English
Based on โWhat the Agent Economy Looks Like From Inside Stripeโ from Every Watch the original video
The Invisible Hand of AI: Stripeโs View of the Agent Economy
The internet, a realm once exclusively shaped by human interaction, is undergoing a profound transformation. A new class of actors โ autonomous AI agents โ is emerging, poised to become the dominant force in online commerce. From within Stripe, the company famously tasked with increasing the internetโs GDP, Emily, Head of Data and AI, offers an unparalleled, macro-level perspective on this seismic shift. Her insights reveal not just a technological evolution, but a complete rethinking of economic infrastructure, payment models, and the very nature of online fraud.
The Dawn of an Autonomous Internet
For decades, the internet operated on a foundational assumption: the primary actor was a human sitting in front of a screen, browsing, filling forms, clicking, and ultimately, making decisions. This assumption is now breaking. โThe internet economy is becoming more autonomous,โ Emily explains. โSometimes the human is still totally in control, but theyโre interacting through an AI interface. Sometimes the agent is acting on their behalf. And then sometimes software now is just out interacting directly with other software.โ
This shift isnโt merely about AI improving existing services like search or coding; itโs about the introduction of a fundamentally new type of participant. These agents, over time, are expected to become the predominant actors on the internet, necessitating an evolution at every layer of the digital stack. For Stripe, this means not only becoming โagent readyโ itself but also empowering businesses to navigate this new landscape. The questions are profound: How are products discovered and bought? What should developer tools look like? And crucially for Stripe, what is the underlying economic infrastructure โ payments, billing, fraud detection, identity โ needed in a world where actors are no longer just humans?
The New Frontier of Fraud: Stealing Compute
One of the most immediate and striking impacts of the agent economy is the radical redefinition of fraud. Historically, payment fraud involved stealing money or card credentials. Today, โfraudsters are stealing compute,โ Emily reveals. This is a very different problem with existential implications for many AI businesses.
In traditional SaaS models, a free tier or trial cost the provider very little. Stealing access to a basic software seat wasnโt particularly valuable for fraudsters. However, in the AI world, every prompt, every image generated, every API request has a very real, often substantial, cost attached. Compute is expensive, and โfree compute is kind of the new CACโ (Customer Acquisition Cost) for many AI companies, serving as a major lever for growth through free trials and credits. This high-value, high-cost resource has opened up three primary fraud vectors:
- Multi-Account Abuse: Bad actors sign up repeatedly with new identities and email addresses to claim new user credits, staying ahead of detection. Stripe data shows that across AI companies using their platform, about 7% of sign-ups are these multi-account users โ a non-trivial share.
- Free Trial Abuse: This is often the most urgent issue due to its rapid impact on unit economics. Emily shares a startling anecdote: one large AI company saw only 4% of its free trials convert to paid, with each trial costing $25 in LLM spend. This meant they were effectively spending $625 per payer before a single dollar of revenue came in. The vast majority of these โtrial usersโ were abusers with no intent to pay. This vector has seen exponential growth, quadrupling in the last six months. For one large AI user on Stripe, an astonishing 250,000 fraudulent free trials are blocked per week.
- Non-Payment Abuse: Customers consume thousands or even tens of thousands of dollars in compute over a month, day, or even an hour, only to fail payment when billed. By the time payment fails, the loss has already occurred, leaving AI companies holding the bag.
This shift means fraud is no longer a transaction-level problem occurring solely at checkout. โFraud used to be a transaction thing. Now it is a customer thing. It is a full-funnel thing,โ Emily emphasizes. It starts at sign-up, extends through credit allocation and free trials, and continues with overage management. The cost of having compute stolen is, for many, an existential risk.
The AI Arms Race
The rise of AI-powered fraud has ignited an โarms raceโ between fraudsters and those building defenses. Fraudsters now leverage AI to their advantage, but so do detection systems. Stripe, processing 2% of global GDP and growing rapidly, possesses a unique advantage: vast amounts of data. โLuckily not only do we have AI on our side just like they do, but we also have data on our side,โ Emily notes.
Stripeโs fraud protection product, Radar, has evolved dramatically. Itโs moved beyond just card transactions to encompass all payment methods, including crypto, and has extended its capabilities to the Radar API, allowing businesses to screen transactions even if they arenโt processed directly on Stripe. The company views fraud mitigation as a โpublic good,โ investing disproportionately to protect the internet economy. While new, creative fraud vectors constantly emerge, AI agents on Stripeโs network can now identify anomalous patterns across processors, payment methods, and merchants, burning down threats quickly. Emily remains โoverall bullish, but certainly not complacent.โ
The Unprecedented Growth of AI Companies
Beyond fraud, Stripeโs vantage point reveals other striking trends in the AI economy. These new AI companies are growing at an unprecedented pace. Looking at the top 100 AI companies on Stripe, those that reach $30 million in Annual Recurring Revenue (ARR) do so in about 18 months โ three times faster than the top 100 SaaS companies from 2018. This rapid scaling is evident even at earlier stages, like reaching $1 million or $5 million ARR.
Initially, much of this AI growth represented โnet new spend being pumped into the economy.โ It wasnโt directly substituting traditional SaaS or headcount because AI was starting near zero, and organizations were slow to drop existing contracts. However, the dynamics are changing. Emily predicts that going forward, some AI spend will substitute for traditional SaaS, especially as existing SaaS companies reinvent themselves with AI-first solutions. Moreover, companies will inevitably start to view AI spend as a substitute for headcount opex. If an engineerโs true cost now includes $30,000 in LLM spend, budget decisions will adjust accordingly.
Interestingly, while the overall category of AI tools (e.g., AI dev tools, AI coding assistants) shows higher retention rates than traditional SaaS, individual provider retention can be slightly lower. This is intuitive: once a user adopts an AI coding assistant, they love it and wonโt stop using such a tool. However, they are highly likely to โiterate across providersโ as new models emerge or quality varies, reflecting a โhigh percentage of curious travelers hopping from one thing to the next within a category.โ
Evolving Monetization Models
The rapid iteration isnโt limited to product features; itโs fundamentally reshaping monetization. Traditional SaaS relied heavily on fixed monthly subscriptions and seat-based usage, which made sense when marginal costs were near zero and humans were the primary users. Now, with very real inference costs, usage-based billing has become paramount. Companies are metering tokens, API calls, workflows, and even outcomes, charging with high precision based on units that reflect both customer value and cost structure.
Hybrid models are prevalent: subscriptions with usage overages, prepaid credits that burn down, or real-time top-ups. Stripeโs โtoken billingโ solution directly addresses the volatility of underlying model costs. For โwrapperโ companies built on top of LLMs, if their pricing doesnโt keep pace with changing model costs, margins can quickly disappear. Token billing allows businesses to track and price to underlying token costs in real-time, with a set markup, ensuring profitability.
Looking ahead, Emily foresees a significant shift away from seat-based pricing, particularly in the enterprise context. While consumers might prefer fixed monthly fees, businesses will increasingly demand outcome-based pricing for vertical AI solutions. For example, a customer service AI platform might charge per resolved case, aligning revenue with the value delivered and holding the provider accountable for ROI. โI just have a hard time imagining, you know, a year from now most vertical providers are literally charging on tokens,โ Emily states, predicting a move towards complex objective functions that measure quality, complexity, and even customer satisfaction for resolved outcomes. The concept of charging per developer seat for a tool that makes developers 10x more productive will likely become โsilly.โ
Stripeโs Adaptation: Building for Agents
Stripeโs own product development reflects this profound shift. The traditional notion of โdeveloper experienceโ focused on making it easier for human engineers at a keyboard. While clear APIs, better docs, and less setup still matter, the โdeveloperโ persona has broadened to include non-technical founders, coding assistants, and agents provisioning infrastructure on a humanโs behalf.
A telling metric: LLM traffic to Stripe docs is up 10x year-over-year. This isnโt just a substitution for human use (which remains flat to climbing); it signifies that machines are becoming direct users of developer infrastructure, including Stripeโs. Humans still sanity-check, but agents are integrating.
To address the growing friction in provisioning modern software stacks, Stripe launched โStripe Projects.โ This initiative allows users and their agents to create and manage parts of their software stack directly from the command line, automating account creation, credential management, and setup across multiple services. The overwhelming interest from over a hundred companies immediately after launch underscores the real need for tools that ease the setup process, even as coding becomes easier with AI.
Understanding Agentic Commerce
The term โagentic commerceโ can be misleading, often conjuring images of extreme, fully autonomous systems. Emily prefers to view it as a spectrum, starting with AI simply removing friction from existing internet commerce. This includes AI assisting with research, comparisons, and form filling, where the human still makes the final decision. The next stage involves descriptive search, where users articulate complex needs (e.g., โI need a summer camp for my kids within this budget and driving radiusโ). Real delegation, where the system makes purchases based on constraints, marks the โminimum viable barโ for agentic commerce. The furthest, most futuristic end of the spectrum is โambient commerce,โ where the system anticipates needs without explicit prompts.
Regardless of where an interaction falls on this spectrum, the underlying economic infrastructure needs to change. The old model โ humans creating accounts, choosing plans, filling forms, and entering card details โ is no longer universally applicable. Agent-assisted buying, where discovery, checkout, and payment occur within AI interfaces like Gemini or ChatGPT, rather than on a merchantโs website, requires a redesign of payment flows. The AI agent needs to understand merchant products, prices, and checkout flows to act on behalf of the consumer, and new trust mechanisms are paramount.
The internet is no longer just a marketplace for humans; itโs rapidly evolving into an autonomous economic ecosystem where AI agents play an increasingly central role. Stripe, with its unique position at the heart of internet payments, is not just observing this transformation but actively building the foundational infrastructure that will enable the next wave of digital commerce. The agent economy is here, and its implications are reshaping everything from how businesses grow to how fraud is fought, and how we all interact with the digital world.
Based on โWhy Even Some Democrats Hate Californiaโs Billionaire Tax Proposalโ from New York Times Podcasts Watch the original video
Californiaโs Billionaire Blitz: Why a Progressive Tax Proposal Faces Unexpected Democratic Resistance
In the heart of California, a battle is brewing that pits the promise of economic equality against the fear of driving away the stateโs wealthiest residents. A landmark proposal to levy a one-time 5% tax on the net worth of Californiaโs billionaires has reached a critical milestone, garnering enough signatures to put the measure to a vote. Proponents champion it as a necessary step to โright the wrongs of an unequal system,โ arguing that those who have โprospered from here in California can afford to invest a little more in keeping California running.โ Yet, this seemingly populist move has ignited a โfirestormโ of opposition, not just from the expected ultra-wealthy and Republicans, but surprisingly, from a significant contingent of Democrats.
This contentious proposal, emerging from a state that boasts more billionaires than any other, is poised to become a national touchstone in the ongoing debate over economic inequality. As Laurel Rosenhall, a New York Times colleague, explains, Californiaโs journey to this moment, and the intricate web of alliances and opposition it has created, offers crucial insights into the broader fight over wealth distribution across the nation.
