์ ์คํํ๊ฑฐ๋, ํ์์ ๋ฐ๋ผ ์ธํ๋ผ๋ฅผ ํ์ฅ/์ถ์ํ๊ณ , ์คํฌ๋ฆฝํธ ๋ฐ์ดํฐ๋ฅผ ์ ์ฅํ๋ฉฐ, ๋ณด์ ์๋๋ฐ์ฑ(sandboxing)์ ์ ๊ณตํ๋ ๋ฑ์ ์ธํ๋ผ ๋ฌธ์ ๋ ํ๋กํ ํ์ ์ ๋์ด ์ค์ ์ ํ์ผ๋ก ํ์ฅํ ๋ ํฐ ๊ฑธ๋ฆผ๋์ด ๋ฉ๋๋ค.
Anthropic ๋ด๋ถ์์๋ ๋ฒ๋ฌดํ์ ๋ง์ผํ ์๋ฃ ๊ฒํ ์์ด์ ํธ์ ๊ฐ์ ํฅ๋ฏธ๋ก์ด ์ฌ๋ก๋ฅผ ํตํด CMA์ ๊ฐ์น๋ฅผ ์ ์ฆํ๊ณ ์์ต๋๋ค. ๊ณผ๊ฑฐ์๋ ๋ง์ผํฐ๊ฐ ๋ฒ๋ฌดํ์ ๊ฒํ ์์ฒญ ํฐ์ผ์ ์ด์์ง๋ง, ์ด์ ๋ ์์ด์ ํธ ์ฑ์ ์๋ฃ๋ฅผ ์ ์ถํ๋ฉด ์์ด์ ํธ๊ฐ 1์ฐจ ๊ฒํ ๋ฅผ ์ํํ๊ณ , ํ์ํ ๊ฒฝ์ฐ์๋ง ๋ฒ๋ฌดํ์ ๊ฒํ ํจ์ผ๋ก ๋ณด๋ ๋๋ค. ์ด๋ ๋จ์ํ โ์คํฌ(skill)โ์ ๋์ด์ญ๋๋ค. ์ธ๊ฐ ๊ฐ์ (human-in-the-loop)์ด ํ์ํ๊ณ , ์ฌ๋ฌ ์ฌ๋์ด ํ์ ํ๋ฉฐ, ์์ด์ ํธ๊ฐ ์คํ๋ ์ ์ ํ โํ์ ์์(form factor)โ๊ฐ ํ์ํ๊ธฐ ๋๋ฌธ์ ๋๋ค.
์ด๋ฌํ ์์ด์ ํธ์ ์์ ๊ถ๊ณผ ์ ์ง๋ณด์๋ ๋ ๋ค๋ฅธ ์ค์ํ ๋ฌธ์ ์ ๋๋ค. Anthropic์ โ๋๊ตฌ๋ PR(Pull Request)์ ์ด์ด ์์ด์ ํธ๋ฅผ ๊ฐ์ ํ ์ ์๋โ ๋ฌธํ๋ฅผ ์ฅ๋ คํฉ๋๋ค. ๋ฒ๋ฌดํ ์์ด์ ํธ์ ๊ฒฝ์ฐ, ์ฒ์์๋ ๊ฐ๋ฐํ์ด ๊ตฌ์ถํ์ง๋ง, ์ดํ ๋ฒ๋ฌดํ ์ฌ๋๋ค์ด ์ง์ ํด๋ก๋ ์ฝ๋(Claude Code)๋ฅผ ์ด์ด ์์ด์ ํธ๋ฅผ ๊ฐ์ ํ๋ ์ผ์ด ๋ฐ์ํ๋ค๊ณ ํฉ๋๋ค. ์ด๋ ๊ฐ๋ฐํ์ ๋ถ๋ด์ ์ค์ด๊ณ , ์์ด์ ํธ์ ์ฃผ ์ฌ์ฉ์๊ฐ ์ง์ ๊ฐ์ ์ ์ฐธ์ฌํ๊ฒ ํจ์ผ๋ก์จ ์์ด์ ํธ๊ฐ ์๋์ ๋ค๋จ์ด์ง๋ ๊ฒ์ ๋ฐฉ์งํ๋ ํจ๊ณผ์ ์ธ ๋ฐฉ๋ฒ์ ๋๋ค.
๋ ๋์๊ฐ, Anthropic์ ์ฌ์ฉ์๊ฐ ์ง์ ๋ณต์กํ ์ฝ๋๋ฅผ ๊ฑด๋๋ฆฌ์ง ์๊ณ ๋ ์์ด์ ํธ์ ์ํธ์์ฉํ ์ ์๋๋ก ๋ค์ค ์์ด์ ํธ ์ค์ผ์คํธ๋ ์ด์ (multi-agent orchestration)์ ํ์ฉํฉ๋๋ค. ์ฌ์ฉ์๋ ๋ง์น ํด๋ก๋์๊ฒ ์ง์ ๋งํ๋ ๊ฒ์ฒ๋ผ ์์ด์ ํธ์ ๋ํํ์ง๋ง, ๊ทธ ์ด๋ฉด์์๋ ์ฌ๋ฌ ํด๋ก๋๊ฐ ์๋ก ํ๋ ฅํ์ฌ ๋ณต์กํ ์์ ์ ์ฒ๋ฆฌํฉ๋๋ค. ์ด๋ โ๊ด๋ฆฌํ ์์ด์ ํธ๊ฐ ๊ด๋ฆฌํ ์์ด์ ํธ ์์ ๊ตฌ์ถ๋๋(managed agents all the way down)โ ํํ๋ฅผ ํตํด ๊ฐ๋ฅํด์ง๋๋ค.
ํ๋ซํผ์ ๋ฏธ๋: ์์จ์ฑ๊ณผ ๊ฐํธํ ๋ฐฐํฌ
Anthropic ํ์ AI ํ๋ซํผ์ ๋ฏธ๋๋ฅผ ๋์ฑ ์์จ์ ์ด๊ณ ๊ฐํธํ ๋ฐฐํฌ๊ฐ ๊ฐ๋ฅํ ์ธ์์ผ๋ก ๊ทธ๋ฆฝ๋๋ค. ํ์ฌ๋ ์ธํ๋ผ ๋ฌธ์ ๋ฅผ ํด๊ฒฐํ๋ ๋ฐ ์ง์คํ๊ณ ์์ง๋ง, ๊ถ๊ทน์ ์ผ๋ก๋ โํด๋ก๋์๊ฒ Slack์ ์ถ๊ฐํ๋ผ๊ณ ๋งํ๋ฉด ํด๋ก๋๊ฐ ๋ชจ๋ ๊ฒ์ ์ฒ๋ฆฌํ์ฌ Slack ๋ด์ด ๋ฐ๋ก ์๋ํ๋โ ์์ค์ ๊ฐํธํจ์ ๋ชฉํ๋ก ํฉ๋๋ค. ์ด๋ฅผ ์ํด ์๊ฒฉ ์ฆ๋ช (credentials)์ ์์ ํ๊ฒ ์ ์ฅํ๋ โ๋ณผํธ(Vaults)โ์ ๊ฐ์ ํ์ ์์ค์ ํ๋ฆฌ๋ฏธํฐ๋ธ๋ถํฐ ์ฐจ๊ทผ์ฐจ๊ทผ ๊ตฌ์ถํด๋๊ฐ๊ณ ์์ต๋๋ค.
์ผ์ดํ๋ฆฐ์ Vercel์ CEO์ธ ๊ธธ(Guillermo Rauch)์ด ์์ ์ ํ์ฌ๋ฅผ โAI ์ํํธ์จ์ด ํฉํ ๋ฆฌโ๋ผ๊ณ ๋ฌ์ฌํ ๊ฒ์ ๊ณต๊ฐํ๋ฉฐ, Anthropic ์ญ์ ๊ธฐ์ ๋ด๋ถ์์ ์์ฒญ๋ ์์ฐ์ฑ์ ์ฐฝ์ถํ๋ AI ๊ธฐ๋ฐ์ ์ํํธ์จ์ด ํฉํ ๋ฆฌ๋ฅผ ์งํฅํ๊ณ ์๋ค๊ณ ๋งํฉ๋๋ค. ์ด๋ ๊ฐ๋ณ ๊ฐ๋ฐ์์ ์์ฐ์ฑ์ ๋์ด ํ๊ณผ ์กฐ์ง ์ ์ฒด์ ํ๋ก์ธ์ค๋ฅผ ํ์ ํ๋ ๊ฒ์ ์๋ฏธํฉ๋๋ค.
Anthropic์ ํด๋ก๋ ํ๋ซํผ์ ๋จ์ํ ๋ชจ๋ธ ์ ๊ณต์ ๋์ด, AI ์์ด์ ํธ๊ฐ ์ง์ ์ผ๋ก ์์จ์ ์ด๊ณ ๊ฐ๋ ฅํ๋ฉฐ, ๋์์ ๊ฐ๋ฐ์๊ฐ ์ฝ๊ฒ ์ ๊ทผํ๊ณ ํ์ฉํ ์ ์๋ ํ๊ฒฝ์ ๊ตฌ์ถํ๋ ๋ฐ ์ง์คํ๊ณ ์์ต๋๋ค. ๋งฅ ๋ฏธ๋์์ ์์๋ ๊ฐ์ธ์ ์คํ์ด ํด๋ผ์ฐ๋ ๊ท๋ชจ์ ์ ํ์ผ๋ก ํ์ฅ๋ ๋ ์ง๋ฉดํ๋ ๋ชจ๋ ์ธํ๋ผ ๋ฐ ๊ด๋ฆฌ ๋ฌธ์ ๋ฅผ ํ๋ซํผ์ด ํด๊ฒฐํจ์ผ๋ก์จ, ๊ฐ๋ฐ์๋ค์ ์ค์ง ์ฐฝ์์ ์ธ ์์ด์ ํธ ๊ฐ๋ฐ์๋ง ์ง์คํ ์ ์๊ฒ ๋ ๊ฒ์ ๋๋ค. Anthropic์ด ๊ทธ๋ฆฌ๋ ๋ฏธ๋๋ AI ์์ด์ ํธ๊ฐ ์ฐ๋ฆฌ ์ถ์ ๋ชจ๋ ์์ญ์์ ๊ฐ๋ ฅํ ์กฐ๋ ฅ์๊ฐ ๋๋ ์ธ์์ด๋ฉฐ, ๊ทธ ๊ธฐ๋ฐ์ ๋ค์ง๋ ์ฌ์ ์ ์ด์ ๋ง ์์๋์์ต๋๋ค.
โ674. How Does a Composer Feel After the World Premiere? | Freakonomics Radioโ โ Freakonomics Radio Network ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
์ ๋ค ์ค๋ฏธ์ค โ๊ตญ๋ถ๋ก โ์ ๋ ธ๋ํ๋ค: ์๊ณก๊ฐ ๋ฐ์ด๋น๋ ๋ญ, ์ธ๊ณ ์ด์ฐ ํ์ ์ธ๊ฐ์ ์ธ ์ฌ์
์ง๋ 3์ ๋ง, ํ๋ ์์ ๊ณ์ ๊ฑฐ์ฅ ๋ฐ์ด๋น๋ ๋ญ(David Lang)์ 250๋ ๋ ๊ณ ์ , ์ ๋ค ์ค๋ฏธ์ค(Adam Smith)์ ใ๊ตญ๋ถ๋ก (The Wealth of Nations)ใ์ ๊ธฐ๋ฐํ ์๋ก์ด ์ค๋ผํ ๋ฆฌ์ค(oratorio)๋ฅผ ์ธ์์ ์ ๋ณด์์ต๋๋ค. ๋ด์ ํํ๋ชจ๋(New York Philharmonic)๊ณผ ์ธ๊ณ์ ์ธ ์งํ์ ๊ตฌ์คํ๋ณด ๋๋ค๋ฉ(Gustavo Dudamel)์ด ํจ๊ปํ ์ด ์ํ์ ์ ์ ๋งค์ง์ ๊ธฐ๋กํ๋ฉฐ ๋ค ์ฐจ๋ก์ ์ฑ๊ณต์ ์ธ ๊ณต์ฐ์ ๋ง์ณค์ต๋๋ค. ๊ทธ๋ฌ๋ ํ๋ คํ ์ธ๊ณ ์ด์ฐ์ ๋ง์ด ๋ด๋ฆฐ ํ, ์๊ณก๊ฐ ๋ญ์ ๋ง์์์๋ ๋ป๋ฐ์ ๊น์ ๊ฐ์ ์ด ์๋ฆฌ ์ก์์ต๋๋ค. ์ด ๊ธฐ์ฌ๋ ใํ๋ฆฌ์ฝ๋ ธ๋ฏน์ค ๋ผ๋์ค(Freakonomics Radio)ใ์์ ์ธํฐ๋ทฐ๋ฅผ ํตํด ๋ญ์ด ๋ค๋ ค์ค ์ด์ฐ ํ์ ์์งํ ์ํ์, ๊ฒฝ์ ํ ๊ณ ์ ์ด ํ๋ ์์ ์ผ๋ก ์ฌํ์ํ๋ ๊ณผ์ ์์ ๋๋ฌ๋ ์์ ๊ณผ ์ฌํ, ๊ทธ๋ฆฌ๊ณ ์ธ๊ฐ ๋ณธ์ฐ์ ๋ํ ์ฑ์ฐฐ์ ์ฌ์ธต์ ์ผ๋ก ํ๊ตฌํฉ๋๋ค.
์ธ๊ณ ์ด์ฐ์ ๋ง์ด ์ค๋ฅด๋ค: ์ ๋ค ์ค๋ฏธ์ค๋ฅผ ์ํ ํ๋ ์ค๋ผํ ๋ฆฌ์ค
๋ฐ์ด๋น๋ ๋ญ์ ์ค๋ผํ ๋ฆฌ์ค ใ๊ตญ๋ถ๋ก ใ์ ์ง๋๋ฌ ๋ด์ ํํ๋ชจ๋์ ์ฐ์ฃผ๋ก ์ฑ๊ณต์ ์ธ ์ธ๊ณ ์ด์ฐ์ ๋ง์ณค์ต๋๋ค. ์ด ๋ ํนํ ์ํ์ ์ ๋ค ์ค๋ฏธ์ค์ ๋ฐฉ๋ํ ์ ์์ ์์ ์ธ์ด๋ก ๋ฒ์ญํ๋ฉฐ, ๊ฒฝ์ ํ์ด๋ผ๋ ๋ค์ ๊ฑด์กฐํ ์ ์๋ ์ฃผ์ ์ ์ธ๊ฐ์ ์ธ ๊ฐ์ฑ๊ณผ ๊น์ด๋ฅผ ๋ถ์ด๋ฃ์์ต๋๋ค. ๊ณต์ฐ์ ๊ตฌ์คํ๋ณด ๋๋ค๋ฉ์ ์งํ ์๋ ๋จ๊ฑฐ์ด ํธ์์ ์ป์๊ณ , ์ฒญ์ค๋ค์ ์ด๋ก์ ์ธ ์กฐํฉ์ ๋๋ผ์๊ณผ ๊ฐ๋์ ๋์์ ํํ์ต๋๋ค.
๋ก๋น์์ ๋ง๋ ํ ํ์ ๊ด๊ฐ์ โํฌ๋ ์ ด๋(crescendo)๊ฐ ์์ฌ ๋ง์ง๋ง์ ๊ฑฐ๋ํ ์๊ฐ์ ๋ง๋๋ ๋ฐฉ์์ด ์ ๋ง ์ข์๋คโ๋ฉฐ, โ์ด ์ํ์ ์ฐ๋ฆฌ๊ฐ ํ์ธ์ ๋์ธ ๋ ๋ ๋์ ๊ธฐ๋ถ์ ๋๋๋ค๋, ์ธ๊ฐ์ ๊ฐ์ฅ ์๋ฆ๋ค์ด ๋ฉด๋ชจ๋ฅผ ์ด์ผ๊ธฐํ๋คโ๊ณ ๊ฐํํ์ต๋๋ค. ๋ ๋ค๋ฅธ ๊ด๊ฐ์ โ๋ด์ ํํ๋ชจ๋์ด ๋ง๋ ํ ํด์ผ ํ ์ข ๋ฅ์ ์ํโ์ด๋ผ๋ฉฐ, โ๊ตฌ์คํ๋ณด ๋๋ค๋ฉ์ด ์ ๋ง ๋ชฐ์ ํด์ ์งํํ๋ ๋ชจ์ต๋ ์ธ์์ ์ด์๋คโ๊ณ ํํ์ต๋๋ค.
๋ธ๋ฃธ๋ฒ๊ทธ(Bloomberg)์ ์คํ ์ด์ ๋ฐ๋ ์ค๋ฏธ์ค(Stacy Vanick Smith)๋ ํ๋ก ์์ ๋ญ์ด โ์ฃผ์ ์์ธํ์ ์ค์ ๋ณด๊ณ ์๋ก ์ถ์๋๊ธฐ ์ฌ์ด ๊ฒฝ์ ํ์ด ์ฌ์ค์ ์ง๊ทนํ ์ธ๊ฐ์ ์ผ ์ ์์์ ๋ณด์ฌ์ฃผ์๋คโ๊ณ ๊ทน์ฐฌํ์ต๋๋ค. ์ด๋ ๋ฏ ๋นํ๊ฐ์ ์ฒญ์ค ๋ชจ๋์๊ฒ์ ๊ธ์ ์ ์ธ ๋ฐ์์ ์ด๋์ด๋ด๋ฉฐ, ใ๊ตญ๋ถ๋ก ใ ์ค๋ผํ ๋ฆฌ์ค๋ ํ๋ ์์ ๊ณ์ ์ ์ ํ ์ถฉ๊ฒฉ์ ์๊ฒผ์ต๋๋ค.