The Anatomy of an Unusual Tax Proposal
At its core, the initiative seeks to impose a 5% tax on the net worth of any billionaire who was a California resident as of January 1st of this year. The revenue generated, estimated to affect roughly 200 individuals (though some experts believe the number could be higher), would be channeled into a special state fund, with 90% specifically earmarked for healthcare services.
What makes this proposal particularly unusual, and a source of much of the controversy, are two key structural elements:
- Taxing Assets, Not Income: Unlike traditional taxation, which typically targets income, this measure would tax an individualโs accumulated assets. This includes a wide array of wealth, from stocks and investment portfolios to luxury items like jewelry, cars, and even valuable paintings. Notably, it specifically exempts residential property, a curious carve-out in a state grappling with a housing crisis.
- A Dedicated Healthcare Fund: Rather than funneling the tax revenue into the stateโs general fund, where elected officials have discretion over its allocation, 90% of the billionaire tax proceeds would be statutorily dedicated to healthcare. This specific allocation is intended to offset massive funding cuts to Medicaid and Obamacare subsidies made by the federal government, which supporters argue are essential to โkeep hospitals open, to keep emergency rooms operating, and to keep the healthcare system that Californians rely on operating.โ
The Genesis of the Initiative: A Maverick Unionโs Fight
The driving force behind this unprecedented tax proposal is the Service Employees International Union United Healthcare Workers West (SEIU-UHW). This union, representing a broad spectrum of hospital workers from janitors to healthcare technicians, is deeply concerned about the ramifications of federal Medicaid cuts. Their fear is that these cuts will lead to widespread closures of emergency rooms and hospitals, directly impacting their members and the communities they serve.
The unionโs leader, Dave Regan, is described as a โmaverick operatorโ with a long history of utilizing ballot initiatives as leverage in labor negotiations. He conceived of the billionaire tax, collaborating with respected economists to craft the proposal largely on his own. Despite starting without a massive coalition, SEIU-UHW began pouring millions into a signature-gathering campaign, training volunteers who became a common sight at farmers markets and grocery stores across California. The initiative also gained significant momentum with the public backing of Senator Bernie Sanders, who flew to Los Angeles to lead a well-attended rally in support of the measure.
The Billionairesโ Gambit: Exit Strategies and Mitigation Tactics
The prospect of a wealth tax sent shockwaves through Californiaโs ultra-wealthy circles. The January 1st residency deadline spurred a flurry of activity, with some high-profile billionaires making immediate moves to change their official residency. Among them were Google co-founders Sergey Brin, who relocated to the Nevada side of Lake Tahoe, and Larry Page, who left for Florida. Influential tech investors like David Sacks moved to Texas, and Peter Thiel to Miami. While wealthy individuals often threaten to leave in the face of tax increases, in this instance, a notable handful followed through, at least officially, despite often retaining multiple properties in the Golden State.
For those determined to remain in California, the focus shifted to mitigating potential tax liabilities. A revealing anecdote from a conference in Orange County highlighted the lengths to which wealth managers and tax advisors were going, offering sessions on โhow to help your clients reduce their net worth.โ Suggestions included physically relocating valuable assets โ โmove your Picasso out of your house in Beverly Hills to your house in Aspenโ โ or even reducing insurance policies on high-value items like jewelry, as the tax would be based on actual worth.
However, not all billionaires are uniformly opposed. Some acknowledge the stateโs role in their prosperity, expressing a willingness to contribute more. Yet, the broader opposition is rooted in deeper concerns than mere financial loss.
Beyond โDonโt Take My Moneyโ: The Core Arguments Against the Tax
Opponents of the tax are not simply arguing against paying more; they are โdrilling into the specifics of the way that itโs structured,โ raising fundamental objections:
- Structural Objections to Asset Taxation: Critics contend that the very concept of taxing assets, rather than income, is inherently flawed and โwrong.โ They argue that requiring every billionaire to declare the value of all their property and assets is an intrusive and unprecedented demand.
- Constitutional Concerns: The retroactive nature of the tax, applying to residents as of January 1st, is being challenged as potentially unconstitutional.
- Economic Devastation and Innovation Flight: Perhaps the most potent argument from Silicon Valley and business leaders is that the tax would โabsolutely devastate Californiaโs ability to continue dominating in the world of innovation.โ They warn that the state stands to lose far more than just the tax revenue from billionaires; it risks losing the entrepreneurial spirit and the broader economic ecosystem that these individuals foster.
- Loss of Income Tax Revenue: Opponents highlight that the ultra-wealthy already contribute significantly through income taxes, which fund a vast array of essential state services like schools, public safety, and infrastructure. Governor Gavin Newsom, a vocal opponent, expresses concern that a โshort-term bump in tax revenue from the billionaire taxโ could lead to a โlong-term loss from that ongoing loss of income taxes from people who leave.โ
The Progressive Paradox: Why Even Democrats Are Wary
The most surprising element of this debate is the significant opposition and hesitancy from within the Democratic Party and influential unions, groups typically aligned with progressive policies aimed at wealth redistribution.
- Governor Gavin Newsomโs Tightrope Walk: Newsom finds himself in a precarious position. As a potential 2028 presidential contender, he cannot afford to alienate billionaires, yet he must also avoid appearing anti-populist in an era of potent economic discontent. He attempts to โthread a needle,โ vocally opposing a state-specific wealth tax while simultaneously stating, โWe ought to have a conversation about a national wealth tax.โ He also emphasizes Californiaโs already progressive income tax structure, arguing the state already levies significant taxes on its wealthiest earners. This stance allows him to agree with the โmoral argument to tax the richโ while opposing the current iteration of the proposal.
- Union Reservations: Even some of the stateโs most influential unions, including the teachersโ union and the umbrella group for SEIU, have not publicly endorsed the measure. Privately, they express concerns about the โspecial boxโ nature of the fund, worrying that locking up money for a single purpose might divert funds from other critical programs that most tax dollars typically support.
- Democratic Candidatesโ Hesitancy: Many Democratic candidates running for governor have also voiced reservations, even while affirming the general principle that the wealthy should pay more. Their concerns, much like Newsomโs, often hinge on the โparticularities of how this bill came about and who specifically benefits from it,โ particularly the ring-fencing of funds for healthcare.
This stands in stark contrast to figures like Bernie Sanders, who advocates for a more straightforward approach: โDonโt let the perfect be the enemy of the good here. Taxing billionaires is goodโฆ overall, we support this concept.โ For Sanders, deeply invested in fighting oligarchy, the specific structural โparticulars are of less concern.โ
The Road Ahead: A โStrange Bedfellowsโ Coalition?
The future of Californiaโs billionaire tax is far from certain. The stateโs unique ballot initiative system often serves as leverage, allowing political actors to negotiate outcomes outside of a direct vote. There is a possibility of โbackroom negotiationsโ in the coming months, potentially involving Governor Newsom, to try and get the measure off the ballot. Sergey Brin, for instance, has funded committees supporting other ballot measures that could effectively negate the billionaire tax, which could be used as bargaining chips.
Alternatively, all proposed measures could proceed to the ballot, leading to an โexpensive campaign season where voters are subjected to conflicting ballot measures on these tax issues.โ This scenario opens the door for a โstrange bedfellows kind of coalitionโ to emerge, where billionaires could align with unexpected allies โ perhaps unions concerned about the special fund, or even fiscally conservative Democrats โ to defeat the tax.
Lessons for a Nation Grappling with Inequality
The California billionaire tax fight offers several profound lessons for other states and the nation as a whole:
- Popularity vs. Practicality: The idea of โsticking it to the rich is very popular,โ reflecting widespread anxiety about economic inequality. However, translating that popularity into policy is โstill really hard to pass and get across the finish line,โ largely due to complex structural particulars and the immense influence of the billionaire class.
- The Power of Wealth: The swift and coordinated response from billionaires, including actual relocations and sophisticated mitigation strategies, underscores the formidable power of wealth in shaping policy debates.
- The Need for Broad Consensus: For such a transformative policy to succeed, it requires โa really broad base of support to push it across the finish line,โ extending beyond initial public enthusiasm to a more robust and unified political coalition.
Ultimately, this measure represents an โexistential threatโ for many billionaires, as it โwould set a precedent of going after the wealth that people have earned and taxing their possessions in a way that really has never been done before.โ For proponents, itโs an opportunity to โshift the paradigmโ in how government taxes wealth and how the wealthy contribute to society, especially in an era where many have utilized various accounts for tax avoidance.
Californiaโs billionaire tax proposal is more than just a local tax debate; itโs a microcosm of a larger national reckoning. It is โforcing a lot of people into the uncomfortable territory of reckoning with the moment weโre in,โ pushing the โanxiety about the billionaire class and economic inequality right to the forefrontโ of public discourse. The outcome in the Golden State will undoubtedly reverberate far beyond its borders, offering a blueprint โ or a cautionary tale โ for how societies might attempt to address the widening chasm of wealth.
ํ๊ตญ์ด
โWhat the Agent Economy Looks Like From Inside Stripeโ โ Every ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
AI ์๋, ์ธํฐ๋ท ๊ฒฝ์ ์ ์๋ก์ด ์ฃผ์ญ โ์์ด์ ํธโ์ ๋ฑ์ฅ๊ณผ ์คํธ๋ผ์ดํ์ ๋์
์ธํฐ๋ท ๊ฒฝ์ ์ ํจ๋ฌ๋ค์ ์ ํ: ์ธ๊ฐ์์ ์์ด์ ํธ๋ก
์ธํฐ๋ท์ ์ง๊ธ๊ป ์ฐ๋ฆฌ๊ฐ ๊ฒฝํํ๋ ๊ฒ๊ณผ๋ ์ ํ ๋ค๋ฅธ ์๋ก์ด ์ข ๋ฅ์ ํ์์๋ฅผ ๋ง์ดํ๊ณ ์์ต๋๋ค. ์คํธ๋ผ์ดํ(Stripe)์ ๋ฐ์ดํฐ ๋ฐ AI ์ด๊ด ์๋ฐ๋ฆฌ ์๋(Emily Shiner)๋ โ์๊ฐ์ด ์ง๋จ์ ๋ฐ๋ผ ์ด๋ฌํ ํ์์, ์ฆ ์์ด์ ํธ(agents)๊ฐ ์ธํฐ๋ท์ ์ฃผ์ ํ์์๊ฐ ๋ ๊ฒโ์ด๋ผ๊ณ ๋จ์ธํฉ๋๋ค. ์ด๋ ๋จ์ํ ๊ฒ์์ด ๋ ๋์์ง๊ฑฐ๋ ์ฝ๋ฉ์ด ๋ ์ฌ์์ง๋ ์์ค์ ๋์ด, ์ธํฐ๋ท ๊ฒฝ์ ์ ๊ทผ๋ณธ์ ์ธ ๊ตฌ์กฐ๊ฐ ๋ณํํ๊ณ ์์์ ์๋ฏธํฉ๋๋ค.