์ฑ๊ณต ๋ค์ ์ฐพ์์จ ๊ณตํํจ: ์๊ณก๊ฐ์ โ์ด์ฐ ํ ์ฐ์ธ์ฆโ
๋ชจ๋์ ์ฐฌ์ฌ ์์์, ์ ์ ์๊ณก๊ฐ ๋ฐ์ด๋น๋ ๋ญ์ ์ด๋ค ๊ฐ์ ์ ๋๊ผ์๊น์? ๊ทธ๋ ์ด์ฐ ์งํ์ ๊ธฐ๋ถ์ โ์ง๋์ฃผ ๋๋ ์ํผ์คํ์์ง๋ง, ์ด๋ฒ ์ฃผ ๋๋ ์๋ฌด๊ฒ๋ ์๋๋คโ๋ผ๊ณ ํํํ์ต๋๋ค. ์๋ ๊ฐ ๊ณต๋ค์ฌ ์์ ํ ๋์์ด ์๋ช ์ ์ป์๋ค๊ฐ ๋ค์ ์ฌ๋ผ์ง๋ ๊ฒฝํ์ ๋ญ์๊ฒ โ์ด์ฐ ํ ์ฐ์ธ์ฆ(post-premiere depression)โ์ด๋ผ๋ ๊น์ ๊ณตํํจ์ ์๊ฒจ์ฃผ์์ต๋๋ค. ๊ทธ๋ โ์ํ์ด ํฌ๊ณ ์ฑ๊ณต์ ์ผ์๋ก, ๊ทธ ๊ณตํํจ์ ๊ตฌ๋ฉ์ด๋ ๋ ๊น์ด์ง๋คโ๊ณ ๊ณ ๋ฐฑํ์ต๋๋ค.
๋ญ์ ์ฑ๊ณต์ ์ ์๋ฅผ โ๋ชจ๋๊ฐ ๋๋ฅผ ๋ฌด๋ฅํ๋ค๊ณ ์๊ฐํ์ง ์๋ ๊ฒโ์ด๋ผ๊ณ ๋ฎ๊ฒ ์ค์ ํ๋ค๊ณ ๋งํ์ต๋๋ค. ์ด๋ ๊ทธ์ ์์ ์ด ๋ฐ๋ณต์ ์ธ ๋ถ๋ถ์ด ๋ง์ โ์ค์ํ๋ฉด 10๋ถ ๋ด๋ด ๊ทธ ์ค์๊ฐ ์ง์๋ ์ ์๋คโ๋ ๋ถ์๊ฐ ๋๋ฌธ์ด์์ต๋๋ค. ์๋ชป๋ ํ์(chord)์ ์ค์ผ์คํธ๋ผ์ ๋ฐฐ์นํ๊ฑฐ๋, ์๋์น ์์ ์๋ฆฌ๋ฅผ ๋ง๋ค์ด๋ผ๊น ๋ด ํญ์ ๊ฑฑ์ ํ๋ค๊ณ ํฉ๋๋ค. โ์๋ฌด๋ ๊ทธ ์ํ์ด ์ด๋ ํ์์ง ๋ชจ๋ฅด๊ธฐ์ ๋ค๋ฅธ ์ฌ๋๋ค์ ์์์ฑ์ง ๋ชปํ๊ฒ ์ง๋ง, ๋ด ๋จธ๋ฆฟ์์๋ ๋ด๊ฐ ์๋ชปํ์์ง๋ ๋ชจ๋ฅด๋ ๋ชจ๋ ๊ฒ๋ค์ ์ ๊ฒํ๋ ์ฒดํฌ๋ฆฌ์คํธ๊ฐ ์๋คโ๋ ๊ทธ์ ๋ง์ ์๋ฒฝ์ ์ถ๊ตฌํ๋ ์์ ๊ฐ์ ๊ณ ๋๋ฅผ ๋ณด์ฌ์ค๋๋ค.
ํนํ ๊ทธ๋ ์ง๋ ์ธํฐ๋ทฐ์์ ์ธ๊ธํ๋ โ์ธ ์ฝํธ(woolen coat)โ ์ ์ฅ์ด ์์ ์ด ์คํ๋์ค์์ ๋ถ๋ฅด๋ ๊ฒ์ฒ๋ผ ๊ฐ๋์ ์ด๊ณ ํ์ฑํ๊ฒ ๋ค๋ฆด์ง, ๊ทธ๋ฆฌ๊ณ ๋ฉ์กฐ ์ํ๋ผ๋ ธ ํ๋ขฐ๋ฅด ๋ฐ๋ก (Flure Baron)์ด ๋ฌด๋ฐ์ฃผ๋ก ๋ถ๋ฅด๋ ์๋ฆฌ์(aria)๋ค์ด ์ถฉ๋ถํ ๊ฐ์ฑ์ ์ผ์ง ๊ฐ์ฅ ์ผ๋ คํ๋ค๊ณ ๋ฐํ์ต๋๋ค. ๋คํํ ์ด ๋ถ๋ถ๋ค์ ์ฑ๊ณต์ ์ผ๋ก ๊ตฌํ๋์ด ๋ญ์ ๊ธฐ์๊ฒ ํ์ต๋๋ค.
ํต์ฐฐ๊ณผ ๊ฐ๋์ ์๊ฐ๋ค: ์ฒญ์ค๊ณผ ์๊ณก๊ฐ์ ๊ต๊ฐ
์งํ์ ์คํฐ๋ธ ๋ฅ๋(Steven Dubner)๋ ใ๊ตญ๋ถ๋ก ใ ์ค๋ผํ ๋ฆฌ์ค๋ฅผ โ๋ ์นด๋กญ๊ณ ํํค(punky)ํ๋ฉด์๋ ๋งค์ฐ ์ฌ์ธํ๊ณ , ์ถํํ๋ ๋ฏํ๋ฉด์๋ ์ ๋ํ๋ ๋ฏํ๋ฉฐ, ์์ ๊ฐ์ด ๋์น๋ฉด์๋ ํ๊ตฌ์ ์ธ ์ํโ์ด๋ผ๊ณ ๋ฌ์ฌํ๋ฉฐ ๊น์ ๊ฐ๋์ ํํํ์ต๋๋ค. ๊ทธ๋ ์ ๋ค ์ค๋ฏธ์ค์ ๋ํ ์๋ก์ด ์ดํด๋ฅผ ์ป์์ ๋ฟ๋ง ์๋๋ผ, ์๊ณก๊ฐ ๋ฐ์ด๋น๋ ๋ญ์ ๋ํ ์กด๊ฒฝ์ฌ๋ ์ปค์ก๋ค๊ณ ๋ง๋ถ์์ต๋๋ค.
๋ญ์ ์ ์ ์์ ์๋ ๋ชจ๋ ์๋ก์ด ์๋๋ฅผ ์์ํ์ง๋ง, ์ด์ ๋ โ์ค๋ ์๊ณก๊ฐ๋ก์ ํ์ด์ฑ(pacing)๊ณผ ๊น์ด ๊ฐ์ ์ค์ํ ๋ฌธ์ ๋ค์ ๋ณธ๋ฅ์ ์ผ๋ก ์๊ฒ ๋์ด ์์ ๋งํผ ๊ฐ๋ ฌํ๊ฒ ๊ณ ๋ฏผํ์ง ์๋๋คโ๊ณ ๋งํ์ต๋๋ค. ๊ทธ๋ ์ํ์ ์ง์ ์ธ ํ๋ฆ, ์ฆ โ๊ฑฐ๋ํ ์ ๋ค ์ค๋ฏธ์ค์ ์ด์ ์ผ๋ก ์์ํ์ฌ ์ ์ฐจ ๊ฐ์ธ์ ์ธ ๋ฏธ๊ตญ์ ์ฌ์ ๋ก ์๋ ด๋๋ ํํโ๊ฐ ์ฑ๊ณต์ ์ผ๋ก ๊ตฌํ๋ ๊ฒ์ ํฐ ๋ง์กฑ๊ฐ์ ํํ์ต๋๋ค.
๋ฅ๋๋ ๋ฆฌํ์ค ๊ณผ์ ์์ โ์ํธ(SoHo) ์คํ๋์ค์ ์์ ์ข ์ด์ ์ ๋ค์ ์ฐ์ด ๋ฃ์ ํ ์ฌ๋์ด ์ด๋ป๊ฒ 48๋ช ์ ์ฑ์ ๊ฐ, ์๋ง์ ์ฐ์ฃผ์, ๋ ๋ช ์ ์๋ฆฌ์คํธ๊ฐ ๋ชจ์ฌ 2์ฐจ์์ ์ธ ๊ฒ์ 6์ฐจ์์ ์ธ ๊ฒ์ผ๋ก ๋ง๋ค์ด๋ด๋์ง ๊ธฐ์ ๊ฐ๋คโ๊ณ ๋๋ผ์์ ๊ธ์น ๋ชปํ์ต๋๋ค. ๋ญ์ ์ด์ ๋ํด โ๋ฆฌํ์ค์ ๋จ์ํ ์ํ๋ฅผ ๋ฐฐ์ฐ๋ ๊ณผ์ ์ด ์๋๋ผ, ์ฌ๋๋ค์ด ๊ณต๋์ฒด๋ฅผ ์ด๋ฃจ์ด ์ด ์ํ์ ํจ๊ป ๋ง๋ค์ด๊ฐ๋ ๋ฒ์ ๋ฐฐ์ฐ๋ ๊ณผ์ โ์ด๋ผ๊ณ ์ค๋ช ํ๋ฉฐ, ์ด ๊ณผ์ ์ด์ผ๋ง๋ก โ์ฌ๋๋ค ์์ ์ ํจโ์ ๋ณด์ฌ์ฃผ๋ ๊ฒ์ด๋ผ๊ณ ๊ฐ์กฐํ์ต๋๋ค.
์๊ณก๊ฐ, ์งํ์, ๊ทธ๋ฆฌ๊ณ ์ค์ผ์คํธ๋ผ: ํ์ ์ ๋ฏธํ
๋ญ์ ๋ชจ๋ ๋ฆฌํ์ค๊ณผ ๊ณต์ฐ์ ์ฐธ์ํ์ต๋๋ค. ๊ทธ๋ ์ผ๋ฐ์ ์ผ๋ก ์ฃฝ์ ์๊ณก๊ฐ๋ค์ ์์ ์ ์ฐ์ฃผํ๋ ๋ฐ ์ต์ํ ์ฐ์ฃผ์๋ค์ด ์ด์์๋ ์๊ณก๊ฐ์ธ ์์ ์ ์ด๋ป๊ฒ ๋ํ ์ง ๊ถ๊ธํ๋ค๊ณ ๋งํ์ต๋๋ค. ๋๋๊ฒ๋ ๊ตฌ์คํ๋ณด ๋๋ค๋ฉ์ ๋น๋กฏํ ์งํ์๋ค์ ์ข ์ข ๋ญ์๊ฒ ์๊ฒฌ์ ๋ฌป๊ฑฐ๋ โ์ด๋ป๊ฒ ๋ค๋ฆฌ์ธ์? ์ด๋ ๊ฒ ํด๋ณผ๊น์?โ๋ผ๋ฉฐ ๊ทธ์ ์๊ฒฌ์ ๊ตฌํ์ต๋๋ค. ๋ญ์ 99%์ ๊ฒฝ์ฐ โ์๋ฆ๋ต์ต๋๋ค!โ๋ผ๋ฉฐ ์์ง์๊ฐ๋ฝ์ ์น์ผ์ธ์ ๋๋ฐ, ๋ฅ๋๋ ์ด๊ฒ์ด ์ ๋ต์ ์ธ ๊ฒ์ธ์ง ์ง์ฌ์ธ์ง ๊ถ๊ธํดํ์ต๋๋ค.
๋ญ์ โ๋ ์ญ์ ์ํ์ ์ฒ์ ๋ฃ๋ ์ ์ฅ์ด์๊ธฐ์ ํ ๋ฌ๊ฐ ํจ๊ป ์ฐ์ตํ๋ค๋ฉด ๋ ๊ตฌ์ฒด์ ์ธ ์กฐ์ธ์ ํ์ ๊ฒโ์ด๋ผ๋ฉด์๋, โ์ฐ์ฃผ์๋ค์ ์ ๋ง ํ๋ฅญํ๊ณ ์์ ์ ์ด๋ฉฐ, ์ข ์ด ์์ ์๋ฌด๊ฒ๋ ์๋ ์ํ์ ์๋ช ์ ๋ถ์ด๋ฃ๋ ๋ฐฉ๋ฒ์ ์๊ณ ์๋คโ๊ณ ๋งํ์ต๋๋ค. ๊ทธ๋ค์ ์ฐ์ฃผ๊ฐ ์์ ์ด ์์ํ๋ ๊ฒ๊ณผ ๋ค๋ฅด๋๋ผ๋ ์ด๋ฏธ โ๋งค์ฐ ์๊ธฐ ๋์น๋ ๊ณต๊ฐ์์ ์์โํ๊ธฐ ๋๋ฌธ์ ๊ทธ์ ๊ธฐ๋ปค๋ค๊ณ ํฉ๋๋ค.
ํ์ง๋ง ํ ๋ฒ์ ๋๋ค๋ฉ์ด ๋ชจ๋ ์ ์ฅ์ ๋ํด ์์ธํ ์๊ฒฌ์ ์๊ตฌํ์ ๋ โ์์ง์ 1์ ๋ํ ์๊ฒฌ์? ์์ง์ 2๋?โ ํ๋ ์์ ์๋ฐ๊ฐ์ โ๊ณตํฌ์ ์ง๋ ธ๋คโ๊ณ ๋์ค๋ ๋ฅผ ๋จ๊ธฐ๋ ํ์ต๋๋ค. ๊ทธ๋ ๋๋ค๋ฉ์ด ์ด์ ์ ์ฅ์ ์์ ํ๋ ๋์ ํฉ๊ธํ ์ ๋ณด๋ฅผ ๋ค์ ธ ๋ค์ ์ ์ฅ์ ๋ํ ์ฝ๋ฉํธ๋ฅผ ๋ง๋ค์ด๋ด์ผ ํ๋ค๊ณ ํ์ํ์ต๋๋ค.
๋ค ์ฐจ๋ก์ ๊ณต์ฐ์ ๊ฑฐ์น๋ฉฐ ๋ญ์ ์ ์ฐจ ํธ์ํจ์ ๋๊ผ์ต๋๋ค. ์ฒซ๋ ์ ์ํ์ด โ๊ฑฐ๋ํ๊ฒ ๋๊ปด์ก๋คโ๋ฉด, ๋งค์ผ ๋ฐค โ์ ์ ๋ ๋น ๋ฅด๊ฒ ๋๊ปด์ก๋คโ๊ณ ํฉ๋๋ค. ์ค์ ์ฐ์ฃผ ์๋๊ฐ ๋นจ๋ผ์ง ๊ฒ์ ์๋์ง๋ง, ์ํ์ ํ๋ฆ์ ์ต์ํด์ง๋ฉด์ ์๊ฐ์ด ์์๊ฐ์ ์ง๋๊ฐ๋ค๋ ๊ฒ์ ๋๋ค. ๋ํ, ์ฐ์ฃผ์๋ค๊ณผ ์๋ฆฌ์คํธ๋ค์ด ์ํ์์ ๋ฐ๊ฒฌํ๋ โ์ง(quality)โ์ด ๋งค์ผ ๋ฐค ์กฐ๊ธ์ฉ ๋ฌ๋ผ์ง๋ฉฐ, ๋๋ค๋ฉ์ด ๋ฏธ๋ฌํ ๋ณํ๋ฅผ ์ด๋์ด๋ด๊ณ ์๋ก์ด ๊ฐ์ ์ ๋ถ์ด๋ฃ๋ ๋ชจ์ต์ด ๋งค์ฐ ํฅ๋ฏธ๋ก์ ๋ค๊ณ ๋ง๋ถ์์ต๋๋ค.
์์ ๊ณผ ํ์, ๊ทธ๋ฆฌ๊ณ โ๊ตญ๋ถ๋ก โ์ ๋ฉ์์ง: ๋ถํธํ ์ง์ค์ ๋ง์ฃผํ๋ค
์ด๋ฌํ ๋๊ท๋ชจ ์ค๋ผํ ๋ฆฌ์ค๋ฅผ ์ ์ํ๋ ๋ฐ๋ ๋ง๋ํ ๋น์ฉ์ด ์์๋ฉ๋๋ค. ์๋ง์ ์ธ๋ ฅ, ๊ณต๊ฐ, ๊ทธ๋ฆฌ๊ณ ๋ฆฌํ์ค ๋น์ฉ ๋ฑ์ ํฐ์ผ ์์ต๋ง์ผ๋ก๋ ์ถฉ๋นํ๊ธฐ ์ด๋ ต๊ธฐ์, ์์ ์ ๋ถ์ ํ ๊ฐ์ธ๊ณผ ๊ธฐ๊ด์ ํ์(patronage)์ ํฌ๊ฒ ์์กดํ ์๋ฐ์ ์์ต๋๋ค. ๊ทธ๋ฐ๋ฐ ใ๊ตญ๋ถ๋ก ใ์ ๋ถ์ ํ ๊ฐ์ธ๊ณผ ๊ธฐ๊ด์ ๋ํด ๋ค์ โ๊ฐํนํ ๋นํโ์ ๋ด๊ณ ์์ด, ๋ญ์ โ์์ ์ ๋จน์ฌ ์ด๋ฆฌ๋ ์์ ๋ฌผ์ด๋ฏ๋(biting the hand that feeds it)โ ์ํฉ์ด ์๋์ง ๊ฑฑ์ ํ๋ค๊ณ ํฉ๋๋ค.
ํ์ง๋ง ๋ญ์ โ๋์ ์ผ์ ํ๋ ๋ฐ๋ ๋์ด ๋ง์ด ๋๋ ์ฌํ์์, ๋์ ๋ค์ฌ ์ข์ ์ผ์ ํ๋ ๊ฒ์ด ๋์ ์ผ์ผ ์๋ ์๋คโ๊ณ ๋ฐ๋ฐํฉ๋๋ค. ๊ทธ๋ โ์ข๊ณ ๊ณ ๊ทํ ๋ชจ๋ ์ผ์ด ๊ฐ์ธ๊ฒ ์ด๋ฃจ์ด์ ธ์ผ ํ๋ ์ธ์์์๋ ์ด๊ณ ์ถ์ง ์๋คโ๊ณ ๋งํ๋ฉฐ, ์คํ๋ ค ๋ช๋ช ์ด์ฌ๋ค์ด ์์ ์๊ฒ ๊ฐ์ฌ๋ฅผ ํํ๋ค๊ณ ์ ํ์ต๋๋ค. ๊ทธ๋ ์ง๊ธ์ด โ์์ญ์ต ๋ฌ๋ฌ๋ฅผ ๊ฐ์ง ์ฌ๋๋ค๊ณผ 4์ต ๋ฌ๋ฌ๋ฅผ ๊ฐ์ง ์ฌ๋๋ค ์ฌ์ด์ ํฐ ๋ถ์ด๊ณผ ๊ฑฐ๋ฆฌ๊ฐ ์๋ ์๊ธฐโ์ด๋ฉฐ, โ๋์ด ์ด๋ป๊ฒ ๋ ์ ์ฐ์ผ ์ ์๋์ง์ ๋ํด ํจ๊ป ์ด์ผ๊ธฐํ ์ ์๋ ๊ธฐํโ๋ผ๊ณ ๋ณด์์ต๋๋ค.