๊ณผ๊ฑฐ ์ธํฐ๋ท ๊ฒฝ์ ๋ โ์ธ๊ฐโ์ด๋ผ๋ ๊ทน๋๋ก ๋จ์ํ ๊ฐ์ ์ ๊ธฐ๋ฐ์ผ๋ก ๊ตฌ์ถ๋์์ต๋๋ค. ์ฌ๋์ ์คํฌ๋ฆฐ ์์ ์์ ์น์ฌ์ดํธ๋ฅผ ํ์ํ๊ณ , ์์์ ์์ฑํ๋ฉฐ, ๊ฒฐ์ ๋ฒํผ์ ํด๋ฆญํ์ต๋๋ค. ํ์ง๋ง ์ด์ ์ด๋ฌํ ๊ฐ์ ์ ๋ฌด๋์ง๊ธฐ ์์ํ์ต๋๋ค. ์ธ๊ฐ์ด AI ์ธํฐํ์ด์ค๋ฅผ ํตํด ์ํธ์์ฉํ๊ฑฐ๋, ์์ด์ ํธ๊ฐ ์ธ๊ฐ์ ๋์ ํด ํ๋ํ๋ฉฐ, ์ฌ์ง์ด ์ํํธ์จ์ด๊ฐ ๋ค๋ฅธ ์ํํธ์จ์ด์ ์ง์ ์ํธ์์ฉํ๋ ์๋๊ฐ ๋๋ํ ๊ฒ์ ๋๋ค.
์๋ฐ๋ฆฌ ์๋๋ โ์ด๋ฌํ ๋ณํ๊ฐ ๋ชจ๋ ๊ณ์ธต์์ ๋ฐ์ํ๊ธฐ ์์ํ๋ฉด์, ๋ง์ ๊ฒ๋ค์ ์ฌ๊ณ ํด์ผ ํ๋คโ๊ณ ๋งํฉ๋๋ค. ์ ํ์ด ์ด๋ป๊ฒ ๋ฐ๊ฒฌ๋๊ณ ๊ตฌ๋งค๋๋์ง๋ถํฐ ๊ฐ๋ฐ ๋๊ตฌ์ ๋ชจ์ต, ๊ทธ๋ฆฌ๊ณ ๊ฒฐ์ , ์ฒญ๊ตฌ, ์ฌ๊ธฐ ํ์ง, ์ ์ ํ์ธ ๋ฑ ์คํธ๋ผ์ดํ๊ฐ ๋ด๋นํ๋ ๊ทผ๋ณธ์ ์ธ ๊ฒฝ์ ์ธํ๋ผ์ ์ด๋ฅด๊ธฐ๊น์ง ๋ชจ๋ ๊ฒ์ด ์งํํด์ผ ํ๋ค๋ ๊ฒ์ ๋๋ค. ์คํธ๋ผ์ดํ์๊ฒ๋ โ์์ด์ ํธ ๋ ๋(agent ready)โ๊ฐ ๋๋ ๊ฒ, ๊ทธ๋ฆฌ๊ณ ๋ค๋ฅธ ๊ธฐ์ ๋ค์ด ์์ด์ ํธ ๋ ๋๊ฐ ๋๋๋ก ๋๋ ๊ฒ์ด ํต์ฌ ๊ณผ์ ๊ฐ ๋์์ต๋๋ค.
AI๊ฐ ๋ถ๋ฌ์จ โ์ ์ข ์ฌ๊ธฐโ์ ๊ทธ๋ฆผ์: ์ปดํจํธ ์ ๋๋ ์๋ก์ด ์ํ
์์ด์ ํธ ๊ฒฝ์ ์ ๋ถ์์ ์๋ก์ด ๊ธฐํ์ ํจ๊ป ์ ์ ์๋ ์ํ๋ ๊ฐ์ ธ์์ต๋๋ค. ํนํ ์ฌ๊ธฐ(fraud) ๋ถ์ผ์์ ๊ทธ ๋ณํ๋ ๋์ฑ ๊ทน๋ช ํฉ๋๋ค. ์๋ฐ๋ฆฌ ์๋๋ โAI๋ ๋งค์ฐ ๋ค๋ฅธ ์ฌ๊ธฐ ๋ฌธ์ ๋ฅผ ์ผ๊ธฐํ๋คโ๊ณ ์ง์ ํ๋ฉฐ, ๊ณผ๊ฑฐ์ ์ฌ๊ธฐ๊ฐ โ๊ฒฐ์ ์ฌ๊ธฐโ ์ฆ ๋์ด๋ ์นด๋ ์ ๋ณด๋ฅผ ํ์น๋ ํ์์๋ค๋ฉด, ์ด์ ๋ โ์ปดํจํธ ์ ๋(compute theft)โ๊ฐ ์ฃผ์ ๋ฌธ์ ๊ฐ ๋๊ณ ์๋ค๊ณ ์ค๋ช ํฉ๋๋ค.
AI ์๋์ ์ปดํจํธ๋ ์๋ก์ด ๊ณ ๊ฐ ํ๋ณด ๋น์ฉ(CAC, Customer Acquisition Cost)๊ณผ ๊ฐ์ต๋๋ค. ๊ธฐ์กด SaaS ๋ชจ๋ธ์์๋ ๋ฌด๋ฃ ๊ณ์ธต(free tier)์ ์ ๊ณตํ๋ ๊ฒ์ด ํฐ ๋น์ฉ์ด ๋ค์ง ์์์ง๋ง, AI์์๋ ๋ชจ๋ ํ๋กฌํํธ, ์ด๋ฏธ์ง ์์ฑ, API ์์ฒญ์ ์ค์ ๋น์ฉ์ด ๋ฐ์ํฉ๋๋ค. ๋ฐ๋ผ์ ๋ฌด๋ฃ ํฌ๋ ๋ง, ํ๋ฆฌ๋ฏธ์ ์๋น์ค, ๋ฌด๋ฃ ์ฒดํ ๋ฑ์ ๋๋๋ง๋ ๊ฒ์ AI ๊ธฐ์ ์๊ฒ ์ค์กด์ ์ธ ์ํ์ด ๋ฉ๋๋ค.
์คํธ๋ผ์ดํ๋ ์ด๋ฌํ ์ปดํจํธ ์ ๋ ์ฌ๊ธฐ๋ฅผ ์ธ ๊ฐ์ง ์ฃผ์ ์ ํ์ผ๋ก ๋ถ๋ฅํฉ๋๋ค:
- ๋ค์ค ๊ณ์ ์ ์ฉ(Multi-account Abuse): ์ ์ฑ ํ์์๊ฐ ์ฌ๋ฌ ๊ฐ์ ๊ฐ์ง ์ ์์ ๋ง๋ค์ด ๋ฌด๋ฃ ํฌ๋ ๋ง์ ๋ฐ๋ณต์ ์ผ๋ก ํ๋ํ๋ ๊ฒฝ์ฐ์ ๋๋ค. ์คํธ๋ผ์ดํ์์ ์ด์๋๋ AI ๊ธฐ์ ๋ค์ ๊ฐ์ ์ค ์ฝ 7%๊ฐ ์ด๋ฌํ ๋ค์ค ๊ณ์ ์ฌ์ฉ์์๊ฒ์ ๋ฐ์ํฉ๋๋ค.
- ๋ฌด๋ฃ ์ฒดํ ์ ์ฉ(Free Trial Abuse): ๋ฌด๋ฃ ์ฒดํ ๊ธฐ๊ฐ ๋์ ์๋น์ค๋ฅผ ์ ์ฉํ๊ณ ๊ฒฐ์ ์์ฌ ์์ด ์ปดํจํธ๋ฅผ ํ์น๋ ๊ฒฝ์ฐ์ ๋๋ค. ํ ๋ํ AI ๊ธฐ์ ์ ๊ฒฝ์ฐ, ๋ฌด๋ฃ ์ฒดํ ์ฌ์ฉ์์ 4%๋ง์ด ์ ๋ฃ๋ก ์ ํ๋์๊ณ , ๊ฐ ๋ฌด๋ฃ ์ฒดํ์ 25๋ฌ๋ฌ์ LLM ์ฌ์ฉ ๋น์ฉ์ด ๋ฐ์ํ์ฌ ์ ๋ฃ ๊ณ ๊ฐ ํ ๋ช ์ ํ๋ณดํ๋ ๋ฐ 625๋ฌ๋ฌ๊ฐ ์์๋์์ต๋๋ค. ์ด๋ค ์ค ๋๋ค์๋ ์ค์ ์๋น์ค ์ฌ์ฉ์๊ฐ ์๋ ์ฌ๊ธฐ๊พผ์ด์์ต๋๋ค. ์ง๋ 6๊ฐ์ ๋์ ๋ฌด๋ฃ ์ฒดํ ์ ์ฉ์ 4๋ฐฐ ์ฆ๊ฐํ์ผ๋ฉฐ, ์คํธ๋ผ์ดํ์ ํ ๋ํ AI ๊ณ ๊ฐ์ฌ์ ๊ฒฝ์ฐ ๋งค์ฃผ 25๋ง ๊ฑด์ ์ฌ๊ธฐ์ฑ ๋ฌด๋ฃ ์ฒดํ์ ์ฐจ๋จํ๊ณ ์์ต๋๋ค.
- ๋ฏธ์ง๋ถ ์ ์ฉ(Non-payment Abuse): ์๋น์ค ๊ณผ๋ค ์ฌ์ฉ(overage) ํ ์ฒญ๊ตฌ์๋ฅผ ์ง๋ถํ์ง ์๋ ๊ฒฝ์ฐ์ ๋๋ค. ๊ณ ๊ฐ์ด ํ ๋ฌ ๋๋ ํ๋ฃจ, ์ฌ์ง์ด ํ ์๊ฐ ๋ง์ ์์ฒ, ์๋ง ๋ฌ๋ฌ ์๋น์ ์ปดํจํธ๋ฅผ ์ฌ์ฉํ๊ณ ๊ฒฐ์ ์ ์คํจํ๋ฉด, AI ๊ธฐ์ ์ ๋ง๋ํ ์์ค์ ๋ ์๊ฒ ๋ฉ๋๋ค.
์ด๋ฌํ ๋ณํ์ ๋ฐ๋ผ ์ฌ๊ธฐ ํ์ง๋ ๋จ์ํ ๊ฑฐ๋ ์์ ์ด ์๋, ๊ณ ๊ฐ ์์ ์ฃผ๊ธฐ ์ ์ฒด์ ๊ฑธ์น โํ ํผ๋(full-funnel)โ ๋ฌธ์ ๋ก ์งํํ์ต๋๋ค. ์คํธ๋ผ์ดํ์ ์ฌ๊ธฐ ๋ฐฉ์ง ์ ํ์ธ ๋ ์ด๋(Radar)๋ ๊ณผ๊ฑฐ ๊ฒฐ์ ์์ ์๋ง ์๋ํ์ง๋ง, ์ด์ ๋ ๊ฐ์ ์์ ๋ถํฐ ํตํฉ๋์ด ๊ณ ๊ฐ์ ์ ์ฉ๋๋ฅผ ํ๋จํ๊ณ , ๊ณผ๋ค ์ฌ์ฉ ์ ์๋น์ค ์ ํ ์ฌ๋ถ๋ฅผ ๊ฒฐ์ ํ๋ ๋ฐ๊น์ง ๊ด์ฌํฉ๋๋ค.