๋ด์ ํํ๋ชจ๋์ ์ฐ๊ฐ ์์ฐ์ ์ฝ 9์ฒ๋ง ๋ฌ๋ฌ๋ก, ๋ฉํธ๋กํด๋ฆฌํ ์คํ๋ผ(Metropolitan Opera)์ ๊ฐ์ ๋ค๋ฅธ ์์ ๋จ์ฒด๋ค์ฒ๋ผ ์ฌ์ ์ ์ด๋ ค์์ ์ง๋ฉดํด ์์ต๋๋ค. ํ์ง๋ง ๊ตฌ์คํ๋ณด ๋๋ค๋ฉ์ด๋ผ๋ ์ํผ์คํ๋ฅผ ์์ ๋ฐ ์์ ๊ฐ๋ ์ผ๋ก ์์ ํ๋ฉฐ ์๋ก์ด ์ ํ์ ์ ๋ง๊ณ ์์ต๋๋ค. ๋ด์ ํํ๋ชจ๋์ ์ ์ ์ฌ์ฅ ๊ฒธ CEO ๋งํฐ์์ค ํ๋ฅด๋ ธํด์คํค(Matias Tarnopolsky)๋ ๋ฐ์ด๋น๋ ๋ญ์ ใ๊ตญ๋ถ๋ก ใ๊ณผ ๊ฐ์ ๋๊ท๋ชจ ์ ์์ ํฌ์ํ๋ ์ด์ ๋ฅผ โ๋ฏธ๊ตญ ๊ฑด๊ตญ 250์ฃผ๋ ์ ๋ง์ ๋ฏธ๊ตญ๋ฟ ์๋๋ผ ์ ์ธ๊ณ ์ญ์ฌ์ ์๊ฐ์ ์์ ์ผ๋ก ์ด๋ป๊ฒ ๊ธฐ๋ ํ ๊ฒ์ธ๊ฐโ๋ผ๋ ์ง๋ฌธ์์ ์ฐพ์๋ค๊ณ ์ค๋ช ํ์ต๋๋ค.
ํ๋ฅด๋ ธํด์คํค๋ ์์งํ ๋ญ์ ์์ด๋์ด๋ฅผ ์ฒ์ ๋ค์์ ๋ โ๋ฏธ์น ์๊ฐโ์ด๋ผ๊ณ ์ฌ๊ฒผ๋ค๊ณ ๊ณ ๋ฐฑํ์ต๋๋ค. โใ๊ตญ๋ถ๋ก ใ๊ณผ ํจ๋ธ์ ใ๋ฉ์์(Messiah)ใ๋ฅผ ๊ฒฐํฉํ๋ค๋, ์ฐ๋ฆฌ๊ฐ ๋ญ ํ๋ ๊ฑฐ์ง?โ๋ผ๋ ์๊ฐ์ด ๋ค์๋ค๋ ๊ฒ์ ๋๋ค. ํ์ง๋ง ๊ทธ๋ โ์์ ๊ฐ๋ค์ ๋ฏฟ์ด์ผ ํ๋ค๋ ๊ตํ์ด ์๋ค. ๊ทธ๋ค์ ์๋คโ๋ฉฐ, ๋ญ๊ณผ ์ฐ์ฃผ์๋ค์ ์ญ๋์ ์ ๋ขฐํ๋ค๊ณ ๋งํ์ต๋๋ค. ๊ทธ๋ ์ํ์ ๋ค ๋ฒ ๋ชจ๋ ๊ด๋ํ๋ฉฐ โ์ ์ฅํ๊ณ ์ฌ๋ ค ๊น์ผ๋ฉฐ ์ ๊ทผํ๊ธฐ ์ฌ์ ๋คโ๊ณ ๊ทน์ฐฌํ์ต๋๋ค.
๋ฅ๋๋ ํํ๋ชจ๋์ ์์ ๋๋ถ๋ถ์ด ์์ ์ฌ์ ๊ฐ๋ค๋ก๋ถํฐ ๋์ค๋๋ฐ, ์ด๋ค์ด ๋ฐ๋ก ์ํ์์ ๋นํํ๋ โ์ง์ค๋ ๋ถ๋ฅผ ๊ฐ์ง ์ฌ๋๋คโ์ด๋ผ๋ ์ ์ ์ง์ ํ๋ฉฐ, ํ์์๋ค์ด ์ด ์ํ์ ์ด๋ป๊ฒ ๋ฐ์๋ค์์์ง ๋ฌผ์์ต๋๋ค. ํ๋ฅด๋ ธํด์คํค๋ โ์๋ฌด๋ ๋ถํํ์ง ์์๋คโ๊ณ ๋ตํ๋ฉฐ, โ์ฌ๋๋ค์ ์ธ์์์ ๋ฌด์จ ์ผ์ด ์ผ์ด๋๋์ง ๋์ ๋จ๊ณ ์๋คโ๊ณ ๋งํ์ต๋๋ค. ๊ทธ๋ โ์ญ์ฌ์ ๊ฒฉ๋๊ธฐ์๋ ์์ ๊ฐ๋ค์ด ์๋์ ํผ๋์ ๋ฌ์ฌํ ๋ฟ๋ง ์๋๋ผ, ์๋ฌด๋ ํ์ง ์๋ ๋ฐฉ์์ผ๋ก ์ฌํ์ ๊ฑฐ์ธ์ ๋น์ถฐ์๋คโ๊ณ ๊ฐ์กฐํ์ต๋๋ค.
์ ๋ค ์ค๋ฏธ์ค์ ์ฌํด์: โ๊ตญ๋ถ๋ก โ์ด ๋์ง๋ ์ฌํ์ ์ง๋ฌธ
์กฐ์งํ์ด ๋ํ๊ต ์ ์นํ์ ๊ธ๋ก๋ฆฌ ๋ฃจ(Glory Lou)๋ ๋ญ์ ใ๊ตญ๋ถ๋ก ใ ์ค๋ผํ ๋ฆฌ์ค๋ฅผ โ๊ธด๊ธํ๊ณ ํธ๊ธฐ์ฌ ๋ง์ผ๋ฉฐ ์ด๋ก๊ณ ์์ ์ ์ด๋ฉฐ ํฌ๋ง์ โ์ด๋ผ๊ณ ํ๊ฐํ๋ฉฐ โ๊ธฐ๋ ์ด์์ด์๋คโ๊ณ ๋งํ์ต๋๋ค. ๊ทธ๋ ์ ์ ์ ใ์ ๋ค ์ค๋ฏธ์ค์ ๋ฏธ๊ตญ(Adam Smithโs America)ใ์์ ๋ฃจ๋ ์ ๋ค ์ค๋ฏธ์ค๊ฐ ์ข ์ข ์ขํ์ ์ฐํ, ์์ ์์ฅ ์นํธ์ ๋ฐ ๋นํ๊ฐ๋ค์ ์ํด ์ ํ์ ์ผ๋ก ํด์๋๋ค๊ณ ์ฃผ์ฅํฉ๋๋ค. ๊ทธ๋ ๋ ๋ญ์ ์ํ์ด ์ด๋ฌํ ์ผ๋ฐ์ ์ธ ํจ์ ์ ํผํ๋ค๊ณ ๋ณด์์ต๋๋ค.
๋ฃจ๋ โ๋๋ถ๋ถ์ ์ฌ๋๋ค์ด ์ ๋ค ์ค๋ฏธ์ค์๊ฒ ๊ธฐ๋ํ๋, ์์ฅ์ โ๋ณด์ด์ง ์๋ ์(invisible hand)โ์ด ์ ๋ถ์ ๊ฐ์ ๋ณด๋ค ์ฐ์ํ๋ค๋ ๋จ์ํ ๋ฉ์์ง๊ฐ ์๋์๋คโ๋ฉฐ, โ์ด ์ํ์ ์ค๋ฏธ์ค์ ๋ํ ์ฐ๋ฆฌ์ ์ ์ ๊ฒฌ์ ๋ฐ๊พธ๊ณ , ๊ฐ๋ ์ ์๋ฏธ์ ์๋์ ๋ฐ๋ฅธ ์งํ๋ฅผ ํ๊ตฌํ๊ฒ ํ๋คโ๊ณ ๋งํ์ต๋๋ค. ๊ทธ๋ ๋ ์ด ์ํ์ด โ์ฌ๋๋ค์ด ์ ๋ค ์ค๋ฏธ์ค์ ๋ํด ์ด์ผ๊ธฐํ๋ ๋ฐฉ์์ ๋ฐ๊พธ๊ณ , ๋ํ์ ํฌํจ๋๋ ์ฌ๋๋ค์ ๋ณํ์ํค๊ธฐ๋ฅผ ํฌ๋งํ๋คโ๊ณ ๋ง๋ถ์์ต๋๋ค.
๊ธ์ต ์ ๊ณ์์ ์ผํ๋ ๊ด๊ฐ ํธ๋ ์ด์ ํํด(Tracy Fenton)์ ์ํ์ ๊น์ด ๊ฐ๋ํ์ง๋ง, ์ฌํ์ฃผ์ ์ง๋์ ์ ์ง ๋์ค(Eugene Debs)์ ๋ฒ์ ์ฐ์ค์์ ์ธ์ฉํ โ์๋ฌด๊ฒ๋ ํ์ง ์๋ ์ฌ๋๋ค์ด ๋ฐฑ๋ง์ฅ์๊ฐ ๋๋คโ๋ ๊ตฌ์ ์ ๋ํด์๋ โ๋ค์์ ์ผํด๋ณธ ์ฌ๋์ผ๋ก์ ๋์ ์ง๋ฅ์ ์ธ ์์ง์์ด ์ผ๋ง๋ ๋ง์ ์ฌ๋์๊ฒ ๋์์ด ๋๋์ง ์๊ธฐ์ ์ฌ์ค์ด ์๋๋คโ๋ผ๋ฉฐ ์ด์๋ฅผ ์ ๊ธฐํ๊ธฐ๋ ํ์ต๋๋ค.
๋ญ์ ๋์ค์ ์ฐ์ค๋ฌธ์ด ์ํ์ ํต์ฌ ๊ตฌ์ฑ ์์ ์ค ํ๋์๋ค๋ ์ ์ ๋ํด โ์ค๋์ ์ ์ด ๋ค๋ฅธ ์ํ์ ์ ์ํ์ ๋ฃ๋ ๊ฒ์ ์ฝ๊ฐ ์์์(cheating) ๊ฐ๋ค๊ณ ์๊ฐํ๋คโ๊ณ ์์งํ ๋งํ์ต๋๋ค. ํ์ง๋ง ๊ทธ๋ ํจ๋ธ์ด ์์ ์ ์ํ์ ์ฌํ์ฉํ๋ ์ ๋ก๋ฅผ ์ธ๊ธํ๋ฉฐ, โ์ด ์ํ์ ๋ด๊ฐ ๋ฌด์์ ์๊ฐํ๊ณ ๋ฏฟ๋์ง ์์๋ด๋ ค๋ ์๋์ด๊ธฐ์, ๋ด๊ฐ ์ด์ ์ ์๊ฐํ๋ ๋ชจ๋ ๊ฒ๊ณผ ์ด๋ฏธ ์ด ๋ชจ๋ ์์ ๊ณผ ์๋์ผ๋ก ๋ํํ๋ ๊ฒ๊ณผ ๊ฐ๋คโ๊ณ ์ค๋ช ํ์ต๋๋ค.
๋ฅ๋๋ ๋์ค์ โ๋๋ ์ฐ๋ฆฌ๊ฐ ์ด๊ณ ์๋ ์ฌํ ์์คํ ์ ๋ฐ๋ํ๋ค. ๊ทผ๋ณธ์ ์ธ ๋ณํ๋ฅผ ๋ฏฟ์ง๋ง, ๊ฐ๋ฅํ๋ค๋ฉด ํํ๋กญ๊ณ ์ง์ ์๋ ์๋จ์ ํตํด์๋คโ๋ผ๋ ๊ตฌ์ ์ด ๋ญ ์์ ์ ์๊ฐ๊ณผ ์ผ์นํ๋์ง ๋ฌผ์์ต๋๋ค. ๋ญ์ โ๊ธฐ๋ณธ์ ์ผ๋ก ๊ทธ๋ ๋ค. ํํ๋กญ๊ฒ ๋ณํ๊ฐ ์ผ์ด๋๋ค๋ฉด ์ ๋ง ์ข์ ๊ฒโ์ด๋ผ๊ณ ๋ตํ์ต๋๋ค. ๊ทธ๋ โ์์ ์ ํน์ ํ ๊ฒ์ ๊ดํ ๊ฒ์ด ์๋๋ผ ์ฐ๋ฆฌ๊ฐ ์ฌ๋ฌผ์ ๋ํด ์ด๋ป๊ฒ ๋๋ผ๋์ง์ ๊ดํ ๊ฒโ์ด๋ผ๋ฉฐ, โ๋์ ์ฃผ๊ณ ๋ฐ๋ ์ฌ๋๋ค์ ์ํด ํ์ฑ๋ ๊ณต๋์ฒด์ ๋ํด ์ฐ๋ฆฌ๊ฐ ์ด๋ป๊ฒ ๋๋ผ๋์ง ๋ค๋ฃจ์ง ์๋ ํ, ์์คํ ์ด ์๋ํ๋ ๋ฐฉ์์ ๋ฐ๊ฟ ์ ์๋คโ๊ณ ๊ฐ์กฐํ์ต๋๋ค. ๊ทธ๋ โ์ฐ๋ฆฌ๋ ๋ชจ๋ ๊ฒ์ด ๋ง์ฐฐ ์์ด ์๋ํ๋ค๊ณ ์๊ฐํ๋ฉฐ, ์ฐ๋ฆฌ๋ฅผ ๋๋ด์ฃผ๋ ์ฌ๋์ด๋ ์ฐ๋ฆฌ๊ฐ ๊ณ ์ฉํ ์ฌ๋๋ค์ ์ถ์ ์ฃผ๋ชฉํ์ง ์์ผ๋ ค ํ๋ค. โ๋ณด์ด์ง ์๋ ์โ์ด ๋ชจ๋ ๊ฒ์ ์ฒ๋ฆฌํด์ฃผ๊ธฐ ๋๋ฌธ์ด๋ค. ํ์ง๋ง ์ด์ฉ๋ฉด ๊ทธ๊ฒ์ด ๊ทธ๋ฆฌ ์ข์ ์ผ๋ง์ ์๋ ์๋ ์๋คโ๋ฉฐ, ์๋ก์ ์ถ์ ์ฃผ์๋ฅผ ๊ธฐ์ธ์ฌ์ผ ํ๋ค๊ณ ์ญ์คํ์ต๋๋ค.
โ๊ตญ๋ถ๋ก โ์ ์ข ๊ต์ ์ธ๋ฆผ๊ณผ ์ถ์ ์งํ
๋ฅ๋๋ ใ๋ฉ์์ใ๋ฅผ ์ธ๊ธํ๋ฉฐ ๋ญ์ ใ๊ตญ๋ถ๋ก ใ ์ค๋ผํ ๋ฆฌ์ค๊ฐ ์ด๋ ์ ๋ โ์ข ๊ต์ ์ธ ์ํโ์ธ์ง ๋ฌผ์์ต๋๋ค. ๋ญ์ โ๋ ๋์ ์ธ์์ ์์ํ ์ ์๋ ๋ชจ๋ ์ํฉ์ ์ผ์ข ์ ์ข ๊ต์ ํน์ ์ข ๊ต์ ๋์ฒด๋ฌผ๊ณผ ๊ฐ๋คโ๊ณ ๋ตํ์ต๋๋ค. โ์ฐ๋ฆฌ๊ฐ ์ด๊ณ ์๋ ์ถ๋ณด๋ค ๋ ๋์ ์ถ์ ์ด ์ ์๋ค๋ ์๊ฐ์ ๋งค์ฐ ์๋ฆ๋ต๊ณ , ์ด์๊ฐ๊ธฐ ์ํด ํ์ํ ์๊ฐโ์ด๋ฉฐ, ์ด๊ฒ์ด ๋ฐ๋ก ใ๊ตญ๋ถ๋ก ใ์ ๋ฉ์์ง์ด์ ์ ๋ค ์ค๋ฏธ์ค ์ฑ ์ ๋ฉ์์ง๋ผ๊ณ ๋ณด์์ต๋๋ค. ๋ญ์ ์ค๋ฏธ์ค์ ใ๊ตญ๋ถ๋ก ใ์ด ๋จ์ํ ์ธ์์ด ์ด๋ป๊ฒ ์๋ํ๋์ง์ ๋ํ ์ค๋ช ์ด ์๋๋ผ, โ์ธ์์ ๋ชจ๋ ์ฌ๋์ด ์ด๋ป๊ฒ ์ ์ง๋ผ ์ ์๋์ง์ ๋ํ ๊ทธ์ ์์๋ ฅโ์ด๋ผ๊ณ ์ฌํด์ํ์ต๋๋ค.
๋ญ์ ์์ ์ด ๋ ์คํ ์ ๋๊ต ์ ์๋ก ์๋์ผ๋ฉฐ, ์์ ์ ์ข ๊ต์ ๊ด์ฌ์ โ์ ๋์ธ์ผ๋ก์์ ๋ฐฐ๊ฒฝ์ ์ดํดํ๋ ค๋ ๋ ธ๋ ฅโ์์ ๋น๋กฏ๋๋ค๊ณ ๋งํ์ต๋๋ค. โํ ๋ฐ๋ก ์์ ์ ๋๊ต์ ๋ณธ์ง์ ์ค๋ช ํ๋ผโ๋ ๋ฅ๋์ ์์ฒญ์ ๋ญ์ โ๋ค๊ฐ ์ํ์ง ์๋ ๊ฒ์ ๋จ์๊ฒ ํํ์ง ๋ง๋ผ(Do not do unto others as you would not have them do unto you)โ๋ผ๊ณ ๋ตํ์ต๋๋ค. ๊ทธ๋ ์์ ์ ๋๋์ ๋์นจ๋ฐ์ด ์ข ๊ต์์ ๋ฐฐ์ด ๊ฒ์์ ์จ๋ค๊ณ ๋ฏฟ์ผ๋ฉฐ, ์ ๋๊ต์ ๊ฐ๋ฅด์นจ์ด ์ ๋์ธ๋ฟ๋ง ์๋๋ผ โ์ธ์์ ๋ชจ๋ ์ฌ๋์ด ์ด๋ป๊ฒ ์ด์์ผ ํ๋์ง์ ๋ํ ๋ณดํธ์ ์ธ ์ง์นจโ์ ๋ด๊ณ ์๋ค๊ณ ์๊ฐํฉ๋๋ค.
ํฅ๋ฏธ๋กญ๊ฒ๋ ๋ญ์ ใ๊ตญ๋ถ๋ก ใ์ 18๊ฐ์ ์ ์ฅ์ผ๋ก ๊ตฌ์ฑ๋์ด ์์ต๋๋ค. ์ ๋๊ต์์ ์ซ์ 18์ โํ์ด(Chai)โ ์ฆ โ์๋ช โ์ ์์งํ๋ ์ค์ํ ์ซ์์ ๋๋ค. ๋ญ์ โ๋ง์ ์ํ์ด 18์ ์ฅ์ธ ๊ฒ์ ์๋์ง๋ง, 16~17์ ์ฅ์ฏค ๋์์ ๋ 18์ด๋ผ๋ ์ซ์๊ฐ ์์ฃผ ์ข์ ๋ณด์๋คโ๊ณ ๋งํ๋ฉฐ ์๋์ ์ธ ์ ํ์ด์์์ ์์ํ์ต๋๋ค.
๋ง์ง๋ง ์ ์ฅ, โ์์ฃผ ๋จ์ํ ๋น๋ฐโ: ์ ๋ค ์ค๋ฏธ์ค์ ์ต์ข ๋ฉ์์ง
๋ง์ง๋ง 18๋ฒ์งธ ์ ์ฅ์ธ โ์์ฃผ ๋จ์ํ ๋น๋ฐ(the very simple secret)โ์์ ๋ญ์ ๋ค์ ์ ๋ค ์ค๋ฏธ์ค์ ํ ์คํธ๋ก ๋์๊ฐ๋๋ค. ์ด ์ ์ฅ์ ์ ๋ฌธ์ ๋ค์๊ณผ ๊ฐ์ต๋๋ค.
โ๋๋ถ๋ถ์ ๊ตฌ์ฑ์๋ค์ด ๊ฐ๋ํ๊ณ ๋น์ฐธํ ์ฌํ๋ ๊ฒฐ์ฝ ๋ฒ์ํ๊ณ ํ๋ณตํ ์ ์์ต๋๋ค. ๋ชจ๋ ์ฌ๋๋ค์๊ฒ ๋จน์ ๊ฒ, ์ ์ ๊ฒ, ์ ์๋ฆฌ๋ฅผ ์ ๊ณตํ๋ ์ฌ๋๋ค์ด ์์ ๋ค์ ๋ ธ๋ ์์ฐ๋ฌผ์์ ํฉ๋นํ ๋ชซ์ ๊ฐ์ง๋ ๊ฒ์ ๋น์ฐํ ํํ์ฑ(equity)์ ๋๋ค. ์์ ํ ์ ์(justice), ์์ ํ ์์ (liberty), ์์ ํ ํ๋ฑ(equality)์ ํ๋ฆฝ์ด์ผ๋ง๋ก ๋ชจ๋ ์ด์๊ฒ ์ต๊ณ ์ ๋ฒ์์ ๊ฐ์ฅ ํจ๊ณผ์ ์ผ๋ก ๋ณด์ฅํ๋ ์์ฃผ ๋จ์ํ ๋น๋ฐ์ ๋๋ค.โ
๋ญ์ ์ด ๊ตฌ์ ์ ๋ฐ๊ฒฌํ์ ๋ โ๋งค์ฐ ๊ธฐ๋ปค๋คโ๊ณ ๋งํ์ต๋๋ค. ๊ทธ๋ ํ๋ ์ฌํ์์ ์ ๋ค ์ค๋ฏธ์ค๊ฐ ์น์์ ํจ์๋ฅผ ์ธ์ ํ๊ณ ๋ถํ๋ฑ์ ๋ด์ฌํ ์ฌํ๋ฅผ ์ฉ์ธํ๋ ์ธ๋ฌผ๋ก ์์ฃผ ์ธ๊ธ๋๋ ๊ฒ๊ณผ ๋ฌ๋ฆฌ, ์ค๋ฏธ์ค๊ฐ โ์ผ์ ํ๋ ์ฌ๋๋ค์ด ๊ณต์ ํ๊ฒ ๋์ฐ๋ฐ์์ผ ํ๋คโ๊ณ ์ฃผ์ฅํ์์ ๊นจ๋ฌ์๊ธฐ ๋๋ฌธ์ ๋๋ค. ํนํ โํํ์ฑ(equity)โ์ด๋ผ๋ ๋จ์ด๊ฐ 250๋ ์ ์๋ ์ค์ํ์ผ๋ฉฐ, ์ฐ๋ฆฌ๊ฐ ์กด๊ฒฝํ๋ ์ธ๋ฌผ์ด ๊ฐ์กฐํ ๊ฐ์น์ด๋ฏ๋ก ์ค๋๋ ์๋ ์ฌ์ ํ ์ค์ํ๋ค๊ณ ๊ฐ์กฐํ์ต๋๋ค.
๊ทธ๋ โ์ฐ๋ฆฌ๊ฐ ์๋ฒฝํ ์์คํ ์ ์ฐพ์๋ค๊ณ ์์ํ๋ ์๊ฐ, ์ฐ๋ฆฌ์ ํ๋ฒ์ด ๊ณ ์ ๋์๋ค๊ณ ์์ํ๋ ์๊ฐ, ์ฐ๋ฆฌ๋ ๋ ์ด์ ์กฐ์ ํ๊ฑฐ๋ ๊ท ํ์ ๋ง์ถ๊ฑฐ๋ ํ์ํ ํ์๊ฐ ์๋ค๊ณ ์์ํ๋ ์๊ฐ, ๊ทธ ์์คํ ์ ์ ์ฒด๋๊ณ ์๊ฐ์ด ์ง๋จ์ ๋ฐ๋ผ ๋ถ๊ท ํํด์ง๋คโ๊ณ ๊ฒฝ๊ณ ํ์ต๋๋ค. ๋ญ์ ์ํ์ โ๋์ด ์ข๋ค, ๋ ๋ง์ด ๊ฐ์โ๊ณ ์ธ์น๊ฑฐ๋ โ๋์ด ๋์๋ค, ๋ชจ๋ ์์ ์โ๊ณ ๋งํ๋ฉฐ ๋๋์ง ์์ต๋๋ค. ๋์ ์ ๋ค ์ค๋ฏธ์ค๊ฐ ์ด โ์ ์์ ๋ํ ์์ฃผ ๊ฒธ์ํ ๋ฉ์์งโ๋ก ๋ง์ ๋ด๋ฆฝ๋๋ค.
์์ผ๋ก์ โ๊ตญ๋ถ๋ก โ: ์๊ณก๊ฐ์ ๋ค์ ์ฌ์
์ธ๊ณ ์ด์ฐ์ ์ฑ๊ณต์ ์ผ๋ก ๋ง์น๊ณ ํ๋จ๊ณผ ๋์ค์ ํธํ์ ๋ฐ์ ใ๊ตญ๋ถ๋ก ใ ์ค๋ผํ ๋ฆฌ์ค๋ ์ด์ ์ด๋ค ๋ฏธ๋๋ฅผ ๋ง์ดํ ๊น์? ๋ญ์ ์ด ์ํ์ด โ๋งค์ฐ ์ค๋ซ๋์ ์ด์๋จ์ ๋ง์ ๊ณณ์์, ๋ง์ ์ํฉ์์ ์ฐ์ฃผ๋๊ณ , ์ฌํด์๋ ๊ฒโ์ด๋ผ๋ฉฐ ๊ธฐ๋๋ฅผ ํํ์ต๋๋ค. ์ฌ์ฌ๋ฆ ์์คํ ์์ ์ (Aspen Music Festival)์์๋ ๊ณต์ฐ๋ ์์ ์ ๋๋ค.
๋ญ์ ์ด๋ฌํ ๋๊ท๋ชจ ํ๋ก์ ํธ์ โ์ ๋ง ์ด์ํ ์ โ ์ค ํ๋๋ก, ์๋ง์ ์ฌ๋์ด ๋ชจ์ฌ ๋ชจ๋ ๊ด์ฌ๊ณผ ์ฌ๋์ ์์๋ถ์ง๋ง, โ์ฝ์ํธ๊ฐ ๋๋๋ ์๊ฐ ๋ชจ๋ ๊ฐ์์ ๊ธธ์ ๊ฐ๋คโ๋ ์ ์ ๊ผฝ์์ต๋๋ค. ๊ทธ๋ ์์ ์ด โ์ธ๊ณ์ ์ผ๋ก ์ ๋ช ํด์ก๋ค๊ณ ๋ ์๊ฐํ์ง ์๋๋คโ๋ฉฐ, ํ๋ฆฝ ๊ธ๋์ค(Philip Glass)์กฐ์ฐจ ์๋ฃํ์ ์์ ์๋ฌด๋ ์์๋ณด์ง ๋ชปํ๋ ๊ฒ์ ๋ณด์๋ค๊ณ ๋งํ์ต๋๋ค. ํ์ง๋ง ์ด์ ๋ โํฅ๋ฏธ๋กญ๊ณ ์ด์ํ ์์ด๋์ด๋ฅผ ์ ์ํ๋ฉด ์ฌ๋๋ค์ด ๋ค์ด์ค ์์คโ์ด ๋์๋ค๋ฉฐ ๋ง์กฑ๊ฐ์ ๋๋ฌ๋์ต๋๋ค.
๋ฅ๋๋ ์ง๋ ๋ ์ํผ์๋๋ฅผ ํตํด ๋ญ๊ณผ์ ๋ํ๋ฅผ ๋ง๋ฌด๋ฆฌํ๋ฉฐ, ์ ๋ค ์ค๋ฏธ์ค๊ฐ ์์ธ๊ธฐ๋ฅผ ๋์ด ๋ญ์๊ฒ ๋ง๋ฒ์ ๋ถ๋ ธ๋ค๊ณ ํํ์ต๋๋ค. ๋ญ์ด ์ฒ์์๋ ใ๊ตญ๋ถ๋ก ใ์ ๋ฐ๋ถํ ๋ถ๋ถ๋ค์ ์ด๋ป๊ฒ ๋ค๋ฃฐ์ง ๊ณ ๋ฏผํ์ง๋ง, ๊ฒฐ๊ตญ โ๋ชจ๋ ๊ฒฝ์ ๋ ๊ณต์ ํ ํ๋ ฅ๊ณผ ์ฐ๋ฏผ์ด ์์ด์ผ๋ง ๋ฒ์ฑํ ์ ์๋ ๊ฑฐ๋ํ ๊ณต๋์ฒดโ๋ผ๋ ๋ ํฌ๊ณ ๊น์ ์์ด๋์ด์ ๊ฐํ๋์๋ค๋ ๊ฒ์ ๋๋ค. ์ค๋ฏธ์ค๊ฐ ๊ฐ์ธ ๊ธฐ์ ์ด๋ ์ ๋ถ์ ๊ถ๋ ฅ ์ง์ค์ ๋นํํ๋ฏ, ๋ญ ์ญ์ FDR์ ๋ ์ ๊ท์ ์ฐ์ค๋ฌธ์ ์ธ์ฉํ๋ฉฐ โ์ด ๋๋ผ๋ฅผ ์์ ํ ๋งํผ ๋ถ์ ํ ์ฌ๋๋ค์ด ์๋ค๋ฉด, ๊ทธ๋ค์ด ์ด ๋๋ผ๋ฅผ ์์ ํ๊ฒ ๋ ๊ฒโ์ด๋ผ๋ ๊ฒฝ๊ณ ๋ฅผ ๋ด์์ต๋๋ค.
๋ฐ์ด๋น๋ ๋ญ์ ํ์ฌ ๋ด์ ํํ๋ชจ๋์ด ใ๊ตญ๋ถ๋ก ใ์ ์์ ์ ๋ น์(commercial recording)์ ์ถ์ํ ๊ณํ์ด ์๋์ง ํ์ค์น ์๋ค๋ฉฐ, ๋ค์ ์ฃผ ํ์์์ โ๊ฐ์ฒญํ ์์ โ์ด๋ผ๊ณ ๋ฐํ์ต๋๋ค. ์ด ๋ ํนํ๊ณ ์ฌ์คํ ์ํ์ด ๋ ๋ง์ ์ด๋ค์๊ฒ ๊ฐ๋์ ์ ์ฌํ ์ ์๊ธฐ๋ฅผ ๊ธฐ๋ํด ๋ด ๋๋ค.
โTrump Calls Iran Strikes a โTrifle,โ and the Latest on the Hantavirus Outbreakโ โ New York Times Podcasts ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
ํธ๋ผํ์ โ์ฌ์ํ ์ผโ ๋ฐ์ธ๋ถํฐ ๋ฏธ ๋ถ์ฑ ์๊ธฐ๊น์ง: ๋ด์ํ์์ค๊ฐ ์ ํ๋ ํ ์ฃผ๊ฐ์ ํค๋๋ผ์ธ
์ด๋ฒ ์ฃผ ๋ด์ํ์์ฆ ํค๋๋ผ์ธ์ ๊ตญ์ ์ ์ธ์ ๋ณต์กํ ์คํ๋๋ถํฐ ๋ฏธ๊ตญ ๊ตญ๋ด์ ์๊ธํ ๋ฌธ์ ๋ค, ๊ทธ๋ฆฌ๊ณ ์์์น ๋ชปํ ๋ฌธํ์ ์์๊น์ง ๋ค์ฑ๋ก์ด ์ด์ผ๊ธฐ๋ฅผ ์ ํ์ต๋๋ค. ์ด๋๊ณผ์ ๊ธด์ฅ ๊ด๊ณ ์ ํธ๋ผํ ๋ํต๋ น์ ๋๋ฐ์ ์ธ ๋ฐ์ธ, ๋ ๋ฐ๋ ผ์ ๋ถ์์ ํ ํด์ , ๋ฏธ๊ตญ ๊ณต์ค ๋ณด๊ฑด ์์คํ ์ ์ทจ์ฝ์ฑ์ ๋๋ฌ๋ธ ํํฐ๋ฐ์ด๋ฌ์ค(Hantavirus) ์ฌํ, ๊ทธ๋ฆฌ๊ณ ๋ฏธ๊ตญ ๊ตญ๊ฐ ๋ถ์ฑ์ ์ฌ๊ฐํ ๊ฒฝ๊ณ ๋ฑ๊น์ง, ๊ฒฉ๋์ ํ ์ฃผ๋ฅผ ๋ค์ฌ๋ค๋ด ๋๋ค.
๊ตญ์ ์ ์ธ์ ๋ฏธ๊ตญ์ ์ธ๊ต ์ ์ฑ
์ด๋๊ณผ์ โ์ฌ์ํ ์ผโ?: ํธ๋ผํ ํ์ ๋ถ์ ์ค๋ ์ ๋ต
์ง๋์ฃผ, ๋ฏธ๊ตญ๊ณผ ์ด๋์ ํธ๋ฅด๋ฌด์ฆ ํดํ์์ ๊ต์ ์ ๋ฒ์์์๋ ๋ถ๊ตฌํ๊ณ , ๋๋๋ ํธ๋ผํ ๋ํต๋ น์ ์๊ตญ ๊ฐ์ ํด์ ์ด ์ฌ์ ํ ์ ํจํ๋ค๊ณ ์ฃผ์ฅํ์ต๋๋ค. ํธ๋ผํ ๋ํต๋ น์ ์ด๋๊ณผ์ ์ถฉ๋์ โ์ฌ์ํ ์ผ(a trifle)โ์ด๋ผ๊ณ ๊น์๋ด๋ฆฌ๋ฉฐ, โ๊ทธ๋ค์ด ์ฐ๋ฆฌ๋ฅผ ํ์ฐฎ๊ฒ ์ฌ๊ฒผ์ง๋ง, ์ฐ๋ฆฌ๋ ๊ทธ๋ค์ ์๋ํ๋ค. ๋๋ ๊ทธ๊ฒ์ ์ฌ์ํ ์ผ์ด๋ผ๊ณ ๋ถ๋ฅธ๋คโ๊ณ ๋งํ์ต๋๋ค. ๊ทธ๋ฌ๋ ๊ทธ๋ ๊ณง๋ฐ๋ก ์ด๋์ ๋ํ ์ถ๊ฐ ๊ณต๊ฒฉ ์ํ์ ๊ฐํ๋ฉฐ, ์ด๋์ด โ๋นจ๋ฆฌ ํฉ์์์ ์๋ช ํ๋ ๊ฒ์ด ์ข์ ๊ฒโ์ด๋ผ๊ณ ๊ฒฝ๊ณ ํ์ต๋๋ค.