์ฌ๊ธฐ๊พผ๋ค ๋ํ AI๋ฅผ ํ์ฉํ๊ธฐ ์์ํ๋ฉด์, โAI ๋ฌด์ฅ ์ฌ๊ธฐ๊พผโ๊ณผ์ ๊ตฐ๋น ๊ฒฝ์์ด ์ฌํ๋๊ณ ์์ต๋๋ค. ํ์ง๋ง ์คํธ๋ผ์ดํ๋ ์ ์ธ๊ณ GDP์ 2%์ ๋ฌํ๋ ๋ฐฉ๋ํ ๊ฑฐ๋ ๋ฐ์ดํฐ๋ฅผ ๊ธฐ๋ฐ์ผ๋ก AI ํ์ง ๋ชจ๋ธ์ ๊ณ ๋ํํ๋ฉฐ ์ด ๊ฒฝ์์์ ์์๋๊ฐ๊ณ ์์ต๋๋ค. ์๋ฐ๋ฆฌ ์๋๋ โ์ฌ๊ธฐ ๋ฐฉ์ง ๋ฐ ์ํ๋ ๊ณต๊ณต์ฌ์ ๊ฐ๋คโ๋ฉฐ, ์ธํฐ๋ท ๊ฒฝ์ ์ ์ฑ์ฅ์ ์ํด ์ฌ๊ธฐ ๋ฐฉ์ง์ ๋ง๋ํ ํฌ์๋ฅผ ์๋ผ์ง ์๊ณ ์๋ค๊ณ ๊ฐ์กฐํฉ๋๋ค.
AI ๊ธฐ์ ์ ์ด๊ณ ์ ์ฑ์ฅ๊ณผ ์์ต ๋ชจ๋ธ์ ํ์
์คํธ๋ผ์ดํ๊ฐ ๋ชฉ๊ฒฉํ๋ ๋ ๋ค๋ฅธ ์ค์ํ ํธ๋ ๋๋ AI ๊ธฐ์ ๋ค์ ํญ๋ฐ์ ์ธ ์ฑ์ฅ ์๋์ ์์ต ๋ชจ๋ธ์ ๊ธ์ง์ ์ธ ๋ณํ์ ๋๋ค. ์๋ฐ๋ฆฌ ์๋๋ โAI ๊ธฐ์ ๋ค์ ์ฐ๋ฆฌ๊ฐ ์ง๊ธ๊น์ง ๋ณธ ์ด๋ค ์ด์ ์ฝํธํธ(cohort)๋ณด๋ค๋ ๋น ๋ฅธ ์๋๋ก ๋งค์ถ์ ์ฑ์ฅ์ํค๊ณ ์๋คโ๊ณ ๋งํฉ๋๋ค.
์คํธ๋ผ์ดํ ์์ 100๊ฐ AI ๊ธฐ์ ์ ๋ถ์ํ ๊ฒฐ๊ณผ, ์ฐ๊ฐ ๋ฐ๋ณต ๋งค์ถ(ARR, Annual Recurring Revenue) 3์ฒ๋ง ๋ฌ๋ฌ์ ๋๋ฌํ๋ ๋ฐ ํ๊ท 18๊ฐ์์ด ์์๋์์ต๋๋ค. ์ด๋ 2018๋ ์์ 100๊ฐ SaaS ๊ธฐ์ ๋ณด๋ค 3๋ฐฐ๋ ๋น ๋ฅธ ์๋์ ๋๋ค. 1๋ฐฑ๋ง ๋ฌ๋ฌ ๋๋ 5๋ฐฑ๋ง ๋ฌ๋ฌ ARR ๋๋ฌ ์๋ ๋ํ ํจ์ฌ ๋น ๋ฆ ๋๋ค.
์ด๋ฌํ ์ด๊ณ ์ ์ฑ์ฅ์ ๋ท๋ฐ์นจํ๋ ๊ฒ์ ์์ตํ ๋ชจ๋ธ์ ๋น ๋ฅธ ์คํ๊ณผ ์งํ์ ๋๋ค.
- ์ฌ์ฉ๋ ๊ธฐ๋ฐ ์ฒญ๊ตฌ(Usage-based Billing)์ ๋ถ์: ๊ธฐ์กด SaaS๊ฐ ์ข์ ์(seat-based)๋ ๊ณ ์ ์์ ์ก์ ๊ธฐ๋ฐํ๋ค๋ฉด, AI ๊ธฐ์ ๋ค์ ํ ํฐ(tokens), API ํธ์ถ, ์ํฌํ๋ก์ฐ(workflows), ์ฌ์ง์ด ๊ฒฐ๊ณผ๋ฌผ(outcomes) ๋ฑ ๊ณ ๊ฐ ๊ฐ์น์ ์ค์ ๋น์ฉ ๊ตฌ์กฐ๋ฅผ ๊ฐ์ฅ ์ ๋ฐ์ํ๋ ๋จ์๋ฅผ ๊ธฐ์ค์ผ๋ก ๊ณผ๊ธํฉ๋๋ค.
- ํ์ด๋ธ๋ฆฌ๋ ์์ต ๋ชจ๋ธ: ๊ตฌ๋ (subscription) ๋ชจ๋ธ์ด ์ฌ๋ผ์ง ๊ฒ์ ์๋์ง๋ง, ์ฌ์ฉ๋ ์ด๊ณผ์ ๋ฐ๋ฅธ ์ถ๊ฐ ์๊ธ, ์ ๋ถ ํฌ๋ ๋ง ์์ง ๋ฐฉ์, ์ค์๊ฐ ์ถฉ์ (real-time top-ups) ๋ฑ ๋ค์ํ ํ์ด๋ธ๋ฆฌ๋ ํํ๋ก ์งํํ๊ณ ์์ต๋๋ค.
- ํ ํฐ ๋น๋ง(Token Billing): ํ์ LLM ๋ชจ๋ธ ๋น์ฉ์ด ์ค์๊ฐ์ผ๋ก ๋ณ๋ํ๋ ์ํฉ์์, ๋ํผ(wrapper) ๋๋ ์์ง AI ์๋ฃจ์ ์ ๊ณต์ ์ฒด๊ฐ ๋ง์ง์ ๋ณดํธํ๊ธฐ ์ํด ๋์ ํ ๋ฐฉ์์ ๋๋ค. ๊ธฐ๋ณธ ํ ํฐ ๋น์ฉ์ ๋น์ฆ๋์ค์์ ์ค์ ํ ๋งํฌ์ ์ ๋ํด ์ค์๊ฐ์ผ๋ก ์ถ์ ํ๊ณ ๊ฐ๊ฒฉ์ ์ฑ ์ ํฉ๋๋ค.
๋ฏธ๋์ ๊ฐ๊ฒฉ ๋ชจ๋ธ: ์ข์ ์์์ ๊ฒฐ๊ณผ๋ฌผ๋ก
์๋ฐ๋ฆฌ ์๋๋ ๋ฏธ๋์ ๊ฐ๊ฒฉ ๋ชจ๋ธ์ ๋ํด ํฅ๋ฏธ๋ก์ด ํต์ฐฐ์ ์ ์ํฉ๋๋ค. ๋ชจ๋ธ ์ ๊ณต์ ์ฒด(์: OpenAI, Claude)๋ ๊ณ์ํด์ ํ ํฐ ์ฌ์ฉ๋์ ๊ธฐ์ค์ผ๋ก ๊ณผ๊ธํ๊ฒ ์ง๋ง, ์์ง AI ์๋ฃจ์ (vertical AI solutions)์ ๊ฒฝ์ฐ ๊ถ๊ทน์ ์ผ๋ก โ๊ฒฐ๊ณผ๋ฌผ(outcomes)โ์ ๊ธฐ์ค์ผ๋ก ๊ณผ๊ธํ๊ฒ ๋ ๊ฒ์ด๋ผ๋ ์์ธก์ ๋๋ค. ๊ณ ๊ฐ ์๋น์ค ํ๋ซํผ Finn์ด๋ Intercom์ด ํด๊ฒฐ๋ ํฐ์ผ ์์ ๋ฐ๋ผ ๊ณผ๊ธํ๋ ๊ฒ์ฒ๋ผ, AI ์๋ฃจ์ ์ ๊ณ ๊ฐ์๊ฒ ์ ๊ณตํ๋ ์ค์ ROI์ ๊ธฐ๋ฐํ์ฌ ๋น์ฉ์ ์ฒญ๊ตฌํ๊ฒ ๋ ๊ฒ์ด๋ผ๋ ์ค๋ช ์ ๋๋ค.
๋ ๋์๊ฐ, ์๋ฐ๋ฆฌ ์๋๋ ๊ธฐ์ ํ๊ฒฝ์์ โ์ข์ ๊ธฐ๋ฐ(seat-based)โ ๋ผ์ด์ ์ค๊ฐ ์ฌ๋ผ์ง ๊ฒ์ด๋ผ๋ ๊ณผ๊ฐํ ์์ธก์ ๋ด๋์ต๋๋ค. AI ์์ด์ ํธ๊ฐ ๊ฐ๋ฐ์์ ์์ฐ์ฑ์ 10๋ฐฐ ๋์ผ ์ ์๋ค๋ฉด, ์ ๊ฐ๋ฐ์ ์์ ์์ต์ ๊ณ ์ ํด์ผ ํ๋๊ฐ์ ๋ํ ์๋ฌธ์ด ์ ๊ธฐ๋๋ค๋ ๊ฒ์ ๋๋ค. ๋ฌผ๋ก ๊ฐ์ธ ์๋น์ ์์ฅ์์๋ ๊ณ ์ ์์ ์ก ๋ชจ๋ธ์ด ์ ํธ๋ ์ ์์ง๋ง, ๊ธฐ์ ์์ฅ์์๋ 6๊ฐ์ ์์ ์ข์ ๊ธฐ๋ฐ ๋ผ์ด์ ์ค๊ฐ ์ ๋ฐ์ผ๋ก ์ค์ด๋ค ์๋ ์๋ค๊ณ ์ ๋งํฉ๋๋ค.
์ฑ์ฅ ๋๋ ฅ์ ๋ณํ: ์ ๊ท ์ง์ถ์์ ๋์ฒด ์ง์ถ๋ก
AI ์ฑ์ฅ์ ์ด๊ธฐ ๋จ๊ณ์์๋ ๋๋ถ๋ถ โ์์ ์ ๊ท ์ง์ถ(net new spend)โ์ด ๊ฒฝ์ ์ ํฌ์ ๋์์ต๋๋ค. AI๊ฐ ๊ฑฐ์ ์ ๋ก์์ ์์ํ๊ธฐ ๋๋ฌธ์, ์ ํต์ ์ธ SaaS๋ ์ธ๊ฑด๋น(headcount opex)๋ฅผ ๋์ฒดํ๊ธฐ๋ณด๋ค๋ ์๋ก์ด ์์ญ์ ๋ํ ํฌ์๊ฐ ์ฃผ๋ฅผ ์ด๋ฃจ์์ต๋๋ค.
๊ทธ๋ฌ๋ ์์ผ๋ก๋ ๋ณํ๊ฐ ์์๋ฉ๋๋ค.