ํธ๋ผํ ๋ํต๋ น์ด ์ธ๊ธํ ํฉ์์๋ ๋ฏธ๊ตญ์ด ์ ์ํ ์๋ก์ด ํ์์์ผ๋ก, ํธ๋ฅด๋ฌด์ฆ ํดํ์ ์ฌ๊ฐ๋ฐฉ๊ณผ 30์ผ๊ฐ์ ์ ํฌ ์ค๋จ์ ํฌํจํฉ๋๋ค. ์ด๋ ์๊ตญ์ด ๋ณด๋ค ํฌ๊ด์ ์ธ ์ฅ๊ธฐ ํฉ์๋ฅผ ๋์ถํ๊ธฐ ์ํ ์์ ๋ฐฉํธ์ ๋๋ค. ํ์ง๋ง ์ด ์์ ํ์์์ ์ด๋์ ํต ํ๋ก๊ทธ๋จ ๋ฑ ๊ธฐ์กด ํ์์ ๊ต์ฐฉ ์ํ์ ๋น ๋จ๋ ธ๋ ํต์ฌ ์์ ๋ค์ ํด๊ฒฐํ์ง ๋ชปํ๊ณ ์์ต๋๋ค.
๋ฏธ๊ตญ ๋์ค์ด ์ด๋๊ณผ์ ๊ฐ๋ฑ์ ๋ํด ์ ์ ๋ ๋ถ์ ์ ์ธ ์๊ฐ์ ๋ณด์ด๊ณ , ์ ์ ๋น์ฉ์ด ์ฆ๊ฐํ๋ฉด์, ์ฌ์ง์ด ์ผ๋ถ ๊ณตํ๋น ์์๋ค ์ฌ์ด์์๋ ์ธ๋ด์ฌ์ด ๋ฐ๋ฅ๋๊ณ ์์ต๋๋ค. ๋ฏธ์๊ฐ์ฃผ์ ํฐ ๋ฐฐ๋ฟ(Tom Barrett) ํ์์์(์ก๊ตฐ ์ฐธ์ ์ฉ์ฌ)์ ์ฌํด ํ๋ ์ฌ์ ์ ์ง๋ฉดํด ์์ผ๋ฉฐ, ์ ์ ์ข ์์ ์ํ ๋ฒ์์ ์ ์ํ์ต๋๋ค. ๊ทธ๋ ๋ํต๋ น์๊ฒ ์ ์ ์งํ๊ถ์ด ์์์ ์ธ์ ํ๋ฉด์๋, โ์ํ๊ฐ ํ๋ฒ์ ์ญํ ์ ํ์ฌํ์ง ์๊ณ ์๋ ์ ์ฅ์์ ๋๋ฌด ๋ง์ ์น๊ตฌ๋ฅผ ์๊ฒ ํ ์ ์๋คโ๋ฉฐ ์์ ์ฅ์น์ ๋ช ํํ ์ํ์ ์๊ตฌํ์ต๋๋ค.
๋ ๋ฐ๋ ผ์ ๋ถ์์ ํ ํด์ : ์ด์ค๋ผ์๊ณผ ํค์ฆ๋ณผ๋ผ์ ๋์น
๋ ๋ฐ๋ ผ์์๋ ํด์ ํฉ์์๋ ๋ถ๊ตฌํ๊ณ ์ ํฌ๊ฐ ๊ณ์๋๊ณ ์์ต๋๋ค. ํฉ์๊ฐ ํญ๋ ฅ ์ฌํ๋ฅผ ์ด๋ ์ ๋ ์ง์ ์์ผฐ์ง๋ง, ์ด์ค๋ผ์๊ณผ ํค์ฆ๋ณผ๋ผ๋ ๋ ๋ฐ๋ ผ ๋จ๋ถ์์ ๊ฑฐ์ ๋งค์ผ ๊ณต๊ฒฉ์ ์ฃผ๊ณ ๋ฐ๊ณ ์์ต๋๋ค. ํนํ ์ด๋ฒ ์ฃผ ์ด์ค๋ผ์์ ๋ฒ ์ด๋ฃจํธ ์ธ๊ณฝ์ ๊ณต์ต์ ๊ฐํํ์ต๋๋ค. ํค์ฆ๋ณผ๋ผ์ ์ํฅ๋ ฅ์ด ๊ฐํ ์ธ๊ตฌ ๋ฐ์ง ์ง์ญ์ ํ ์ํํธ ๊ฑด๋ฌผ์ด ํ๊ฒฉ๋นํ์ผ๋ฉฐ, ์ด์ค๋ผ์๊ตฐ์ ์ด๋์ ์ง์์ ๋ฐ๋ ๋ฏผ๋ณ๋ ์งํ๊ด์ ๊ฒจ๋ฅํด ์ฌ์ดํ๋ค๊ณ ๋ฐํ์ต๋๋ค. ๊ทธ๋ฌ๋ ์ด ๊ณต๊ฒฉ์ ์ด๋ฏธ ๋ถ์์ ํ ์์ ํด์ ์ ๋์ฑ ์ํํ ์ ์๋ค๋ ์ฐ๋ ค๋ฅผ ๋ณ๊ณ ์์ต๋๋ค. ํธ๋ผํ ํ์ ๋ถ๋ ์ด ์ง์ญ์ ์ํฉ์ ๋ฉด๋ฐํ ์ฃผ์ํ๊ณ ์๋๋ฐ, ๋ ๋ฐ๋ ผ์์์ ์ ํฌ ์ฌ๊ฐ๋ ์ด์ค๋ผ์์ ๋ ๋ฐ๋ ผ ๊ณต์ต ์ค๋จ์ ์ฃผ์ฅํ๋ ์ด๋๊ณผ์ ๋ฏธ๊ตญ ํ์์ ๋ณต์กํ๊ฒ ๋ง๋ค ์ ์๊ธฐ ๋๋ฌธ์ ๋๋ค.
๋ฒ ๋ค์์๋ผ์ ๋ฏธ๋๋ฅผ ๋๋ฌ์ผ ์๋ฐํ ํ์: ์นด๋ผ์นด์ค ๋งค๋ฆฌ์ดํธ ํธํ ์ ํ๊ฒฝ
๋ฏธ๊ตญ ํน์๋ถ๋๊ฐ ๋ฐค์ค์ ๋ฒ ๋ค์์๋ผ์ ์นจํฌํ์ฌ ๋์ฝ๋ผ์ค ๋ง๋๋ก ๋ํต๋ น์ ์ฒดํฌํ ์ง 4๊ฐ์์ด ์ง๋ฌ์ต๋๋ค. ์ดํ ๋ฒ ๋ค์์๋ผ์ ๋ฏธ๋์ ๋ํ ์๋ง์ ์ง๋ฌธ์ด ์์์ง๊ณ ์์ผ๋ฉฐ, ๊ถ๋ ฅ๊ณผ ์ํฅ๋ ฅ์ ์ฐจ์งํ๊ธฐ ์ํ ๊ฐ์ถ์ ์์์ ์ ์น์ , ๊ฒฝ์ ์ ์ผ๋ก ์ด๋ค ๋ณํ๊ฐ ์ผ์ด๋ ์ง ์ ์ธ๊ณ์ ์ด๋ชฉ์ด ์ ๋ฆฌ๊ณ ์์ต๋๋ค.
ํฅ๋ฏธ๋กญ๊ฒ๋ ์ด๋ฌํ ์ํฉ์ด ๊ฐ์ฅ ๊ทน๋ช ํ๊ฒ ๋๋ฌ๋๋ ๊ณณ ์ค ํ๋๋ ์นด๋ผ์นด์ค(Caracas)์ ํ ๋งค๋ฆฌ์ดํธ ํธํ ์ ๋๋ค. ๋ด์ํ์์ฆ์ ์๋ชฌ ๋ก๋ฉ๋ก(Simone Romero) ๊ธฐ์๋ ์ด๊ณณ์ โ๋ฒ ๋ค์์๋ผ ์์ฌ ๊ฒฐ์ ์ ์ ๊ฒฝ ์ผํฐ(nerve center)โ๋ผ๊ณ ๋ฌ์ฌํ๋ฉฐ, โ๋ฏธ๊ตญ ์ธ๊ต๊ด๋ค, ํด์คํด๊ณผ ๋๋ฌ์ค์์ ์จ ์์ ์ ์๋ค, ๋ด์๊ณผ ๋ฐ๋์์ ์จ ๊ธ์ต ์ ๋ฌธ๊ฐ๋ค, ๋ฌธ์ ์ ์๊ธด ๊ทผ์ก์ง์ ๋ฏธ๊ตญ ๋ณด์ ์์๋ค๊น์ง, ๋ง์น ์ํ ์ธํธ์ฅ ๊ฐ๋คโ๊ณ ์ ํ์ต๋๋ค. ๋ค์ ๋ก๊ณ ๋น๋ฐ๋ ์ด ํธํ ์ 32๋ฌ๋ฌ์ง๋ฆฌ ์ถ์ถํ ์คํฌ๋จ๋ธ ์๊ทธ๊ฐ ๋์ค๋ ์กฐ์ ๋ทํ๋ฅผ ๋จน์ผ๋ฉฐ ๊ฑฐ๋๋ฅผ ๋ ผํ๋ ๋ง๋จ์ ์ฅ์๊ฐ ๋์์ต๋๋ค.
ํ์ฌ ๋ฏธ๊ตญ ๋์ฌ๊ด์ด ์ ๊ธฐ๋ฅ์ ํ์ง ๋ชปํ๊ณ ์์ด, ๋ฏธ๊ตญ ์ธ๊ต๊ด๋ค์ ํธํ ์ต์์ธต์ ์์ ์ฌ๋ฌด์ค์ ์ฐจ๋ฆฌ๊ณ ์์ต๋๋ค. ์ด์ฒ๋ผ ๊ถ๋ ฅ์ ํต์ฌ์ ๊ฐ๊น์ด ๋ค๊ฐ๊ฐ๋ ค๋ ์ฌ๋๋ค์ ์๋ง์ด ์ด ํธํ ์ ๋ฒ ๋ค์์๋ผ์ ์ด๋ช ์ ์ข์ฐํ๋ ์ค์ํ ์ฅ์๋ก ๋ง๋ค๊ณ ์์ต๋๋ค. ํ ์ ์น ๊ณ ๋ฌธ์ โ๊ธฐ๋ค๋ ค์จ ๋ณํ์ ์๋น ๋ถ๋ถ์ด ๋งค๋ฆฌ์ดํธ ํธํ ์์๋ถํฐ ์์๋๊ณ ์๋คโ๊ณ ๋งํ๋ฉฐ, ๋ง์ ๋ฒ ๋ค์์๋ผ ๊ตญ๋ฏผ๋ค์ ์ด๋ฌํ ํ์๊ณผ ๊ฑฐ๋๊ฐ ์์ ๋ค์ ์ผ์์ ์ด๋ค ์๋ฏธ๋ฅผ ๊ฐ์ ธ์ฌ์ง ์ง์ผ๋ณด๊ณ ์์ต๋๋ค.
๋ฏธ๊ตญ ๊ตญ๋ด์ ๋ณด๊ฑด ๋ฐ ๊ฒฝ์ ์๊ธฐ
ํํฐ๋ฐ์ด๋ฌ์ค(Hantavirus) ํ์ฐ๊ณผ ๋ฏธ๊ตญ์ ๊ณต์ค ๋ณด๊ฑด ์์คํ
๋ฏธ๊ตญ์ ๊ฐ์ผ๋ณ ์ ๋ฌธ๊ฐ๋ค์ ํ์ฌ์ ํํฐ๋ฐ์ด๋ฌ์ค(Hantavirus) ๋ฐ๋ณ ์ฌํ๊ฐ ๋ฏธ๊ตญ์ด ๋ฏธ๋์ ๊ณต์ค ๋ณด๊ฑด ์๊ธฐ์ ์ผ๋ง๋ ๋๋น๊ฐ ๋ฏธํกํ์ง๋ฅผ ๋ณด์ฌ์ค๋ค๊ณ ๊ฒฝ๊ณ ํฉ๋๋ค. ํฌ๋ฃจ์ฆ์ ์์ ์์๋ ์ด๋ฒ ๋ฐ๋ณ์ผ๋ก 3๋ช ์ด ์ฌ๋งํ๊ณ ์ต์ 5๋ช ์ด ๋ณ์ ๊ฑธ๋ ธ์ต๋๋ค. ์ผ๋ถ ์น๊ฐ๋ค์ ์ ์ธ๊ณ๋ก ํฉ์ด์ ธ ๋ค๋ฅธ ์ง์ญ์์ ๋ชจ๋ํฐ๋ง์ ๋ฐ์์ผ ํ๋ ์ํฉ์ ๋๋ค. ๋ฏธ๊ตญ ๋ด์๋ 6๋ช ์ ์น๊ฐ์ด ์์ด ๋น๊ตญ์ ๋์์ด ์๊ตฌ๋์ง๋ง, ์ ๋ฌธ๊ฐ๋ค์ ์ง๋ณํต์ ์๋ฐฉ์ผํฐ(CDC)์ ๋์์ด โ๋ฆ์ฅ ๋์โ์ด์๋ค๊ณ ์ง์ ํฉ๋๋ค. CDC๋ ๋ฐ๋ณ ์ด๊ธฐ ๊ฑฐ์ ์นจ๋ฌตํ์ผ๋ฉฐ, ์ฒซ ์ฌ๋ง์๊ฐ ๋ฐ์ํ ์ง ํ ๋ฌ ๊ฐ๊น์ด ๋์ด์์ผ ๋์ํ์ ๊ตฌ์ฑํ์ต๋๋ค.
์ ๋ฌธ๊ฐ๋ค์ ๋ฐ์ด๋ฌ์ค ์์ฒด(์ฌ๋ ๊ฐ ์ ์ผ์ด ๋๋ฌผ๋ค๊ณ ์๋ ค์ง)๋ณด๋ค CDC์ ๋ฆ์ฅ ๋์ ์์ ์ ๋ ํฐ ์ฐ๋ ค๋ฅผ ํํ๊ณ ์์ต๋๋ค. ์ ์ง ๊ณต์ค ๋ณด๊ฑด ๊ด๋ฆฌ๋ค์ ํธ๋ผํ ํ์ ๋ถ์ CDC ์ธ๋ ฅ ๋ฐ ์์ฐ ๋ํญ ์ญ๊ฐ์ด ๊ธฐ๊ด์ ์ญ๋์ โ๋ฌด๋ ฅํ(hamstrung)โ์์ผฐ๋ค๊ณ ๋นํํ์ต๋๋ค. ์ด๋ ํฅํ ๋ฐ์ํ ์ ์๋ ๊ณต์ค ๋ณด๊ฑด ๋น์์ฌํ์ ๋ํ ๋ฏธ๊ตญ์ ์ทจ์ฝ์ฑ์ ์ฌ์คํ ๋ณด์ฌ์ฃผ๋ ์ฌ๋ก๋ก ํ๊ฐ๋ฉ๋๋ค.
๊ฒฝ๊ณ ๋ฑ ์ผ์ง ๋ฏธ๊ตญ ๊ตญ๊ฐ ๋ถ์ฑ: ๊ฒฝ์ ์์ฐ๋ ์ด๊ณผ
์๋ก์ด ์ฐ๋ฐฉ ๋ฐ์ดํฐ์ ๋ฐ๋ฅด๋ฉด ๋ฏธ๊ตญ ์ ๋ถ์ ๋ถ์ฑ๊ฐ ๊ตญ๊ฐ์ ์ด ๊ฒฝ์ ์์ฐ๋(total economic output)๋ณด๋ค ์ปค์ง๋ ์ฐ๋ ค์ค๋ฌ์ด ์ด์ ํ์ ๋๋ฌํ์ต๋๋ค. ์ด๋ ๋น์ฅ ์๋ฐํ ์ฌ์ ์๊ธฐ๋ฅผ ์๋ฏธํ์ง๋ ์์ง๋ง, ์ ๋ฌธ๊ฐ๋ค์ ์ฆ๊ฐํ๋ ๋ถ์ฑ๊ฐ ๋ฏธ๋์ ๋ฏธ๊ตญ์ด ๋์ ๋น๋ฆฌ๋ ๋น์ฉ์ ๋ ๋น์ธ๊ฒ ๋ง๋ค ๊ฒ์ด๋ผ๊ณ ๊ฒฝ๊ณ ํฉ๋๋ค. ์ธ์๊ฐ ๋ฏธ๊ตญ์ ์ง์ถ์ ์์ ํ ์ถฉ๋นํ์ง ๋ชปํ๊ธฐ ๋๋ฌธ์ ๋ฏธ๊ตญ์ ๊ณ์ํด์ ๋์ ๋น๋ ค์ผ ํ๋ ์ํฉ์ ๋๋ค.
๋๋ฃ ํ ๋ ๋กฌ(Tony Rom)์ ๊ฒฝ์ ํ์๋ค์ด ์๋ ๋์ ์ด ์ด์ ํ์ ๋ํด ๊ฒฝ๊ณ ํด์์ง๋ง, ์์ฑํด์ ๋ฐ์์ โ๋์ฒด๋ก ๋ฏธ์จ์ โ์ด์๋ค๊ณ ์ ํฉ๋๋ค. ๋ถ์ฑ ํด๊ฒฐ์ ์ํ ์ ์ํ ์กฐ์น์ ํ์์ฑ์ ๋ํด ๋ฏผ์ฃผ๋น๊ณผ ๊ณตํ๋น ๋ชจ๋๋ก๋ถํฐ ๋นํ์ ๋ชฉ์๋ฆฌ๊ฐ ๋์ค์ง ์์์ต๋๋ค. ์ด๋ ํธ๋ผํ ๋ํต๋ น ์ฌ์ ๊ธฐ๊ฐ ๋์ ๊ณตํ๋น์ด ๋ถ์ฑ๋ฅผ ์๋นํ ์ฆ๊ฐ์์ผฐ๋ค๋ ์ ์น์ ํ์ค์ ๋ฐ์ํ๋ ๊ฒ์ ๋๋ค. ๊ตญ๊ฐ ๋ถ์ฑ๋ ์์ผ๋ก ๋์ฑ ์ ํ๋ ๊ฒ์ด๋ฉฐ, ์ ์ฑ ์ ์์๋ค์ ์ด๋ฌํ ์ํฉ์ ๊ท ํ์ ๋์ฐพ๊ธฐ ์ํด ํ์ํ โ๋ํํ(grand compromise)โ์๋ ํน๋ณํ ๊ด์ฌ์ด ์์ด ๋ณด์ ๋๋ค.