- ์ ํต SaaS ๋์ฒด: AI ํผ์คํธ ์ ๋ต์ ์ฑํํ SaaS ๊ธฐ์ ๋ค์ด ๊ธฐ์กด SaaS ์์ฅ์ ์ผ๋ถ๋ฅผ ์ ์ํ ๊ฒ์ ๋๋ค.
- ์ธ๊ฑด๋น ๋์ฒด: ๊ธฐ์ ๋ค์ด LLM์ ์๋นํ ์์ฐ์ ์ง์ถํ๊ธฐ ์์ํ๋ฉด์, ์ธ๊ฑด๋น ์์ฐ๊ณผ ROI ๊ฒฐ์ ์ ์ํฅ์ ๋ฏธ์น๊ฒ ๋ ๊ฒ์ ๋๋ค.
- AI ๊ธฐ์ ๊ฐ ๋์ฒด: AI ๋๊ตฌ ๋๋ฉ์ธ ์์ฒด์ ๋ํ ์ฌ์ฉ์ ์ ์ง์จ์ ๋์ง๋ง, ๊ฐ๋ณ AI ์ ๊ณต์ ์ฒด ๊ฐ์ ์ ํ์จ์ ๋ ๋ฎ์์ง ๊ฒ์ ๋๋ค. ์ฌ์ฉ์๋ ํญ์ ์๋ก์ด ๋ชจ๋ธ์ ์๋ํ๊ณ ๋ ๋์ ์๋ฃจ์ ์ ์ฐพ์ ์ด๋ํ๊ธฐ ๋๋ฌธ์ ๋๋ค.
์ ๋ฐ์ ์ผ๋ก AI ๊ธฐ์ ์ ๊ฐ์น๋ ์ฌ์ ํ ๊ธ์ ์ ์ผ๋ก ํ๊ฐ๋์ง๋ง, ์ผ๋ถ ํผ์กํ ์์ญ์์๋ ๊ฒฝ์์ด ์ฌํ๋ ๊ฒ์ผ๋ก ์์๋ฉ๋๋ค.
์์ด์ ํธ ์๋, ์คํธ๋ผ์ดํ์ ์งํ: ๊ฐ๋ฐ์ ๊ฒฝํ์์ ์์ด์ ํธ ๊ฒฝํ์ผ๋ก
์์ด์ ํธ ๊ฒฝ์ ๋ ์คํธ๋ผ์ดํ์ ์ ํ ์ฒ ํ์๋ ๊ทผ๋ณธ์ ์ธ ๋ณํ๋ฅผ ์๊ตฌํ๊ณ ์์ต๋๋ค. ์คํธ๋ผ์ดํ๋ ์ ํต์ ์ผ๋ก โ๊ฐ๋ฐ์ ๊ฒฝํ(Developer Experience, DX)โ์ ์ง์คํด ์์ต๋๋ค. ํ์ง๋ง ์ด์ ๊ฐ๋ฐ์์ ์ ์๋ ํ์ฅ๋์์ต๋๋ค. ์ฝ๋ฉ ๋์ฐ๋ฏธ, ๋น๊ธฐ์ ์ ์ฐฝ์ ์, ์ธ๊ฐ์ ๋์ ํด ์ธํ๋ผ๋ฅผ ํ๋ก๋น์ ๋ํ๋ ์์ด์ ํธ๊น์ง ํฌํจํ๋ ๊ฐ๋ ์ด ๋ ๊ฒ์ ๋๋ค.
์ด๋ฌํ ๋ณํ์ ์ฆ๊ฑฐ๋ ์คํธ๋ผ์ดํ ๋ด๋ถ ๋ฐ์ดํฐ์์๋ ๋ํ๋ฉ๋๋ค. ์คํธ๋ผ์ดํ ๋ฌธ์(Stripe Docs)์ ๋ํ LLM ํธ๋ํฝ์ ์ ๋ ๋๋น 10๋ฐฐ ์ฆ๊ฐํ์ต๋๋ค. ์ด๋ ๊ธฐ๊ณ ๋ํ ์คํธ๋ผ์ดํ์ ๊ฐ๋ฐ์ ์ธํ๋ผ๋ฅผ ํฌํจํ ๊ฐ๋ฐ์ ์ธํ๋ผ์ ์ฌ์ฉ์๊ฐ ๋๊ณ ์์์ ๋ณด์ฌ์ฃผ๋ ์ ์ฉํ ์ ํธ์ ๋๋ค.
์คํธ๋ผ์ดํ๋ ์ด๋ฌํ ๋ณํ์ ๋ง์ถฐ ๋ค์๊ณผ ๊ฐ์ ๋ ธ๋ ฅ์ ๊ธฐ์ธ์ด๊ณ ์์ต๋๋ค.
- Stripe Projects: ์ฝ๋ฉ์ ์ฌ์์ง์ง๋ง, ์ฌ์ ํ ์๋์ ์ธ โ์ค์ (setup)โ ๊ณผ์ ์ ์ด๋ ค์์ ํด๊ฒฐํ๊ธฐ ์ํ ํ๋ก์ ํธ์ ๋๋ค. ๊ฐ๋ฐ์๋ ์์ด์ ํธ๊ฐ ๋ช ๋ น์ค(command line)์์ ์ํํธ์จ์ด ์คํ์ ์ผ๋ถ๋ฅผ ์์ฑํ๊ณ ๊ด๋ฆฌํ ์ ์๋๋ก ํ์ฌ, ๊ณ์ ์์ฑ, ์๊ฒฉ ์ฆ๋ช ๊ด๋ฆฌ, ๋์๋ณด๋ ์ ํ ๋ฑ์ ๋ฒ๊ฑฐ๋ก์์ ์ค์ฌ์ค๋๋ค.
- ์์ด์ ํธ ์ปค๋จธ์ค(Agentic Commerce)์ ์คํํธ๋ผ: ์์ด์ ํธ ์ปค๋จธ์ค๋ ๋จ์ํ AI๊ฐ ๋ชจ๋ ๊ฒ์ ์์์ ๊ตฌ๋งคํ๋ ๊ทน๋จ์ ์ธ ์๋๋ฆฌ์ค๋ง์ ์๋ฏธํ์ง ์์ต๋๋ค. ์๋ฐ๋ฆฌ ์๋๋ ์ด๋ฅผ ์คํํธ๋ผ์ผ๋ก ์ค๋ช
ํฉ๋๋ค.
- ๋ง์ฐฐ ์ ๊ฑฐ(Friction Removal): AI๊ฐ ์ ๋ณด ๊ฒ์, ์ต์ ๋น๊ต, ์์ ์์ฑ ๋ฑ์ ๋์ ์ธ๊ฐ์ ๊ฒฐ์ ์ ์ฉ์ดํ๊ฒ ํฉ๋๋ค.
- ์์ ์ ๊ฒ์(Descriptive Search): ์ฌ์ฉ์๊ฐ ์์ฐ์ด๋ก ๋ณต์กํ ์๊ตฌ์ฌํญ์ ์ค๋ช ํ๋ฉด AI๊ฐ ์ด์ ๋ง๋ ๊ฒฐ๊ณผ๋ฅผ ์ฐพ์์ค๋๋ค.
- ์ค์ง์ ์์(Real Delegation): ์ฌ์ฉ์๊ฐ ์ ์ฝ ์กฐ๊ฑด์ ์ค์ ํ๋ฉด AI๊ฐ ์ด๋ฅผ ๊ธฐ๋ฐ์ผ๋ก ์ง์ ๊ตฌ๋งค๋ฅผ ์คํํฉ๋๋ค.
- ์ฐ๋น์ธํธ(Ambient): AI๊ฐ ์ฌ์ฉ์์ ํ์๋ฅผ ์์ธกํ๊ณ ๋ฅ๋์ ์ผ๋ก ๊ตฌ๋งค๋ฅผ ์ ์ํ๊ฑฐ๋ ์คํํฉ๋๋ค.
์ด๋ฌํ ์คํํธ๋ผ์ ์ด๋ ๋จ๊ณ์ ์๋ , ๊ทผ๋ณธ์ ์ธ ๊ฒฝ์ ์ธํ๋ผ์ ์ฌ์ค๊ณ๊ฐ ํ์ํฉ๋๋ค. ๊ณผ๊ฑฐ ์ธ๊ฐ์ด ๋ธ๋ผ์ฐ์ ์์ ๊ณ์ ์ ๋ง๋ค๊ณ , ํ๋์ ์ ํํ๊ณ , ์์์ ์ฑ์ฐ๊ณ , ์นด๋ ์ ๋ณด๋ฅผ ์ ๋ ฅํ๋ ๋ฐฉ์์ ๋ ์ด์ ์ ํจํ์ง ์์ต๋๋ค. AI ์์ด์ ํธ๊ฐ ์์ธ(merchant)์ ์ ํ, ๊ฐ๊ฒฉ, ๊ฒฐ์ ํ๋ฆ์ ์ดํดํ๊ณ ์๋น์๋ฅผ ๋์ ํด ํ๋ํด์ผ ํ๋ฉฐ, ์ด ๊ณผ์ ์์ โ์ ๋ขฐโ๋ ๋์ฑ ์ค์ํด์ง๋๋ค.
์ด๋ฅผ ์ํด ์คํธ๋ผ์ดํ๋ โ๊ณต์ ๊ฒฐ์ ํ ํฐ(shared payment token)โ์ด๋ผ๋ ์๋ก์ด ๊ธฐ๋ณธ ์์๋ฅผ ์ค๊ณํ์ต๋๋ค. ์ด๋ ์์ด์ ํธ๊ฐ ๊ตฌ๋งค์์ ์ ์ฉ ์ ๋ณด๋ฅผ ํ๋งค์์๊ฒ ์์ ํ๊ฒ ์ ๋ฌํ ์ ์๋๋ก ํ๋ฉฐ, ์ด ๊ณผ์ ์์ ๋ ์ด๋ ์ฌ๊ธฐ ์ ์(Radar fraud score)๋ ํจ๊ป ์ ๋ฌ๋์ด ํ๋งค์๊ฐ ์ ์ ํ ๋์ํ ์ ์๋๋ก ๋์ต๋๋ค.
๊ฒฐ๋ก : ์์จ์ ์ธ ์ธํฐ๋ท ๊ฒฝ์ ๋ฅผ ํฅํ ์คํธ๋ผ์ดํ์ ์ฌ์
์์ด์ ํธ ๊ฒฝ์ ์ ๋๋๋ ์ธํฐ๋ท์ ๊ทผ๊ฐ์ ํ๋๋ ๊ฑฐ๋ํ ๋ณํ์ ๋๋ค. AI ๊ธฐ์ ์ ํญ๋ฐ์ ์ธ ์ฑ์ฅ๊ณผ ์์ต ๋ชจ๋ธ์ ํ์ , ๊ทธ๋ฆฌ๊ณ ์ปดํจํธ ์ ๋์ ๊ฐ์ ์๋ก์ด ์ ํ์ ์ฌ๊ธฐ๋ ์ด ๋ณํ๊ฐ ์ด๋ฏธ ํ์ค์ด ๋์์์ ๋ณด์ฌ์ค๋๋ค. ์คํธ๋ผ์ดํ๋ ์ด๋ฌํ ํ๋ฆ์ ์ต์ ์ ์์, ์ธ๊ฐ ์ค์ฌ์ ๊ฒฝ์ ์ธํ๋ผ๋ฅผ ์์ด์ ํธ ์ค์ฌ์ โ์์จ์ ์ธ ์ธํฐ๋ท ๊ฒฝ์ (autonomous internet economy)โ์ ์ ํฉํ๋๋ก ์ฌ๊ตฌ์ถํ๊ณ ์์ต๋๋ค.
๊ฐ๋ฐ์ ๊ฒฝํ์ ์์ด์ ํธ ๊ฒฝํ์ผ๋ก ํ์ฅํ๊ณ , ์ ์ข ์ฌ๊ธฐ์ ๋์ํ๋ฉฐ, ์๋ก์ด ์์ต ๋ชจ๋ธ์ ์ง์ํ๋ ์คํธ๋ผ์ดํ์ ๋ ธ๋ ฅ์ ์ธํฐ๋ท ๊ฒฝ์ ์ ๋ฏธ๋๋ฅผ ํ์ฑํ๋ ๋ฐ ๊ฒฐ์ ์ ์ธ ์ญํ ์ ํ ๊ฒ์ ๋๋ค. ์๋ฐ๋ฆฌ ์๋์ ๋ง์ฒ๋ผ, ์์ด์ ํธ๊ฐ ์ธํฐ๋ท์ ์ฃผ์ ํ์์๊ฐ ๋ ๋ฏธ๋์ ๋๋นํ์ฌ ์คํธ๋ผ์ดํ๋ โ์ธํฐ๋ท์ GDP๋ฅผ ์ฆ๊ฐ์ํค๋โ ๋ณธ์ฐ์ ๋ฏธ์ ์ ์ง์์ ์ผ๋ก ์ํํด ๋๊ฐ ๊ฒ์ ๋๋ค.
โWhy Even Some Democrats Hate Californiaโs Billionaire Tax Proposalโ โ New York Times Podcasts ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
์บ๋ฆฌํฌ๋์ ์ต๋ง์ฅ์์ธ, ๋ฏผ์ฃผ๋น์กฐ์ฐจ ๋ฐ๋ํ๋ โ๋จ๊ฑฐ์ด ๊ฐ์โ
๋ฏธ๊ตญ ๊ฒฝ์ ์ ์ฌ์ฅ์ด์ ํ์ ์ ์๋์ธ ์บ๋ฆฌํฌ๋์์์ ์ ๋ก ์๋ ์ธ๊ธ ์ ์์ด ๋จ๊ฑฐ์ด ๋ ผ์์ ๋ถ๋ฌ์ผ์ผํค๊ณ ์์ต๋๋ค. ์ต๋ง์ฅ์๋ค์ ์์์ฐ์ 5%์ ์ผํ์ฑ ์ธ๊ธ์ ๋ถ๊ณผํ์๋ ์ด ์ ์์ ๋์ค์ ์๋์ ์ธ ์ง์ง๋ฅผ ๋ฐ๊ณ ์์์๋ ๋ถ๊ตฌํ๊ณ , ๊ณตํ๋น๊ณผ ์ด๊ณ ์ก ์์ฐ๊ฐ๋ค๋ฟ๋ง ์๋๋ผ ๋๋๊ฒ๋ ๋ฏผ์ฃผ๋น ๋ด๋ถ์ ์ฃผ์ ๋ ธ๋์กฐํฉ์์๋ ์ฐ๋ ค์ ๋ชฉ์๋ฆฌ๊ฐ ํฐ์ ธ ๋์ค๊ณ ์์ต๋๋ค. ๊ฒฝ์ ์ ๋ถํ๋ฑ ํด์๋ผ๋ ๋ช ๋ถ ์๋ ์์๋ ์ด โ๋ถ์ ์ฆ์ธโ ์คํ์ ๊ณผ์ฐ ์ด๋ค ์๋ฏธ๋ฅผ ๊ฐ์ง๋ฉฐ, ์บ๋ฆฌํฌ๋์์ ๋ฏธ๊ตญ ์ฌํ์ ์ด๋ค ํ์ฅ์ ์ผ์ผํฌ๊น์? ๋ด์ํ์์ค ํ์บ์คํธ โ๋ ๋ฐ์ผ๋ฆฌโ์ ๋ดํธ๋ฆฌ ํคํธ๋กํ(Natalie Kitroeff)์ ๋ก๋ ๋ก์ ํ(Laurel Rosenhall) ๊ธฐ์๊ฐ ์ด ๋ณต์กํ ์ํฉ์ ์ฌ์ธต ๋ถ์ํ์ต๋๋ค.
์บ๋ฆฌํฌ๋์, ์ต๋ง์ฅ์ ์ธ๊ธ ์ ์์ ์๋ง
์ต๊ทผ ์บ๋ฆฌํฌ๋์์์๋ ์ต๋ง์ฅ์ ์์ฐ์ธ(wealth tax) ๋์ ์ ์ํ ์ฃผ๋ฏผ ๋ฐ์์(ballot initiative)์ด ํฌํ์ ๋ถ์ณ์ง ๋งํผ ์ถฉ๋ถํ ์๋ช ์ ํ๋ณดํ๋ฉฐ ์ค๋ํ ์ ํ์ ์ ๋ง์์ต๋๋ค. ์ด ์ธ๊ธ์ ์ฌํด 1์ 1์ผ ๊ธฐ์ค ์บ๋ฆฌํฌ๋์ ๊ฑฐ์ฃผ ์ต๋ง์ฅ์์ ์์์ฐ(net worth)์ 5%๋ฅผ ์ผํ์ฑ์ผ๋ก ๋ถ๊ณผํ์๋ ๋ด์ฉ์ ๋ด๊ณ ์์ต๋๋ค. ์ง์๋ ์ธ๊ธ์ ์ฃผ ์ ๋ถ์ ํน๋ณ ๊ธฐ๊ธ(special fund)์ผ๋ก ๊ท์๋์ด ์๋ฃ ์๋น์ค์ 90%๊ฐ ์ฌ์ฉ๋๋๋ก ๋ช ์๋์ด ์์ต๋๋ค.
์ด ์ ์์ด ์ฃผ๋ชฉ๋ฐ๋ ์ด์ ๋ ๊ทธ ๊ตฌ์กฐ์ ๋น์ ์์ฑ ๋๋ฌธ์ ๋๋ค. ์ผ๋ฐ์ ์ผ๋ก ์ธ๊ธ์ ์๋์ ๋ถ๊ณผ๋์ง๋ง, ์ด ์ธ๊ธ์ ์ฃผ์, ๋ณด์, ์๋์ฐจ, ๋ฏธ์ ํ ๋ฑ ๊ฐ์ธ์ ์ด์์ฐ์ ๋ถ๊ณผ๋๋ค๋ ์ ์์ ๋งค์ฐ ์ด๋ก์ ์ ๋๋ค. ๋ค๋ง, ์ฃผ๊ฑฐ์ฉ ๋ถ๋์ฐ์ ๊ณผ์ธ ๋์์์ ์ ์ธ๋ฉ๋๋ค. ๋ํ, ๋๋ถ๋ถ์ ์ธ๊ธ์ด ์ฃผ ์ ๋ถ์ ์ผ๋ฐ ๊ธฐ๊ธ(general fund)์ผ๋ก ๋ค์ด๊ฐ ์ ์ถ์ง ์ง๋์๋ค์ด ์ฌ๋๊ป ์ฌ์ฉ๋๋ ๊ฒ๊ณผ ๋ฌ๋ฆฌ, ์ด ์ธ๊ธ์ ํน์ ๋ชฉ์ (์๋ฃ ์๋น์ค)์๋ง ์ฌ์ฉ๋๋๋ก ์ ํ๋์ด ์์ต๋๋ค.
์ง์ง์๋ค์ ์ด ์ธ๊ธ์ด ์ฝ 200์ฌ ๋ช ์ ์ต๋ง์ฅ์์๊ฒ ์ํฅ์ ๋ฏธ์น ๊ฒ์ด๋ผ ์์ํฉ๋๋ค. ์ด๋ค์ ํธ๋ผํ ํ์ ๋ถ์ ๋ฉ๋์ผ์ด๋(Medicaid) ๋ฐ ์ค๋ฐ๋ง์ผ์ด(Obamacare subsidies) ์ง์๊ธ ์ญ๊ฐ์ผ๋ก ์ธํ ์๋ฃ ์์คํ ์ ์๊ธฐ๋ฅผ ๋ง๊ณ , ๋ถํ๋ฑํ ์์คํ ์ ๋ฐ๋ก์ก๊ธฐ ์ํด ํ์์ ์ด๋ผ๊ณ ์ฃผ์ฅํฉ๋๋ค. ๋ณ์๊ณผ ์๊ธ์ค์ ์ ์งํ๊ณ ์บ๋ฆฌํฌ๋์ ์ฃผ๋ฏผ๋ค์ด ์์กดํ๋ ์๋ฃ ์์คํ ์ ์งํค๊ธฐ ์ํ ์ค์ํ ์๊ธ์ด๋ผ๋ ๊ฒ์ ๋๋ค.
๋จ๊ฑฐ์ด ์ฐฌ๋ฐ ๋ ผ์: ๋์ค์ ์ง์ง์ ๋ฏผ์ฃผ๋น์ ์ฐ๋ ค
์ด ์ต๋ง์ฅ์์ธ ์ ์์ 150๋ง ๋ช ์ ์๋ช ์ ๋ฐ์์ ์ ๋๋ก ๋์ค์ ํญ๋์ ์ง์ง๋ฅผ ๋ฐ๊ณ ์์ต๋๋ค. ์ฌ๋ก ์กฐ์ฌ ๊ฒฐ๊ณผ์ ๋ฐ๋ฅด๋ฉด ์ ๊ถ์์ ์ ๋ฐ ์ด์, ๋ฏผ์ฃผ๋น ์ง์ง์์ 70% ์ด์์ด ์ด ์ธ๊ธ์ ์ฐฌ์ฑํ๋ ๊ฒ์ผ๋ก ๋ํ๋ฌ์ต๋๋ค. โ๋ถ์๋ค์๊ฒ ์ธ๊ธ์ ๋งค๊ธฐ์โ๋ ๋จ์ํ ๊ตฌํธ๋ ๊ณต์์ด๋ ์ํผ๋ง์ผ ์์์ ์๋ช ์ด๋์ ํ๋ ์ฌ๋๋ค์๊ฒ ๋ง์ ๊ณต๊ฐ์ ์ป์์ต๋๋ค.