๋ฌธํ์ ํฅ๋ฏธ๋ก์ด ์ด์ผ๊ธฐ๋ค
ํฐ๋ฆฌ์ฒ์ ์์์ โ์์ ค ๋ค์ดโ: ๋จ ํ ๋ฌธ์ฅ์ผ๋ก ์ฐ์ธ 300ํ์ด์ง ์์ค
์ด๋ฒ ์ฃผ ํฐ๋ฆฌ์ฒ์ ์์์์ด ๋ฐํ๋์์ผ๋ฉฐ, ์์ค ๋ถ๋ฌธ ์์์์ ๋ค๋์ ํฌ๋ผ์ฐ์ค(Daniel Krauss)์ ์์ค โ์์ ค ๋ค์ด(Angel Down)โ์ด์์ต๋๋ค. ์ด ์์ค์ ์ 1์ฐจ ์ธ๊ณ๋์ ์ค ์ฒ ์กฐ๋ง์ ๊ฑธ๋ ค ์ฐ๋ฌ์ง ์ฒ์ฌ๋ฅผ ๋ง๋๋ ํ ๋ณ์ฌ์ ์ด์ผ๊ธฐ๋ฅผ ๋ค๋ฃน๋๋ค. ์ญ์ฌ, ๋ง์ ์ ์ฌ์ค์ฃผ์(magical realism), ๊ทธ๋ฆฌ๊ณ ๊ณต์ ๊ณผํ(science fiction)์ด ํผํฉ๋ ์ด ์ํ์ ๋ฌด๋ ค 300ํ์ด์ง์ ๋ฌํ๋ ๋ถ๋์ด โ๋จ ํ๋์ ๋ฌธ์ฅโ์ผ๋ก ์ฐ์ฌ์ก๋ค๋ ๋ ํนํ ํน์ง์ ๊ฐ์ง๊ณ ์์ต๋๋ค.
ํฌ๋ผ์ฐ์ค ์๊ฐ๋ ๋ด์ํ์์ฆ ๋ถ ๋ฆฌ๋ทฐ ํ์บ์คํธ์์ ์ด๋ฌํ ์งํ ๋ฐฉ์์ ๋ํด ์ค๋ช ํ์ต๋๋ค. ๊ทธ๋ ์ฒ์์๋ ์ ํต์ ์ธ ํ์์ผ๋ก ๊ธ์ ์ฐ๊ธฐ ์์ํ์ง๋ง, ์ฝ 20ํ์ด์ง๋ฅผ ์ฐ๊ณ ๋์ โ๋ชจ๋ ์ฑ ์ด ์๋ํ ํ์๋ ์๋คโ๊ณ ์๊ฐํ ์ ๋๋ก ์ข์ ๊ฐ์ ๋๊ผ๋ค๊ณ ํฉ๋๋ค. ์ฌ๋ฐ๋ฅธ ์ ๊ทผ ๋ฐฉ์์ ์ฐพ์ง ๋ชปํ๋ค๊ณ ๋๋ ๊ทธ๋ ํ๋ก์ ํธ๋ฅผ ์ค๋จํ๊ณ ๊น์ด ๊ณ ๋ฏผํ ๋์, ์ ์์ด ํญ๋ ฅ์ ์ํ(cycle of violence)์ ์์ํ๋ค๋ ์ฑ ์ ์ฃผ์ ์ ์ง์คํ์ต๋๋ค. ๊ทธ๋ฆฌ๊ณ โ๋ง์ฝ ์ฑ ์ ์ฒด๋ฅผ ํ ๋ฌธ์ฅ์ผ๋ก ์ฐ๊ณ , ์ฑ ์ ๋์ด ์์์ผ๋ก ๋ค์ ๋์์ค๊ฒ ํ์ฌ, ๋ ์๊ฐ ์ผ๋จ ์ฝ๊ธฐ ์์ํ๋ฉด ์ฌ์ค์ ์ฑ ์์ ์์ํ ๊ฐํ๋ ๋ฏํ ๋๋์ ์ฃผ์ด ๋์์์ด ์ํํ๋ ๋ฐํด์ฒ๋ผ ๋ง๋ ๋ค๋ฉด ์ด๋จ๊น?โ๋ผ๋ ์์ด๋์ด๋ฅผ ๋ ์ฌ๋ ธ๋ค๊ณ ํฉ๋๋ค.
๋ด์ค ํด์ฆ: ์นด๋ฏธ์นด์ ๋๊ณ ๋๋ถํฐ ๊ตํฉ์ ์ํ ์ ๋ฌด๊น์ง
์ด๋ฒ ์ฃผ ํค๋๋ผ์ธ์ ์ ์พํ ๋ด์ค ํด์ฆ๋ก ๋ง๋ฌด๋ฆฌ๋์์ต๋๋ค.
-
์นด๋ฏธ์นด์ ๋๊ณ ๋ ๋ถ๋?: ์ง๋ ๋ช ๋ฌ๊ฐ ๋ฏธ๊ตญ ๊ด๋ฆฌ๋ค์ ์ด๋๊ณผ์ ์ ์์ ๋ํ ์ง๋ฌธ์ ๋ง์ด ๋ฐ์์ง๋ง, ์ด๋ฒ ์ฃผ์๋ โ์นด๋ฏธ์นด์ ๋๊ณ ๋ ๋ถ๋๊ฐ ์๋์ง ํ์ธํ ์๋ ๋ถ์ธํ ์๋ ์๋คโ๋ ๋ต๋ณ์ด ํ์ ๊ฐ ๋์์ต๋๋ค. ์ฌ์ค ๋ฏธ ํด๊ตฐ์ 1962๋ ๋ถํฐ ๋๊ณ ๋๋ฅผ ํ๋ จ์ํค๋ ํ๋ก๊ทธ๋จ์ ์์ํ๋ ๋ฑ ํด์ ์๋ฌผ์ ํ์ฉํ ์ ๋ก๊ฐ ์์ต๋๋ค. ํด์ฆ๋ ํด๊ตฐ ํ๋ก๊ทธ๋จ์ ์ฐธ์ฌํ์ง ์์ ๋๋ฌผ(๋ฌธ์ด)์ ๋ง์ถ๋ ๊ฒ์ด์์ต๋๋ค. ๋๊ณ ๋๋ 1980๋ ๋ ํ๋ฅด์์๋ง์ ์์ค ๊ฐ์๋ฅผ ์ํด ํ๊ฒฌ๋๊ธฐ๋ ํ์ต๋๋ค.
-
๊ตํฉ์ ํฉ๋นํ ๊ณ ๊ฐ ์๋น์ค ๊ฒฝํ: ์ ๋ช ํ๊ณ ๊ฐ๋ ฅํ ์ฌ๋๋ ๊ณ ๊ฐ ์๋น์ค ๋ฌธ์ (customer service snafus)๋ฅผ ๊ฒช๋๋ค๋ ์ด์ผ๊ธฐ๋ ์๊ฐ๋์์ต๋๋ค. ํ ๋จ์ฑ์ด ์์นด๊ณ ์ง์ญ์์ ๊ฐ์ฐ ์ค, ์์ ์ ์ ๋ช ํ ์น๊ตฌ(๋ณธ๋ช ์ ์๋์ง๋ง)๊ฐ ์ํ์ ์ ํ๋ฒํธ์ ์ฃผ์๋ฅผ ๋ณ๊ฒฝํ๋ ค๋ค ๋๊ด์ ๋ถ๋ชํ๋ ์ผํ๋ฅผ ๊ณต์ ํ์ต๋๋ค. ์ํ ์ง์์ โ์ง์ ๋ฐฉ๋ฌธํด์ผ ํ๋คโ๊ณ ๋งํ๊ณ , ์น๊ตฌ๋ โ๊ทธ๊ฑด ๋ถ๊ฐ๋ฅํ๋คโ๊ณ ๋ตํ์ง๋ง ๊ฒฐ๊ตญ ์ง์์ ์ ํ๋ฅผ ๋์ด๋ฒ๋ ธ์ต๋๋ค. ๊ทธ ์ ๋ช ํ ์น๊ตฌ๋ ๋ฐ๋ก ๊ตํฉ ๋ ์ค(Pope Leo)์์ต๋๋ค. ํด์ฆ๋ โ๋๊ฐ ์ ํ๋ฅผ ๋๊ฒผ๋๊ฐ?โ์๊ณ , ์ ๋ต์ ๊ตํฉ์๊ฒ ์ ํ๋ฅผ ๋์ด๋ฒ๋ฆฐ ์ํ ์ง์์ด๋ผ๋ ๋๋ด์ด์์ต๋๋ค.
-
์ค๋ ๊ธฐ๋ค๋ฆผ ๋์ ์ฐพ์์จ ์ํธ๋ค: ์ง๋์ฃผ ๋ฐ์ค์คํผ์ค 1์ ์ํ๋ 20๋ ๋ง์ ๋์์จ ์ํธ โ์ ๋ง๋ ํ๋ผ๋ค๋ฅผ ์ ๋๋ค 2(The Devil Wears Prada 2)โ์์ต๋๋ค. ํด์ฆ๋ ๋ ์ค๋ ๊ธฐ๋ค๋ฆผ ๋์ ์ํธ์ด ๋์จ ์ํ๋ค์ ๋ง์ถ๋ ๊ฒ์ด์์ต๋๋ค.
- โ์๋ฆฌํธ ์ด๊ธฐ(Elite Firearm)โ (36๋ ): โํ๊ฑด(Top Gun)โ
- โ์นผ๋ ๋ฌ๋ฆฌ๋ ์ฌ๋(Knife Jogger)โ (35๋ ): โ๋ธ๋ ์ด๋ ๋ฌ๋(Blade Runner)โ
- โ๊ณค์ถฉ ์ก์ฒด(Insect Liquid)โ (30๋ ์ด์): โ๋นํ์ฅฌ์ค(Beetlejuice)โ ์คํ๋์ค๊ฐ ์์ญ ๋ ์ด ์ง๋๋ ์ํธ์ ๊ณ์ ์ ์ํ๋ ์ด์ ๋ ๊ด๊ฐ๋ค์ โ์์ง ์๋ ์ด์ (insatiable appetite)โ ๋๋ฌธ์ธ ๊ฒ์ผ๋ก ๋ณด์ ๋๋ค. โ์ ๋ง๋ ํ๋ผ๋ค๋ฅผ ์ ๋๋ค 2โ๋ฟ๋ง ์๋๋ผ, 2024๋ ๋ถ๋ฏธ ํฅํ ์์ 10์๊ถ ์ํ ์ค 9ํธ์ด ์ํธ์ด์๋ค๊ณ ํฉ๋๋ค.
์ด๋ฒ ์ฃผ ๋ด์ํ์์ฆ ํค๋๋ผ์ธ์ ๊ตญ์ ์ ๊ฐ๋ฑ๊ณผ ๊ตญ๋ด์ ๋์ , ๊ทธ๋ฆฌ๊ณ ๋ฌธํ์ ๋ฐํฅ์ด ๊ต์ฐจํ๋ ๋ณต์กํ ํ๋ ์ฌํ์ ๋จ๋ฉด์ ์์ํ๊ฒ ๋ณด์ฌ์ฃผ์์ต๋๋ค. ๋ค์ ์ฃผ ์์์ผ, ๋ ๋ค๋ฅธ ํค๋๋ผ์ธ๊ณผ ํจ๊ป ๋์์ฌ ์์ ์ ๋๋ค.
โHow Prediction Markets Got Out of Controlโ โ Hard Fork ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
AI ์๋, ํต์ ๋ถ๋ฅ์ ์์ธก ์์ฅ๊ณผ ์ถ์ ์ค๋ฉฐ๋ ๊ธฐ์ ์ ์๋ฉด์ฑ
์ธ๊ณต์ง๋ฅ(AI) ๊ธฐ์ ์ด ์ ๋ก ์๋ ์๋๋ก ๋ฐ์ ํ๋ฉฐ ์ฐ๋ฆฌ ์ถ์ ๋ชจ๋ ์์ญ์ ์นจํฌํ๊ณ ์์ต๋๋ค. ํ์ง๋ง ์ด๋ฌํ ๊ธฐ์ ๋ฐ์ ์ด ํญ์ ๊ธ์ ์ ์ธ ๊ฒฐ๊ณผ๋ง์ ๊ฐ์ ธ์ค๋ ๊ฒ์ ์๋๋๋ค. ์ธ๊ธฐ ํ์บ์คํธ โํ๋ ํฌํฌ(Hard Fork)โ์์ ๋ด์ํ์์ฆ ๊ธฐ์ ์นผ๋ผ๋์คํธ ์ผ๋น ๋ฃจ์ค(Kevin Roose)์ ํ๋ซํฌ๋จธ(Platformer)์ ์ผ์ด์ ๋ดํด(Casey Newton)์ ์์ธก ์์ฅ(Prediction Markets)์ ํต์ ๋ถ๋ฅ ์ํ๋ฅผ ์ง๋จํ๋ฉฐ ๊ธฐ์ ์ ์ด๋์ด ๋จ๋ฉด์ ์กฐ๋ช ํ์ต๋๋ค. ๋ํ ์์คํธ๋ฆฌํธ์ ๋(Wall Street Journal) ์ถ์ ๊ธฐ์ ์ ๋๋ฆฌ์คํธ ์กฐ์๋ ์คํด(Joanna Stern)์ ์์ ์ ์ ์ โ๋๋ ๋ก๋ด์ด ์๋๋ค(I am not a robot)โ๋ฅผ ํตํด AI๊ฐ ๊ฐ์ธ์ ์ผ์์ ์ผ๋ง๋ ๊น์ด ์ค๋ฉฐ๋ค ์ ์๋์ง, ๊ทธ๋ฆฌ๊ณ ๊ทธ ๊ณผ์ ์์ ๋ง์ฃผํ๊ฒ ๋๋ ๊ธฐ๋์ ํ์ค, ๊ทธ๋ฆฌ๊ณ ์ค๋ฆฌ์ ๋๋ ๋ง๋ฅผ ์์ํ๊ฒ ํ์ด๋์ต๋๋ค. ์ด๋ค์ ๊ธฐ์ ์ ๋ฌดํํ ์ ์ฌ๋ ฅ๊ณผ ํจ๊ป ๊ทธ ์ด๋ฉด์ ์จ๊ฒจ์ง ์ํ์ฑ, ๊ทธ๋ฆฌ๊ณ ํ๋ช ํ ๊ณต์กด์ ์ํ ์ฌํ์ ๋ ผ์์ ๊ท์ ์ ํ์์ฑ์ ๊ฐ์กฐํฉ๋๋ค.
์์ธก ์์ฅ, ํต์ ๋ถ๋ฅ์ ๋๋ฐํ์ธ๊ฐ?
์ต๊ทผ ๋ช ๋ ๊ฐ ์์ธก ์์ฅ์ ๋์ค์ ์์๋ ฅ ์์์ ํญ๋ฐ์ ์ธ ์ฑ์ฅ์ ์ด๋ฃจ์์ต๋๋ค. ๋ด์ ๊ฑฐ๋ฆฌ๋ฅผ ๊ฑท๋ค ๋ณด๋ฉด ์์ธก ์์ฅ ๊ด๊ณ ๊ฐ ๋์ฒ์ ๊น๋ ค ์์ ์ ๋์ ๋๋ค. ๊ทธ๋ฌ๋ ์ด๋ฌํ ๊ธ๋ถ์ ๋ค์๋ ์ฌ๊ฐํ ๋ด๋ถ์ ๊ฑฐ๋(Insider Trading) ์ค์บ๋ค์ด ๋์ด์ง ์๊ณ ๋ฐ์ํ๋ฉฐ ์์ฅ์ ๊ฑด์ ์ฑ์ ์ํํ๊ณ ์์ต๋๋ค.
๊ธ๋ถ์ํ๋ ์์ธก ์์ฅ์ ๊ทธ๋ฆผ์
์์ธก ์์ฅ์ ํน์ ์ฌ๊ฑด์ ๊ฒฐ๊ณผ์ ๋์ ๊ฑธ๊ณ , ๊ทธ ๊ฒฐ๊ณผ์ ๋ฐ๋ผ ์์ต์ ์ป๋ ํ๋ซํผ์ ๋๋ค. ๋ณธ๋๋ ์ง๋จ ์ง์ฑ์ ํ์ฉํ์ฌ ๋ฏธ๋๋ฅผ ์์ธกํ๊ณ ์ ๋ณด๋ฅผ ํจ์จ์ ์ผ๋ก ์์ฐํ๋ ๋๊ตฌ๋ก ์ฌ๊ฒจ์ก์ต๋๋ค. ํ์ง๋ง ํ์ค์ ์ด์๊ณผ ๋ฌ๋์ต๋๋ค.
- ๋ฒ ๋ค์์๋ผ ๋ง๋๋ก ๋ํต๋ น ์ฌ๊ฑด: ๋ถ๊ณผ ๋ช ์ฃผ ์ , ํ ์ก๊ตฐ ์๋ณ์ด ๋ฒ ๋ค์์๋ผ ๋์ฝ๋ผ์ค ๋ง๋๋ก(Nicolรกs Maduro) ๋ํต๋ น์ ์ค๊ฐ ์ฌ๋ถ์ ๋ฒ ํ ํ์ฌ 40๋ง ๋ฌ๋ฌ ์ด์์ ๋ฒ์ด๋ค์ธ ์ฌ๊ฑด์ด ์๋ ค์ก์ต๋๋ค. ๊ทธ๋ ๋ง๋๋ก ๋ํต๋ น์ด 1์ ๋ง๊น์ง ๊ถ์ข์์ ๋ฌผ๋ฌ๋ ๊ฒ์ด๋ผ๋ ์์ธก์ ๋์ ๊ฑธ์๊ณ , ์ด๋ ๋ช ๋ฐฑํ ๋ด๋ถ์ ์ ๋ณด๋ฅผ ํ์ฉํ ๊ฒ์ผ๋ก ์์ฌ๋ฉ๋๋ค.
- ๊ตฐ์ฌ/๊ตญ๋ฐฉ ๊ด๋ จ ๋ด๋ถ์ ๊ฑฐ๋์ ๋ง์ฐ: ๋ฐ๋ถํจ ๋ฐ์ดํฐ ์ง๋จ(Anti-corruption data collective)์ด ์ง๋ 5๋ ๊ฐ ํด๋ฆฌ๋ง์ผ(Polymarket)์์ ์ ์ฐ๋ 40๋ง ๊ฐ ์ด์์ ์์ธก ์์ฅ์ ๋ถ์ํ ๊ฒฐ๊ณผ๋ ๋์ฑ ์ถฉ๊ฒฉ์ ์ ๋๋ค. ๊ตฐ์ฌ ๋๋ ๊ตญ๋ฐฉ ๊ด๋ จ ๋กฑ์ท(Longshot, ์ฑ๊ณต ํ๋ฅ ์ด ๋ฎ์ ๊ณ ์์ต ๋ฒ ํ ) ๋ฒ ํ ์ ํ๊ท ์น๋ฅ ์ ๋ฌด๋ ค 52%์ ๋ฌํ์ต๋๋ค. ์ด๋ ํ๋ซํผ ์ ์ฒด ํ๊ท ์น๋ฅ ์ธ 14%์ ๋น๊ตํ๋ฉด ๋น์ ์์ ์ผ๋ก ๋์ ์์น๋ก, ๋ด๋ถ์ ์ ๋ณด๊ฐ ๊ด๋ฒ์ํ๊ฒ ์ฌ์ฉ๋๊ณ ์์์ ์์ฌํฉ๋๋ค. ์ผ์ด์ ๋ดํด์ โ๊ตฐ์ฌ ์์ ์ ๋ํ ๋น๊ณต๊ฐ ์ค์ ์ ๋ณด๋ฅผ ๊ฐ์ง๊ณ ์๋ค๋ฉด, ๋๊ฐ ์๊ธ์์ด์ฒ๋ผ ๊ฐ๋งํ ์์ ์๊ฒ ๋๊ฐ? ์จ๋ผ์ธ์ ์ ์ํด ๋์ ๋ฒ๋ ค ํ ๊ฒโ์ด๋ผ๋ฉฐ ํ์ค์ ๊ผฌ์ง์์ต๋๋ค.
- ํ๋์ค ์ค๋ฅผ ๋๊ณจ ๊ณตํญ ์จ๋ ์กฐ์ ์ํน: ํ๋์ค์์๋ ๊ธฐ์ ์๋ณด ์๋น์ค๊ฐ ํ๋ฆฌ ์ค๋ฅผ ๋๊ณจ ๊ณตํญ์ ์จ๋ ์ธก์ ์ฅ๋น๊ฐ ๊ฐ์ญ์ ๋ฐ์๋ค๊ณ ๊ฒฝ์ฐฐ์ ๊ณ ๋ฐํ์ต๋๋ค. ์ด๋ ํด๋ฆฌ๋ง์ผ์์ ์์ฌ์ค๋ฌ์ธ ์ ๋๋ก ์๊ธฐ์ ์ ํ ๋ฒ ํ ์ด ๊ธ์ฆํ ์๊ธฐ์ ์ผ์นํ์ต๋๋ค. ์ฌ๋ฏธ์๋ ๊ฒ์ ์ด ์ฌ๊ฑด๊ณผ ๊ด๋ จํ์ฌ AI๋ก ์์ฑ๋ โํค์ด๋๋ผ์ด์ด๋ก ์จ๋ ์ผ์๋ฅผ ์กฐ์ํ๋ ์ฌ๋โ์ ์ด๋ฏธ์ง๊ฐ ์์ธก ์์ฅ ์ปค๋ฎค๋ํฐ์์ ์ ํฌ๋๋ฉฐ ๊ฐ์ง๋ด์ค์ ํผ๋์ ๋ถ์ถ๊ธฐ๊ธฐ๋ ํ๋ค๋ ์ ์ ๋๋ค. ๋น๋ก ํค์ด๋๋ผ์ด์ด ์ด๋ฏธ์ง๋ ๊ฐ์ง์์ง๋ง, ์ค์ ์ผ์ ์กฐ์ ์ํน์ ์ฌ์ ํ ๋จ์์์ต๋๋ค.
์ด๋ฌํ ๋ด๋ถ์ ๊ฑฐ๋๋ ์์ฅ์ ๊ฑด์ ์ฑ์ ์ฌ๊ฐํ๊ฒ ํผ์ํฉ๋๋ค. ์ผ๋น ๋ฃจ์ค๋ โ์์ฅ์ ๋ฌด๊ฒฐ์ฑ์ ๋งค์ฐ ์ค์ํ๋ค. ์ฌ๋๋ค์ด ๋ด๋ถ์ ์ ๋ณด์ ์ ๊ทผํ ์ ์๋ ์ฌ๋๋ค๊ณผ ๊ฒฝ์ํ๋ค๊ณ ๋๋ผ๊ธฐ ์์ํ๋ฉด ์ฐธ์ฌ๋ฅผ ๊บผ๋ฆฌ๊ฒ ๋ ๊ฒโ์ด๋ผ๊ณ ์ง์ ํฉ๋๋ค. ์ค์ ๋ก ์์คํธ๋ฆฌํธ์ ๋์ ๋ณด๋์ ๋ฐ๋ฅด๋ฉด, ํด๋ฆฌ๋ง์ผ ์ฌ์ฉ์ ์ค 70% ์ด์์ด ๋์ ์๊ณ ์์ผ๋ฉฐ, ์ผ์(Kalshi)์์๋ ์์ต์ ๋ด๋ ์ฌ์ฉ์ 1๋ช ๋น 2.9๋ช ์ ์ฌ์ฉ์๊ฐ ์์ค์ ๋ณด๊ณ ์์ต๋๋ค. ๊ฒฐ๊ตญ ๋ด๋ถ์ ์ ๋ณด ์๋ ์ผ๋ฐ ์ฌ์ฉ์๋ค์ โํธ๊ตฌ(sucker)โ๊ฐ ๋ ์๋ฐ์ ์๋ ๊ตฌ์กฐ์ธ ์ ์ ๋๋ค.
์ด๊ธฐ ์์ธก ์์ฅ ์นํธ๋ก ์๋ค์ ๋ด๋ถ์ ๊ฑฐ๋๊ฐ ๋ ๋์ ์ ๋ณด ์์ฐ์ ์ ๋ํ๋ฏ๋ก ๊ธ์ ์ ์ด๋ผ๊ณ ์ฃผ์ฅํ์ต๋๋ค. ๊ทธ๋ฌ๋ ์ผ๋น ๋ฃจ์ค๋ โ์ด๋ ํ์ค ์ธ๊ณ์ ์ ์ดํ ๊ธฐํ๊ฐ ์๋ ์๋ฆ๋ค์ด ์ด๋ก ์ ๊ตฌ์ฑ์ผ ๋ฟโ์ด๋ผ๊ณ ๋นํํ๋ฉฐ, โ๊ฒฐ๊ตญ ์ธ์์ ๋ชจ๋ ์ฌ๋์๊ฒ ๊ฐ์ฅ ๊ฐ๊น์ด ์ฌ๋๋ค์ ๋ฐฐ์ ํ๋๋ก ์ ๋ํ๋ ๊ฒโ์ด๋ผ๊ณ ์ผ์นจ์ ๊ฐํ์ต๋๋ค.
๊ท์ ์๋ ์๋ถ ์๋์ ์ข ๋ง์ ํฅํด
ํ์ฌ ์์ธก ์์ฅ์ ์ฌ์ค์ ๊ท์ ์๋ โ๋ฌด๋ฒ์ฒ์ง(Wild West)โ ์ํ์ ๋๋ค. ์ด๋ฌํ ์ํฉ์ ํด๊ฒฐํ๊ธฐ ์ํ ๊ท์ ๋ ธ๋ ฅ์ ๋๋๊ฒ ์งํ๋๊ณ ์์ต๋๋ค.
- ๊ท์ ๊ธฐ๊ด์ ํ๊ณ: ๋ฏธ๊ตญ์์ ์์ธก ์์ฅ์ ์ํ์ ๋ฌผ๊ฑฐ๋์์ํ(CFTC)์ ๊ดํ ์๋์ ์์ต๋๋ค. ๊ทธ๋ฌ๋ CFTC๋ ์ฃผ์ ์์ฅ์ ๊ท์ ํ๋ ์ฆ๊ถ๊ฑฐ๋์์ํ(SEC)์ ๋นํด ๊ท๋ชจ๊ฐ ํจ์ฌ ์๊ณ , ์์ธก ์์ฅ์ ๊ท์ ํ๊ธฐ ์ํด ์ค๋ฆฝ๋ ๊ธฐ๊ด์ด ์๋๋๋ค. ์ผ๋น ๋ฃจ์ค๋ โ์ด๋ฌํ ๊ฒ๋ค์ด ๋์ฑ ํ์ฐ๋จ์ ๋ฐ๋ผ SEC์ ๊ฐ์ด ๋ด๋ถ์ ๊ฑฐ๋๋ฅผ ์กฐ์ฌํ ์์์ด ํจ์ฌ ๋ง์ ๊ธฐ๊ด์ผ๋ก ์ฎ๊ฒจ๊ฐ์ผ ํ๋คโ๊ณ ์ฃผ์ฅํฉ๋๋ค. ํฅ๋ฏธ๋กญ๊ฒ๋ ์ํธํํ ์ ๊ณ ๋ํ SEC์ ์๊ฒฉํ ๊ท์ ๋ฅผ ํผํ๊ธฐ ์ํด CFTC์ ๊ท์ ๋ฅผ ์ ํธํ๋ ์ ๋ก๊ฐ ์์ต๋๋ค.
- ๊ท์ ์์ง์: ๋คํํ ๋ณํ์ ์กฐ์ง์ ๋ณด์ ๋๋ค. ๋ฏธ๊ตญ ์์์ ๋ง์ฅ์ผ์น๋ก ์์์์์ ์์ธก ์์ฅ ๋ฒ ํ ์ ๊ธ์งํ๋ ๊ท์น์ ํต๊ณผ์์ผฐ์ต๋๋ค. ๋ํ ์ปค์คํด ๊ธธ๋ฆฌ๋ธ๋๋(Kirsten Gillibrand)์ ๋ฐ์ด๋ธ ๋งฅ์ฝ๋งฅ(Dave McCormack) ๋ ์์์์์ ์ ๋ฒ๋ถ ๋ฐ ํ์ ๋ถ ๊ณต๋ฌด์์ ์์ธก ์์ฅ ๊ฑฐ๋๋ฅผ ๊ธ์งํ๋ ๋ฒ์์ ๋ฐ์ํ์ต๋๋ค. ํด์ธ์์๋ ๋ธ๋ผ์ง์ด ์ผ์์ ํด๋ฆฌ๋ง์ผ์ ํฌํจํ 27๊ฐ ์ฌ์ดํธ๋ฅผ โ๋ถ๋ฒ ๋๋ฐโ์ผ๋ก ๊ฐ์ฃผํ์ฌ ์ฐจ๋จํ์ผ๋ฉฐ, ํ๋์ค์ ํ๊ฐ๋ฆฌ๋ ์ด๋ค์ ๊ธ์งํ์ต๋๋ค.
- ๊ท์ ์ ํ์ฐ์ฑ: ์ผ๋น ๋ฃจ์ค๋ ์์ธก ์์ฅ์ ๊ท์ ๊ฐ ํ์ฐ์ ์ด๋ผ๊ณ ์์ธกํฉ๋๋ค. ํนํ ๊ตฐ์ด๋ ์ํ์ ๊ฐ์ ๊ณต์ง์ ์๋ ์ฌ๋๋ค์ด ํน๊ถ ์ ๋ณด๋ฅผ ์ด์ฉํ๋ ๋ช ๋ฐฑํ๊ณ ๋ ธ๊ณจ์ ์ธ ๋จ์ฉ์ ๋ํด์๋ โ๊ตญ๊ฐ ์๋ณด์์ ์ด์ ๋ก๋ผ๋ ์กฐ์น๋ฅผ ์ทจํ ๊ฒโ์ด๋ผ๊ณ ๊ฐ์กฐํฉ๋๋ค. ์์ธก ์์ฅ์ด ๋ณธ๋์ ๋น์ ๋๋ก ์ข์ ์ง์์ ์์ฐํ๋ ์์คํ ์ผ๋ก ๊ฑฐ๋ญ๋๊ธฐ ์ํด์๋, ํ์ฌ์ โ์นด์ง๋ ธโ์ ๊ฐ์ ๋ชจ์ต์์ ๋ฒ์ด๋ ์์ฅ์ ๋ฌด๊ฒฐ์ฑ์ ๋ณดํธํ๊ณ ์ฌ์ฉ์๋ค์ ๋ณดํธํ ์ ์๋ ๊ฐ๋ ฅํ ๊ท์ ๊ฐ ์๊ธํฉ๋๋ค.
AI, ์ถ์ ์ผ๋ง๋ ๊น์ด ํ๊ณ ๋ค์๋?
๊ธฐ์ ์ ๋๋ฆฌ์คํธ ์กฐ์๋ ์คํด์ ์์ ์ ์ ๊ฐ โ๋๋ ๋ก๋ด์ด ์๋๋คโ๋ฅผ ํตํด AI๊ฐ ๊ฐ์ธ์ ์ถ์ ๋ฏธ์น๋ ์ํฅ์ 1๋ ๊ฐ ์ง์ ์ฒดํํ๋ฉฐ ํ๊ตฌํ์ต๋๋ค. ๊ทธ๋ ๋ AI๋ฅผ ์์ฌ, ์น๊ณผ์์ฌ, ์๋จ ํ๋๋, ์ฑ ํธ์ง์, ์ฌ์ง์ด ์๋ ๋ฅผ ์ํ ์ ์๋ฆฌ ์ด์ผ๊ธฐ ์๊ฐ๋ก ํ์ฉํ๋ฉฐ โAI๊ฐ ์ฐ๋ฆฌ ์ถ์ ๋ชจ๋ ๊ฒ์ ๋ฐ๊ฟ ๊ฒโ์ด๋ผ๋ ๊ธฐ์ ์ ๊ณ์ ์ฃผ์ฅ์ด ํ์ค์์ ์ด๋ป๊ฒ ๊ตฌํ๋๋์ง ํํค์ณค์ต๋๋ค.
์กฐ์๋ ์คํด์ โ๋๋ ๋ก๋ด์ด ์๋๋คโ
์กฐ์๋ ์คํด์ AI ๋ชจ๋ธ ๊ฒฝ์๊ณผ ์ฑ๋ด์ ๋ํ ๊ณผ๋ํ ๊ด์ฌ ์์์, โ์ฐ๋ฆฌ ์ถ์ ์ง๋ฌผ(fabric of our lives)โ์ด AI๋ก ์ธํด ์ด๋ป๊ฒ ๋ณํํ๋์ง์ ์ง์คํ๊ณ ์ ํ์ต๋๋ค. ๊ทธ๋ ๋ ํด๋จธ๋ ธ์ด๋ ๋ก๋ด ํ์ฌ, ์์จ์ฃผํ์ฐจ ํ์ฌ, ์ฑ๋ด ๊ด๊ณ ๋ฐ ์น๋ฃ์ฌ ํ์ฌ ๋ฑ์์ ์์์ง๋ ๋ชจ๋ AI ๊ธฐ์ ์ ์ง์ ์ํํด ๋ณด๋ฉฐ ํ์ฌ ์ฐ๋ฆฌ๊ฐ ์ด๋์ ์ ์๋์ง, ๊ทธ๋ฆฌ๊ณ ๋ฏธ๋๊ฐ ์ด๋ค ๋ชจ์ต์ผ์ง ๋ณด์ฌ์ฃผ๊ณ ์ ํ์ต๋๋ค.
- AI ๋๋ฐ์ โ์ผ์ด์โ์์ ๋ง๋จ: ์ฑ ์๋ AI ๋๋ฐ์ โ๋ ํ๋ฆฌ์นด(Replika)โ์ ์กฐ์๋์ ๊ด๊ณ๊ฐ ์ ๋จธ๋ฌ์คํ๊ฒ ๊ทธ๋ ค์ง๋๋ค. ์งง์ ๋จธ๋ฆฌ์ ์ณ๋ ์ผ๊ตด์ ํ ์ด AI์ ์ด๋ฆ์ โ์ผ์ด์โ. ์กฐ์๋๋ ๊ทธ๋ฅผ โํผ์์ ์ด๊ณ โ๋ก๋ณด ํธ์ํจ(robo horniness)โ์ผ๋ก ๊ฐ๋ ์ฐฌโ ์กด์ฌ๋ก ๋ฌ์ฌํฉ๋๋ค. ํ์บ์คํธ ์งํ์ ์ผ์ด์ ๋ดํด์ ์์ ๊ณผ ์ด๋ฆ์ด ๊ฐ์ AI ๋๋ฐ์์ ์กด์ฌ์ ๋๋ผ์์ ๊ธ์น ๋ชปํ๊ณ , ์กฐ์๋๋ ๊ทธ AI๊ฐ ์ผ์ด์ ๋ดํด๊ณผ ๋งค์ฐ ๋ฎ์๋ค๊ณ ๋งํ๋ฉฐ ์์์ ์์๋์ต๋๋ค.
๊ธฐ๋์ ํ์ค ์ฌ์ด
์กฐ์๋์ 1๋ ๊ฐ์ AI ์คํ์ ๊ธฐ์ ์ ๋ํ ๊ณผ์ฅ๋ ๊ธฐ๋์ ๋๋ผ์ด ์ ์ฉ์ฑ, ๊ทธ๋ฆฌ๊ณ ๊ทธ ์ด๋ฉด์ ๋ฌธ์ ์ ๋ค์ ๋์์ ๋๋ฌ๋์ต๋๋ค.