์ด ์ ์์ โ์๋น์ค ๋ ธ๋์ ๊ตญ์ ์ฐํฉ ํตํฉ ์๋ฃ ๋ ธ๋์ ์๋ถ ์ง๋ถ(SEIU-UHW)โ๋ผ๋ ํน์ ๋ ธ๋์กฐํฉ์ ์ฃผ๋๋ก ์์๋์์ต๋๋ค. ๋ณ์ ์ฒญ์๋ถ, ์๋ฆฌ์ฌ, ์๋ฃ ๊ธฐ์ ์ ๋ฑ ๋ณ์ ๋ ธ๋์๋ค์ ๋ํํ๋ ์ด ๋ ธ์กฐ๋ ํธ๋ผํ ๋ํต๋ น์ ๋ฉ๋์ผ์ด๋ ์ญ๊ฐ์ด ๋ณ์ ๋ฐ ์๊ธ์ค ํ์๋ก ์ด์ด์ง ๊ฒ์ ์ฌ๊ฐํ๊ฒ ์ฐ๋ คํ์ต๋๋ค. ๋ ธ์กฐ ์ง๋์ ๋ฐ์ด๋ธ ๋ฆฌ๊ฑด(Dave Regan)์ ์ฃผ๋ฏผ ๋ฐ์์์ ๋ ธ๋ ํ์ ๊ณผ์ ์์ ํ์ฉํ๋ ์ค๋ ์ด๋ ฅ์ ๊ฐ์ง ์ธ๋ฌผ๋ก, ์ ๋ช ํ ๊ฒฝ์ ํ์๋ค๊ณผ ํ๋ ฅํ์ฌ ์ด ์ต๋ง์ฅ์์ธ ์ ์์ ๋ ์์ ์ผ๋ก ์ถ์งํ์ต๋๋ค. ์ด๊ธฐ์๋ ๊ด๋ฒ์ํ ์ฐํฉ ์์ด ๋ ์์ ์ผ๋ก ์์๋์์ผ๋, ๋ฒ๋ ์๋์ค(Bernie Sanders) ์์์์์ด LA ์งํ์ ์ฐธ์ํ์ฌ ์ง์ง๋ฅผ ํ๋ช ํ๋ ๋ฑ ์ ์ฐจ ์ง์ง ๊ธฐ๋ฐ์ ํ์ฅํด ๋๊ฐ์ต๋๋ค.
ํ์ง๋ง ์ด๋ฌํ ๋์ค์ ์ง์ง์๋ ๋ณ๊ฐ๋ก, ์ต๋ง์ฅ์๋ค์ ๋ฐ๋ฐ์ ๊ฑฐ์ ์ต๋๋ค. ์ด๋ค์ ์์ฒ๋ง ๋ฌ๋ฌ๋ฅผ ๋ค์ฌ ์ธ๊ธ ๋์ ์ ๋ง๊ธฐ ์ํ ์ธ์์ ๋ฒ์ด๊ณ ์์ต๋๋ค.
์ต๋ง์ฅ์๋ค์ ๋ฐ๊ฒฉ: ์ดํ๊ณผ ์ ํญ
์ด ์ธ๊ธ ์ ์์ด ๋ณธ๊ฒฉํ๋์ ์ต๋ง์ฅ์๋ค์ ํจ๋์ ๋น ์ก์ต๋๋ค. ํนํ 1์ 1์ผ ๊ธฐ์ค ์บ๋ฆฌํฌ๋์ ๊ฑฐ์ฃผ์์๊ฒ ์ธ๊ธ์ด ๋ถ๊ณผ๋๋ค๋ ์กฐํญ ๋๋ฌธ์, ์๋ ๋ง๋ถํฐ ์ผ๋ถ ๊ณ ์ก ์์ฐ๊ฐ๋ค์ ์บ๋ฆฌํฌ๋์๋ฅผ ๋ ๋๋ ์์ง์์ ๋ณด์์ต๋๋ค. ๊ตฌ๊ธ ๊ณต๋ ์ฐฝ์ ์ ์ธ๋ฅด๊ฒ์ด ๋ธ๋ฆฐ(Sergey Brin)์ ๋ค๋ฐ๋ค์ฃผ ํํธ ํธ์ ์ธ๊ทผ์ผ๋ก, ๋๋ฆฌ ํ์ด์ง(Larry Page)๋ ํ๋ก๋ฆฌ๋ค๋ก ๊ฑฐ์ฃผ์ง๋ฅผ ์ฎ๊ฒผ์ต๋๋ค. ๊ธฐ์ ํฌ์๊ณ์ ๊ฑฐ๋ฌผ์ธ ๋ฐ์ด๋น๋ ์์ค(David Sacks)๋ ํ ์ฌ์ค๋ก, ํผํฐ ํธ(Peter Thiel)์ ๋ง์ด์ ๋ฏธ๋ก ์ด์ฃผํ์ต๋๋ค. ๋ถ์ ์ธต์ด ์ธ๊ธ ์ธ์ ์ํ์ ์ฃผ๋ฅผ ๋ ๋๊ฒ ๋ค๊ณ ๊ณต์ธํ๋ ๊ฒฝ์ฐ๋ ๋ง์ง๋ง, ์ค์ ๋ก ํ๋์ ์ฎ๊ธด ์ฌ๋ก๋ ์ด๋ก์ ์ ๋๋ค.
์บ๋ฆฌํฌ๋์์ ๋จ๊ธฐ๋ก ํ ์ต๋ง์ฅ์๋ค์ ์ธ๊ธ ๋ถ๋ด์ ์ค์ด๊ธฐ ์ํ ๋ฐฉ์์ ๋ชจ์ํ๊ณ ์์ต๋๋ค. ์ธ๋ฌด ์๋ฌธ๊ฐ์ ์์ฐ ๊ด๋ฆฌ์๋ค์ ๊ณ ๊ฐ๋ค์๊ฒ โ๋น๋ฒ๋ฆฌํ์ค ์ ํ์ ํผ์นด์ ๊ทธ๋ฆผ์ ์์คํ ๋ณ์ฅ์ผ๋ก ์ฎ๊ธฐ๊ฑฐ๋, ๊ณ ๊ฐ ๋ค์ด์๋ชฌ๋ ๋ชฉ๊ฑธ์ด์ ๋ณดํ ๊ฐ์ก์ ์ค์ ๊ฐ์น๋ก ์ค์ด๋ผโ๋ ๋ฑ์ ์กฐ์ธ์ ํ๊ณ ์์ต๋๋ค. ์ด๋ ์์ฐ์ธ๋ฅผ ํํผํ๊ธฐ ์ํ ๊ตฌ์ฒด์ ์ธ ์ ๋ต๋ค์ ๋๋ค.
์ด๋ค์ด ์ด ์ธ๊ธ์ ๋ฐ๋ํ๋ ์ฃผ์ฅ์ ๋จ์ํ โ๋ด ๋์ ๊ฐ์ ธ๊ฐ์ง ๋ง๋ผโ๋ ์์ค์ ๋์ด์ญ๋๋ค.
- ์์ฐ์ธ์ ๋ณธ์ง์ ๋ฌธ์ : ์์ฐ์ ํ๊ฐํ๊ณ ๊ณผ์ธํ๋ ๋ฐฉ์ ์์ฒด๊ฐ ๋ฌธ์ ๋ผ๊ณ ์ง์ ํฉ๋๋ค. ๋ชจ๋ ๋ฉ์ธ์๊ฐ ์์ ์ ๋ชจ๋ ์์ฐ ๊ฐ์น๋ฅผ ์ ๊ณ ํด์ผ ํ๋ ๊ฒ์ ๋ถ๋นํ๋ค๋ ๊ฒ์ ๋๋ค.
- ์๊ธ ์ ์ฉ์ ์ํ์ฑ: 1์ 1์ผ ๊ธฐ์ค์ผ๋ก ์๊ธ ์ ์ฉ๋๋ ๊ฒ์ด ์ํ์ ์ด๋ผ๊ณ ์ฃผ์ฅํฉ๋๋ค.
- ํ์ ๊ฒฝ์ ์ ๋ฏธ์น ์น๋ช ์ ์ํฅ: ์ค๋ฆฌ์ฝ๋ฐธ๋ฆฌ์์๋ ์ด ์ธ๊ธ์ด ์บ๋ฆฌํฌ๋์์ ํ์ ์ฃผ๋ ๋ฅ๋ ฅ์ ์ฌ๊ฐํ๊ฒ ํผ์ํ ๊ฒ์ด๋ผ๊ณ ๊ฒฝ๊ณ ํฉ๋๋ค. ์ต๋ง์ฅ์๋ค์ด ๋ ๋๋ฉด ์ฃผ(ๅท)๊ฐ ์์ ๊ฒ์ด ๋ง๋ค๋ ๋ ผ๋ฆฌ์ ๋๋ค.
- ์ฅ๊ธฐ์ ์ธ์ ์์ค ์ฐ๋ ค: ํ์ฌ ์ต๋ง์ฅ์๋ค์ด ๋ด๋ ๋ง๋ํ ์๋์ธ๋ ์บ๋ฆฌํฌ๋์์ ํ๊ต, ๊ณต๊ณต ์์ , ๋๋ก ๋ฑ ๋๋ถ๋ถ์ ๊ณต๊ณต ์๋น์ค๋ฅผ ์งํฑํ๊ณ ์์ต๋๋ค. ์ผํ์ฑ ์์ฐ์ธ๋ก ์ธํ ๋จ๊ธฐ์ ์ธ ์ธ์ ์ฆ๊ฐ๋ ์ต๋ง์ฅ์๋ค์ ์ดํ๋ก ์ธํ ์ฅ๊ธฐ์ ์ธ ์๋์ธ ์์ค๋ก ์ด์ด์ ธ ์ฃผ ์ฌ์ ์ ๋ ํฐ ํ๊ฒฉ์ ์ค ์ ์๋ค๋ ์ฐ๋ ค์ ๋๋ค.
์์ธก ๋ถํ์ ๋๋งน: ๋ฏผ์ฃผ๋น ๋ด์ ๊ท ์ด
์ด๋ฌํ ์ฐ๋ ค๋ ์ต๋ง์ฅ์๋ค๋ง์ ์ ์ ๋ฌผ์ด ์๋๋๋ค. ๋ฏผ์ฃผ๋น ๋ด๋ถ์์๋ ์ด ์์ฐ์ธ์ ๋ํ ํ์์ ์ธ ์๊ฐ์ด ์กด์ฌํฉ๋๋ค.