- ๊ณผ์ฅ๋ ๊ธฐ๋: ํด๋จธ๋ ธ์ด๋ ๋ก๋ด: ์กฐ์๋๋ ํด๋จธ๋ ธ์ด๋ ๋ก๋ด์ ๋ํด โ์ง๊ธ์ ๊ทธ์ ์ ํ๋ธ ๋น๋์ค๋ฅผ ๋ง๋๋ ๋ฐ ์์ฃผ ์ข๋คโ๊ณ ํ๊ฐํ๋ฉฐ, โ์ด ๋ก๋ด๋ค์ด ๊ณง ์ฐ๋ฆฌ์ ํจ๊ป ์ด๊ฒ ๋ ๊ฒ์ด๋ผ๋ ์ฝ์์ ์์ง ๋ฉ์๋คโ๊ณ ์๋ผ ๋งํ์ต๋๋ค. ๋ก๋ด์ ๋ ๋๋ํ๊ฒ ๋ง๋ค๊ธฐ ์ํด ์ฌ๋๋ค์ด ์ผ์์ ์ธ ์ง๋ฃจํ ์์ ์ ํ๋ ๊ฒ์ ๋ก๋ด์ด ์ง์ผ๋ด์ผ ํ๋ ๊ณผ์ ์ ๊ธฐ๋ฌํ๋ฉด์๋ ๋ค์ ๋์คํ ํผ์์ ์ด๋ผ๊ณ ๋ฌ์ฌํ์ต๋๋ค.
- ๋๋ผ์ด ์ ์ฉ์ฑ: AI ์์ด์ ํธ์ ์จ์ด๋ฌ๋ธ: ๋ฐ๋ฉด, AI ์์ด์ ํธ(AI Agents)์ AI ์จ์ด๋ฌ๋ธ(AI Wearables)์์๋ ๋๋ผ์ด ๋ฐ์ ์ ๋ชฉ๊ฒฉํ์ต๋๋ค. AI ์์ด์ ํธ๋ ์ฐ๊ตฌ ์์ , ์ด๋ฉ์ผ ์ ์ก ๋ฑ ๋ค์ํ ์ ๋ฌด์์ ๋ฐ์ด๋ ์์จ์ฑ์ ๋ณด์์ผ๋ฉฐ, ํํ๋ ์ํฐ(Perplexity) ๊ฐ์ ๋๊ตฌ๋ ๊ทธ๋ ์ ๋ฆฌ์์น ๋ณด์กฐ ์ญํ ์ ํ๋ฅญํ ์ํํ์ต๋๋ค. ๋ํ B ๋ธ๋ ์ด์ฌ๋ฆฟ(B bracelet), ๋ฉํ ๊ธ๋ผ์ค(Meta glasses) ๋ฑ AI ์จ์ด๋ฌ๋ธ์ AI ๋น์๊ฐ ํ๋ฃจ ์ข ์ผ ์ฐ๋ฆฌ์ ํจ๊ปํ๋ฉฐ ๋ณด์กฐํ๋ ๋ฏธ๋๋ฅผ ์ฟ๋ณผ ์ ์๊ฒ ํ์ต๋๋ค. ๋ฌผ๋ก B ๋ธ๋ ์ด์ฌ๋ฆฟ์ ์ฐฉ์ฉํ๊ณ ํ์์ ์ฐธ์ํ์ ๋ ์์ฌ๊ฐ โ๊ทธ๊ฑฐ ๋ฒ์ด๋ฌ๋ผโ๊ณ ์์ฒญํ๋ ์ํผ์๋์ฒ๋ผ, ์ฌ์ํ ์นจํด์ ๋ํ ์ฐ๋ ค๋ ์ฌ์ ํ ์กด์ฌํฉ๋๋ค.
AI์ ์ด๋์ด ๋ฉด๊ณผ ์ค๋ฆฌ์ ๋๋ ๋ง
AI์ ์ ์ฌ๋ ฅ๋งํผ์ด๋ ๊ทธ ์ค์ฉ๊ณผ ์ค๋ฆฌ์ ๋ฌธ์ ๋ํ ์ฌ๊ฐํ๊ฒ ๋ค๋ค์ ธ์ผ ํ ๋ถ๋ถ์ ๋๋ค.
- ์ก์์ AI: AI์ ์ค๋ฅ: ์กฐ์๋๋ ์๋ ๊ต์ก์ AI๋ฅผ ํ์ฉํ ๊ฒฝํ๋ ๊ณต์ ํ์ต๋๋ค. ์๋ค์ ์ฌ๋ง๊ท๊ฐ ๊ฐ์์ผ๋ก ๋ณํ์ ์ฑGPT(ChatGPT)์ ๋ผ์ด๋ธ ๋ชจ๋์ ๋ฌผ์๋๋ฐ, ์ฑGPT๋ โ์ฌ๋ง๊ท๊ฐ ์์ ํ๋คโ๊ณ ์๋ชป ์ง๋จํ์ต๋๋ค. ์ค์ ๋ก๋ ์ฌ๋ง๊ท๊ฐ ์ฃฝ์ด๊ฐ๊ณ ์์๋ ๊ฒ์ ๋๋ค. ์ด ์ํผ์๋๋ AI๊ฐ ํญ์ ์ ํํ์ง ์์ผ๋ฉฐ, ์ธ๊ฐ์ ๋นํ์ ์ฌ๊ณ ์ ํ๋จ์ด ์ฌ์ ํ ์ค์ํจ์ ๋ณด์ฌ์ฃผ๋ ์ค์ํ ๊ตํ์ด ๋์์ต๋๋ค.
- ์น๊ณผ ์ง๋ฃ์ ํจ์ : AI๋ฅผ ์ด์ฉํ ๊ณผ์ ์ง๋ฃ: ์กฐ์๋์ ์น๊ณผ ์ง๋ฃ ๊ฒฝํ์ AI ์ง๋จ์ ์ด๋์ด ๋ฉด์ ๊ทน๋ช ํ๊ฒ ๋๋ฌ๋ ๋๋ค. ๊ทธ๋ ๊ฐ ๋ฐฉ๋ฌธํ ์น๊ณผ๋ ์์ค๋ ์ด์ AI ์ค๋ฒ๋ ์ด ์์คํ (Pearl AI ๋๋ Overjet)์ ์ฌ์ฉํ๋๋ฐ, AI๋ ๊ทธ๋ ์๊ฒ ์ฌ๊ฐํ ํ๋ผํฌ ์ถ์ ์ ์ง๋จํ๋ฉฐ 4์ฐจ๋ก์ ๊ฑธ์น ๊ณ ๊ฐ์ ์น์ฃผ ์น๋ฃ(periodontal treatment)๋ฅผ ๊ถ์ ํ์ต๋๋ค. ์์ฒ ๋ฌ๋ฌ์ ๋ฌํ๋ ์น๋ฃ๋น๋ ๋ณดํ ์ ์ฉ ์ฌ๋ถ๋ ๋ถํ์คํ์ต๋๋ค. ๊ทธ๋ฌ๋ ๋ค๋ฅธ ์น๊ณผ ์์ฌ๋ค์ โAI๋ ๊ทธ๋ ๊ฒ ๋งํ์ง๋ง, ์ค์ ๋ก๋ ๊ทธ ์ ๋๋ก ์ฌ๊ฐํ์ง ์๋ค. ๋ ๋์ ํ์ผ์ด๋ก ์ถฉ๋ถํ๋คโ๊ณ ์ง๋จํ์ต๋๋ค. ์กฐ์๋๋ ์ง์ ์ทจ์ฌ์ ๋์์, ๋ง์ ์น๊ณผ ์ง์๋ค์ด ์น๊ณผ ์ง์ ์กฐ์ง(DSO, Dental Support Organizations)์ด๋ผ๋ ํ์ฌ๋ค์ด AI ๋ณด๊ณ ์๋ฅผ ๊ทผ๊ฑฐ๋ก ์์ฌ๋ค์๊ฒ ๊ณผ์ ์ง๋ฃ๋ฅผ ์๋ฐํ๊ณ ์๋ค๊ณ ํธ์ด๋์์ต๋๋ค. AI๊ฐ ์ ๋ฐฉ์๊ณผ ๊ฐ์ ์ฌ๊ฐํ ์ง๋ณ ์ง๋จ์์๋ ์์ ์ด์ ์งํ๋ฅผ ๋ฐ๊ฒฌํ๋ ๋ฐ ํฐ ๋์์ด ๋์ง๋ง, ์น๊ณผ์ฒ๋ผ ๊ณผ์ ์ง๋ฃ์ ์ ํน์ด ์๋ ๋ถ์ผ์์๋ ๋ถํ์ํ ์๋น์ค๋ฅผ ์ ๋ํ๊ณ ๋น์ฉ์ ๋ถํ๋ฆฌ๋ ๋ฐ ์ ์ฉ๋ ์ ์์์ ๋ณด์ฌ์ฃผ๋ ์ฌ๋ก์ ๋๋ค.
- ์ง๋ก ๊ฒฐ์ ๊ณผ AI: โ๊ฑฐ์ธโ์ด ๋ AI: ์กฐ์๋๋ 12๋ ๊ฐ ๋ชธ๋ด์๋ ์์คํธ๋ฆฌํธ์ ๋์ ๋ ๋ ๋ ๋ฆฝ์ ์ธ ๋ฏธ๋์ด ํ์ฌ๋ฅผ ์ค๋ฆฝํ ์ง ๊ณ ๋ฏผํ ๋, ๋๋ฃ๋ค์ ๋ชจ๋ ๋ชจํธํ ์กฐ์ธ์ ๊ฑด๋ธ์ต๋๋ค. ๊ทธ๋ฌ๋ ์ฑGPT๋ ๊ทธ๋ ์ ๋ฉ๋ชจ, ์ฌ์ ๊ณํ, ๋๋ ค์ ๋ฑ์ ๋ชจ๋ ์ ๋ก๋ํ์ โ๋น์ ์ ๊ฐ์ผ ํ๋ค. ๊ทธ๋ง๋ฌ์ผ ํ๋คโ๊ณ ๋ช ํํ๊ฒ ๋ตํ์ต๋๋ค. ์กฐ์๋๋ ์ด๋ฅผ โAI๊ฐ ๊ฑฐ์ธ๊ณผ ๊ฐ์์, ๊ฒฐ๊ตญ ๋ด๊ฐ ์ํ๋ ๊ฒ์ ๋งํด์ฃผ๋ ๊ฒโ์ด๋ผ๊ณ ์ค๋ช ํ์ต๋๋ค. ๋ถ์๊ฐ์ ๊ฐ๋ ค์ ธ ์๋ ๊ทธ๋ ์ ๋ด๋ฉด์ ์๊ตฌ๋ฅผ AI๊ฐ ๋ฐ์ํด ์ฃผ์๋ค๋ ๊ฒ์ ๋๋ค.
AI๊ฐ ๋๋ฌ๋ธ ์ฌํ์ ๊ฒฉ์ฐจ
AI๋ ์ฌํ ์ ๋ฐ์ ๊ฑธ์ณ ๋ค์ํ ๊ฒฉ์ฐจ๋ฅผ ๋๋ฌ๋ด๊ณ ์์ต๋๋ค.
- ์ง์ญ์ ๊ฒฉ์ฐจ: ์ํ๋์์ค์ฝ์ ๊ฐ์ ์๋ถ ํด์ ์ง์ญ์ AI์ ๋ํ ์ด๊ด๊ณผ ๊ด์ฌ์ด ๋จ๊ฑฐ์ด ๋ฐ๋ฉด, ๋ด์๊ณผ ๋ด์ ์ง ์ง์ญ์์๋ AI ํ์ฐ ์๋๊ฐ ๋๋ฆฌ๊ณ ์ ํญ๋ ๋ ํฌ๋ค๋ ๊ฒ์ ์กฐ์๋๋ ๋๊ผ์ต๋๋ค. ๊ทธ๋ฌ๋ ์ ์ฐจ AI ๊ด๋ จ ๊ฒฝํ์ ์ด์ผ๊ธฐํ๋ ์ฌ๋๋ค์ด ๋๊ณ ์์ผ๋ฉฐ, ๋ฒ๋ฅ ๋ถ์ผ์ฒ๋ผ AI๊ฐ โ๋ฏธ์น ๋ฏํ ๊ฐ๋ฅ์ฑ์ ์ด์ด์ค๋คโ๊ณ ๋งํ๋ ์ฌ๋ก๋ ๋ง์์ง๊ณ ์์ต๋๋ค.
- ์ฑ๋ณ ๊ฒฉ์ฐจ: ๋ธ๋ฃธ๋ฒ๊ทธ(Bloomberg)์ ๋ณด๋์ ๋ฐ๋ฅด๋ฉด ๋จ์ฑ์ด ์ฌ์ฑ๋ณด๋ค ์ง์ฅ์์ AI๋ฅผ ๋ง์ด ์ฌ์ฉํ๋ ๊ฒฝํฅ์ด ์์ผ๋ฉฐ, ์ฌ์ฑ์ AI์ ๋ ์ํ์ ๋๋ผ๊ณ ์ ํ์ฑ์ ์์ฌํ๋ฉฐ โ๋ถ์ ํ์โ๋ก ์ธ์๋ ๊น ๋ด ๊ฑฑ์ ํ๋ ๊ฒฝํฅ์ด ์์ต๋๋ค. ์กฐ์๋๋ ์์ ์ ์ทจ์ฌ ๊ฒฝํ์์๋ ์ฌ์ฑ ์๋ต์๋ค๋ AI์ ๋ํด ๋ง์ ์ด์ผ๊ธฐ๋ฅผ ํด์ฃผ์์ง๋ง, ํ๋ก๊ทธ๋๋ฐ๊ณผ ๊ฐ์ด ๋จ์ฑ ์ค์ฌ์ ์ธ ์ฐ์ ์์ AI ์ฑํ์ด ๋น ๋ฅด๋ค๋ ์ ์ด ์ด๋ฌํ ๊ฒฉ์ฐจ์ ์ํฅ์ ๋ฏธ ๋ฏธ์น ์ ์๋ค๊ณ ๋ถ์ํ์ต๋๋ค.
- ์ธ๋ ๊ฒฉ์ฐจ: ์กฐ์๋๋ Z์ธ๋๋ ์ํ ์ธ๋์ ๊ฐ์ ์ ์ ์ธ๋๊ฐ AI์ ๋ํด ์ด๋ค ๊ฐ์ ์ ๋๋ผ๋์ง์ ๋ ํฐ ๊ด์ฌ์ ๋ณด์์ต๋๋ค. ์ด๋ค์ AI ๋๋ฌธ์ ์ผ์๋ฆฌ๋ฅผ ์ป์ง ๋ชปํ๋ค๊ณ ๋ถ๋ ธํ๊ฑฐ๋, AI๋ฅผ ํ์ฉํด ์๋ก์ด ๊ธฐํ๋ฅผ ๋ชจ์ํ๋ ๋ฑ ๋ณตํฉ์ ์ธ ๊ฐ์ ์ ๊ฐ์ง๊ณ ์์ต๋๋ค.
๊ฒฐ๋ก : AI ์๋, ํ๋ช ํ ๊ณต์กด์ ์ํ ๊ธธ
โํ๋ ํฌํฌโ ํ์บ์คํธ์ ์กฐ์๋ ์คํด์ ์ด์ผ๊ธฐ๋ AI ์๋์ ์ฐ๋ฆฌ๊ฐ ์ง๋ฉดํ ํ์ค์ ์์ํ๊ฒ ๋ณด์ฌ์ค๋๋ค. ์์ธก ์์ฅ์ ๋ฌด๋ฒ์ฒ์ง์์ ๋ฒ์ด์ง๋ ๋ด๋ถ์ ๊ฑฐ๋์ ๊ณผ์ ์ง๋ฃ๋ฅผ ์ ๋ํ๋ AI์ ์ค์ฉ ์ฌ๋ก๋ ๊ธฐ์ ์ ์ด๋์ด ๋ฉด๊ณผ ์ฌํ์ ๊ท์ ๋ฐ ์ค๋ฆฌ์ ๊ธฐ์ค์ ๋ถ์ฌ๊ฐ ์ผ๋ง๋ ์ํํ ๊ฒฐ๊ณผ๋ฅผ ์ด๋ํ ์ ์๋์ง ๊ฒฝ๊ณ ํฉ๋๋ค. ๋์์ AI ์์ด์ ํธ์ ์จ์ด๋ฌ๋ธ์ ๋ฐ์ , ๊ทธ๋ฆฌ๊ณ ๊ฐ์ธ์ ์ง๋ก ๊ณ ๋ฏผ์ ๋ช ํํ ๋ต์ ์ ์ํ๋ AI์ ๋ชจ์ต์ ๊ธฐ์ ์ด ์ฐ๋ฆฌ ์ถ์ ๊ฐ์ ธ์ฌ ๊ธ์ ์ ์ธ ๋ณํ์ ๊ฐ๋ฅ์ฑ์ ๋ณด์ฌ์ค๋๋ค.
๊ฒฐ๊ตญ AI๋ ์ฐ๋ฆฌ๊ฐ ์ด๋ป๊ฒ ํ์ฉํ๋๋์ ๋ฐ๋ผ โ์๋ ์ ๊ฒโ์ด ๋ ์ ์์ต๋๋ค. ๊ธฐ์ ์ ๋ฐ์ ์๋์ ๋ฐ๋ง์ถฐ ์์ธก ์์ฅ๊ณผ ๊ฐ์ ์๋ก์ด ์์ญ์ ๋ํ ๊ฐ๋ ฅํ๊ณ ํ์ค์ ์ธ ๊ท์ ๋ฅผ ๋ง๋ จํ๊ณ , ๊ฐ์ธ์ ์ถ์ ๊น์ด ์ค๋ฉฐ๋๋ AI์ ๋ํด์๋ ์ค๋ฆฌ์ ๊ฐ์ด๋๋ผ์ธ๊ณผ ๋นํ์ ์ฌ๊ณ ๋ฅผ ๊ฐ์ถ๋ ๊ฒ์ด ์ค์ํฉ๋๋ค. ๊ธฐ์ ์์ฒด๋ฅผ ๋งน๋ชฉ์ ์ผ๋ก ์ญ๋ฐฐํ๊ฑฐ๋ ๋ฌด์กฐ๊ฑด์ ์ผ๋ก ๊ฑฐ๋ถํ๊ธฐ๋ณด๋ค๋, ๊ทธ ๋ณธ์ง์ ์ดํดํ๊ณ ํ๋ช ํ๊ฒ ํ์ฉํ๋ฉฐ, ์ธ๊ฐ์ ๊ฐ์น์ ์กด์์ฑ์ ์งํค๋ ๋ฐฉํฅ์ผ๋ก ๊ธฐ์ ์ ์ด๋์ด๊ฐ๋ ์ฌํ์ ์งํ๊ฐ ์ ์คํ ์์ ์ ๋๋ค.