๊ฐ๋น ๋ด์ฌ(Gavin Newsom) ์บ๋ฆฌํฌ๋์ ์ฃผ์ง์ฌ๋ ์ด ์ธ๊ธ์ ๋ํด ๊ณต๊ฐ์ ์ผ๋ก ๋ฐ๋ ์์ฌ๋ฅผ ํ๋ช ํ์ต๋๋ค. ๊ทธ๋ โ๋จ๊ธฐ์ ์ธ ์ธ์ ์ฆ๊ฐ๋ณด๋ค๋ ์ฅ๊ธฐ์ ์ธ ์๋์ธ ์์คโ์ ์ฐ๋ คํ๋ฉฐ, ์บ๋ฆฌํฌ๋์์๋ง ์ ์ฉ๋๋ ์์ฐ์ธ๋ ์ฃผ๋ฅผ ๋ถ๋ฆฌํ๊ฒ ๋ง๋ค ๊ฒ์ด๋ผ๊ณ ์ฃผ์ฅํฉ๋๋ค. ๋์ ๊ทธ๋ โ์ ๊ตญ์ ์ธ ์์ฐ์ธ ๋ ผ์โ์ ํ์์ฑ์ ์ธ๊ธํ๋ฉฐ, ์บ๋ฆฌํฌ๋์๊ฐ ์ด๋ฏธ ๋งค์ฐ ์ง๋ณด์ ์ธ ์๋์ธ ๊ตฌ์กฐ๋ฅผ ๊ฐ์ง๊ณ ์์ด ๋ถ์ ์ธต์๊ฒ ๋ง์ ์ธ๊ธ์ ๋ถ๊ณผํ๊ณ ์์์ ๊ฐ์กฐํฉ๋๋ค. ๊ทธ๋ ๋ถ์ ์ธต ๊ณผ์ธ๋ผ๋ ๋๋์ ์ฃผ์ฅ์ ๋์ํ๋ฉด์๋, ์บ๋ฆฌํฌ๋์์ ์ด์ต์ ํด์น์ง ์๋ ์ ์์ ์ ๊ทผํ๋ ค๋ โ์คํ๊ธฐ(threading a needle)โ ์ ๋ต์ ๊ตฌ์ฌํ๊ณ ์์ต๋๋ค.
์ฃผ์ ๋ ธ๋์กฐํฉ์ธ ๊ต์ฌ ๋ ธ์กฐ์ SEIU ์ฐํ ๋จ์ฒด๋ค๋ ์ด ์ธ๊ธ์ ๋ํด ๊ณต๊ฐ์ ์ธ ์ ์ฅ์ ํ๋ช ํ์ง ์๊ณ ์์ต๋๋ค. ์ด๋ค์ ์ธ๊ธ์ด ํน์ ๋ชฉ์ (์๋ฃ)์๋ง ์ฌ์ฉ๋๋ ํน๋ณ ๊ธฐ๊ธ์ผ๋ก ๋ฌถ์ฌ ํ๊ต, ๊ณต๊ณต ์์ , ๋๋ก ๋ฑ ๋ค๋ฅธ ์ค์ํ ํ๋ก๊ทธ๋จ์ ์๊ธ์ด ์ง์๋์ง ์๋ ๊ฒ์ ๋ํด ์ฐ๋ คํ๊ณ ์์ต๋๋ค.
๊ฒฐ๊ณผ์ ์ผ๋ก ์ด ์ฃผ๋ฏผ ๋ฐ์์์ โ์ด์ํ ๋๋งน(strange bedfellows coalition)โ์ ํ์ฑํ ๊ฐ๋ฅ์ฑ์ด ์์ต๋๋ค. ์ต๋ง์ฅ์๋ค์ด ์์ ๋ค์ ์ดํด๊ด๊ณ๋ฅผ ๋๋ณํ๋ ๋ค๋ฅธ ์ง๋จ๋ค๊ณผ ์ฐํฉํ์ฌ ์ ๊ถ์๋ค์๊ฒ ๋ ์ค๋๋ ฅ ์๋ ์ฃผ์ฅ์ ํผ์น ์ ์๋ค๋ ๊ฒ์ ๋๋ค. ๋ํ, ์ฃผ๋ฏผ ๋ฐ์์์ด ํฌํ์ ๋ถ์ณ์ง๊ธฐ ์ ๋ช ๋ฌ ๋์ ๋งํ ํ์์ ํตํด ๋ฐ์์ ์์ฒด๋ฅผ ์ฒ ํ์ํค๋ ค๋ ์๋๊ฐ ์์ ์๋ ์์ต๋๋ค. ์ฃผ์ง์ฌ๊ฐ ๊ฐ์ ํ์ฌ ํ์์ ์ค์ฌํ ๊ฐ๋ฅ์ฑ๋ ์ ๊ธฐ๋ฉ๋๋ค. ๊ทธ๋ ์ง ์์ผ๋ฉด, ์ ๊ถ์๋ค์ ์์ถฉํ๋ ์ฌ๋ฌ ์ธ๊ธ ๊ด๋ จ ๋ฐ์์๋ค ์ฌ์ด์์ ํผ๋์ค๋ฌ์ด ์ ํ์ ํด์ผ ํ๋ ๊ฐ๋น์ผ ์ ๊ฑฐ์ ์ ๋ชฉ๊ฒฉํ๊ฒ ๋ ๊ฒ์ ๋๋ค.
๋ฐ๋ฉด, ๋ฒ๋ ์๋์ค ๊ฐ์ ๋ฏผ์ฃผ๋น์ ์ง๋ณด์ ์ธ์ฌ๋ค์ โ์ข์ ๊ฒ์ ์ ์ด ์๋ฒฝํจ์ด ๋๊ฒ ํ์ง ๋ง๋ผโ๋ฉฐ ์ต๋ง์ฅ์ ๊ณผ์ธ ์์ฒด๋ฅผ ์ง์งํฉ๋๋ค. ๊ทธ๋ค์๊ฒ๋ ์ธ๊ธ์ ์ธ๋ถ ์ฌํญ๋ณด๋ค๋ ๋ถ์ ์ธต์๊ฒ ์ธ๊ธ์ ๋ถ๊ณผํ์ฌ ์์ธ๋ ๊ณ์ธต์ ๋๋๋ค๋ ํฐ ๊ทธ๋ฆผ์ด ๋ ์ค์ํฉ๋๋ค.
์บ๋ฆฌํฌ๋์๋ฅผ ๋์ด์ ์์ฌ์
์บ๋ฆฌํฌ๋์์ ์ต๋ง์ฅ์์ธ ๋ ผ์์ ๋ค๋ฅธ ์ฃผ(๋ด์์ ํผ์๋์ํ ๋ฅด์ธ(pied-ร -terre tax) ๋ ผ์ ๋ฑ)์๋ ์ค์ํ ๊ตํ์ ๋์ ธ์ค๋๋ค.
์ฒซ์งธ, โ๋ถ์๋ค์๊ฒ ์ธ๊ธ์ ๋งค๊ธฐ์โ๋ ์์ด๋์ด๋ ๋งค์ฐ ์ธ๊ธฐ๊ฐ ๋ง์ต๋๋ค. ๊ฒฝ์ ์ ๋ถํ๋ฑ์ ๋ํ ๋ถ์๊ฐ๊ณผ ๋ฐ(ๅ)ํธ๋ผํ ์ ์๊ฐ ๊ฒฐํฉ๋์ด ๋์ค์ ๊ฐ์ ์ ์ง์ง๋ฅผ ์ฝ๊ฒ ์ป์ ์ ์์ต๋๋ค. ๋์งธ, ๊ทธ๋ผ์๋ ๋ถ๊ตฌํ๊ณ ์ด๋ฌํ ์ ์ฑ ์ ์ค์ ๋ก ํต๊ณผ์ํค๊ณ ์คํ์ ์ฎ๊ธฐ๋ ๊ฒ์ ๋งค์ฐ ์ด๋ ต์ต๋๋ค. ๋ณต์กํ ์ธ๊ธ ๊ตฌ์กฐ ๋ฌธ์ ์ ์ต๋ง์ฅ์ ๊ณ์ธต์ ๋ง๋ํ ์ํฅ๋ ฅ ๋๋ฌธ์ ๋๋ค. ์ ์งธ, ์ฑ๊ณต์ ์ธ ํต๊ณผ๋ฅผ ์ํด์๋ ๊ด๋ฒ์ํ๊ณ ๊ฐ๋ ฅํ ์ง์ง ๊ธฐ๋ฐ์ด ํ์์ ์ ๋๋ค. ๋จ์ํ ๋์ค์ ์ธ๊ธฐ์๋ง ์์กดํด์๋ ๊ฐ๋ ฅํ ๋ฐ๋ ์ธ๋ ฅ์ ์ด๊ฒจ๋ด๊ธฐ ์ด๋ ต์ต๋๋ค.
์ต๋ง์ฅ์๋ค์ด ์ด ์์ฐ์ธ๋ฅผ โ์ค์กด์ ์ํ(existential threat)โ์ผ๋ก ์ฌ๊ธฐ๋ ์ด์ ๋ ๋จ์ํ 5%์ ์ผํ์ฑ ์ธ๊ธ์ ๋ด๊ธฐ ์ซ์ด์๊ฐ ์๋๋๋ค. ๊ทธ๋ค์ ์ด ์ธ๊ธ์ด ๋ฏธ๋์ ๋ถ์ ๊ณผ์ธ์ ๋ํ ์ ๋ก๋ฅผ ๋ง๋ค๊ณ , ์ฌ๋๋ค์ด ํ๋ค๊ฒ ๋ฒ ์ฌ์ฐ๊ณผ ์์ ๋ฌผ์ ์ธ๊ธ์ ๋ถ๊ณผํ๋ ์๋ก์ด ํจ๋ฌ๋ค์์ ์ ์ํ ๊น ๋ด ๋๋ ค์ํฉ๋๋ค. ์ด๋ ๋ง์ ๋ถ์ ์ธต์ด ์กฐ์ธ ํํผ๋ฅผ ์ํด ์์ฐ์ ๋ค์ํ ๊ณ์ข๋ก ์ฎ๊ฒจ๋์ ์ํฉ์์, ์ ๋ถ๊ฐ ์ค์ ๋ถ(ๅฏ)์ ์ ๊ทผํ์ฌ ์ฌํ์ ๋ ๋ง์ ๊ธฐ์ฌ๋ฅผ ์๊ตฌํ๋ ์๋ก์ด ๋ฐฉ์์ด ๋ ์ ์๊ธฐ ๋๋ฌธ์ ๋๋ค.
์บ๋ฆฌํฌ๋์์ ์ด ์คํ์ ์ฐ๋ฆฌ ์ฌํ๊ฐ ์ง๋ฉดํ ์ต๋ง์ฅ์ ๊ณ์ธต๊ณผ ๊ฒฝ์ ์ ๋ถํ๋ฑ์ ๋ํ ๋ถํธํ ์ง์ค์ ์ ๋ฉด์ผ๋ก ๋ง์ฃผํ๊ฒ ํฉ๋๋ค. ์ด ๋ ผ์์ ๋จ์ํ ์ธ๊ธ ๋ฌธ์ ์ ๊ทธ์น์ง ์๊ณ , ๋ถ์ ์ฌ๋ถ๋ฐฐ, ์ฌํ์ ์ฑ ์, ๊ทธ๋ฆฌ๊ณ ์ ๋ถ์ ์ญํ ์ ๋ํ ๊ทผ๋ณธ์ ์ธ ์ง๋ฌธ์ ๋์ง๋ฉฐ ๋ฏธ๊ตญ ์ ์ญ์ ๊ฒฝ์ ์ ๋ถํ๋ฑ ๋ ผ์๋ฅผ ์๋ฉด ์๋ก ๋์ด์ฌ๋ฆฌ๊ณ ์์ต๋๋ค. ์บ๋ฆฌํฌ๋์์ ์ ํ์ด ์์ผ๋ก ๋ฏธ๊ตญ ์ฌํ์ ์ด๋ค ๋ณํ๋ฅผ ๊ฐ์ ธ์ฌ์ง ๊ท์ถ๊ฐ ์ฃผ๋ชฉ๋ฉ๋๋ค.