English
Based on โHow Intercom 2Xโd engineering velocity with Claude Code | Brian Scanlanโ from How I AI Watch the original video
Unleashing the Code: How Intercom Doubled Engineering Velocity with AI and Reimagined the Software Factory
In the rapidly evolving landscape of artificial intelligence, many companies are grappling with how to integrate these powerful tools effectively. While some express skepticism about AIโs true impact on productivity, Intercom, the customer messaging platform, stands as a compelling counter-narrative. Under the guidance of senior principal engineer Brian Scanlan, Intercomโs R&D department has achieved a remarkable feat: doubling their engineering throughput in just nine months, largely thanks to a full-throttle adoption of AI, particularly Claude Code.
This isnโt just about incremental gains; itโs a fundamental shift in how software is built, challenging long-held beliefs about developer productivity, code quality, and the very nature of engineering work.
The Moment of Inflection: Imagination, Not Tools, Is the Barrier
For years, the promise of AI in software development felt just out of reach. Code completion tools offered minor conveniences, but nothing truly transformative. Brian Scanlan recounts a pivotal moment around November-December last year, coinciding with the release of more advanced models like Anthropicโs Opus 46 and GPT-4U. โSuddenly you started realizing that you have to think bigger about things,โ Scanlan explains, โor that your imagination is now the barrier, not the tool.โ
This breakthrough wasnโt just internal; it was a collective awakening. Scanlan describes the atmosphere during the Christmas break, with โeverybody go wild on Twitter X,โ witnessing developers achieve unprecedented levels of output. The message was clear: โEverythingโs changed.โ
Intercomโs journey was fueled by an internal AI-first mandate that permeated their product strategy. Having already transformed their customer-facing products with AI, the expectation for similar innovation within engineering was high. This created an environment where impatience for AI adoption was a driving force, pushing the team to go โall-inโ rather than dabble.
Measuring the Unmeasurable: 2x Velocity and the Path to 10x
To quantify their progress, Intercomโs CTO, Darra, set an ambitious goal: double the throughput of R&D. They chose a simple, albeit sometimes debated, metric: merged Pull Requests (PRs) per R&D head. This metric, encompassing product managers, designers, and TPMs alongside engineers, revealed a dramatic โnumber goes upโ chart. Within nine months, the team saw a 2x increase in PR throughput.
This isnโt merely about engineers typing faster; itโs about unlocking โthe physical limits of my ability to type code,โ as Claire Vote, the host of โHow I AI,โ eloquently puts it. The shift is so profound that Scanlan now asks, โWhy canโt it be 10x?โ This ambitious outlook stems from a belief that if an organization fully commits, prepares its team, and refines its codebase, it will see significant improvements in product quality and developer experience.
Addressing the common skepticism that such metrics might be โgamedโ by shipping smaller, less impactful changes, Scanlan emphasizes Intercomโs high-trust environment. The goal isnโt just speed; itโs about enabling engineers to do more meaningful work, faster and with greater enjoyment. Indeed, Scanlan shares, โIโve been having the most amount of fun in my career over the last 3 months.โ
Beyond Speed: The Surprising Uplift in Code Quality
A frequent concern among engineers and leaders alike is whether increased velocity comes at the cost of quality. Will AI-generated code be โslopโ or โgarbageโ? Intercomโs experience suggests the opposite.
Theyโve observed a consistent decrease in the time it takes from the first line of code written to a feature being announced on their news channel. More significantly, Intercom collaborated with a research group at Stanford, providing them with their internal data. The Stanford teamโs independent measures of code quality indicated that code quality was actually improving.
This counter-intuitive finding highlights a crucial benefit of AI-driven development: the capacity to tackle technical debt and improve the codebase. As Scanlan and Vote discuss, businesses often have limited capacity for โinternal projectsโ like improving code quality, as these donโt directly generate revenue. However, when the โcost of doing that compressesโ due to AIโs efficiency, organizations can afford to invest in developer experience, security, compliance, maintainability, flaky tests, and CI/CD improvements.
This leads to a powerful piece of advice for CTOs and VPs of engineering: โEverything you hate about the codebase, go spend a month fixing and see how fast we can speedrun that. Thatโs going to feel really good.โ This strategic investment in core engineering health, made feasible by AI, ultimately unlocks even greater velocity and higher quality.
The Software Factory: Engineering the Golden Path
Intercomโs success isnโt just about adopting AI; itโs about intentionally engineering their entire software delivery process around it, moving towards what Scanlan calls a โsoftware factory.โ This involves treating the engineering organization itself like a product, complete with rigorous design, measurement, and continuous improvement.
One striking example is how Intercom addressed the issue of pull request (PR) descriptions. Initially, AI-generated PR descriptions were โterrible,โ merely regurgitating code changes rather than explaining intent โ the truly valuable information for human reviewers. Intercomโs solution involved:
- Defining Quality: They established what a โgoodโ PR description should look like.
- LLM Judge: An AI model was trained to evaluate PR descriptions, revealing a negative trend in quality with early AI adoption.
- Custom Skill: They developed a โcreate PRโ skill for Claude Code that leverages session context to generate high-quality, intent-driven descriptions.
- Enforcement: This skill was integrated as a mandatory hook. If an engineer or agent tries to open a PR without using the skill, itโs blocked.
This approach mirrors the determinism of modern CI/CD pipelines, but applied upstream to the code-writing process itself. โWeโre on this movement towards a software factory,โ Scanlan explains, where predictability, reliability, and consistent quality are paramount. This ensures that while engineers move faster, they still adhere to the high standards of a mature, 15-year-old SaaS company.
The Invisible Infrastructure: Telemetry and Personalized Feedback
Underpinning Intercomโs โsoftware factoryโ is a sophisticated system for monitoring and improving AI usage. They donโt โfly blind.โ
- Skill Telemetry with Honeycomb: Every internal AI skill is instrumented with event-level telemetry, sending data to Honeycomb. This allows individual skill developers to see how often their skills are invoked, by whom, and when, fostering a data-driven approach to skill development.
- Session Data Analysis with S3: All raw Claude Code session data (the chat logs) is anonymized, uploaded to S3, and then analyzed. This allows Intercom to identify broader organizational trends, common pitfalls, and areas where training or new skills are needed.
- Personalized Feedback: Based on this session data, Intercom developed a simple internal tool that provides personalized feedback to individual engineers on their Claude Code usage. This helps new hires or those struggling to understand how to optimize their interactions, supporting a culture of continuous learning and self-improvement.
This robust telemetry system ensures that Intercom can identify bottlenecks, improve their internal tools, and provide targeted support, preventing the common problem of โthrowing an API key and saying best of luck.โ
The Elephant in the Room: AI Costs
The exponential growth in AI usage inevitably raises questions about cost. Scanlan readily admits that their AI bill โlooks exactly like thisโ (referencing the steep velocity growth chart). โItโs like hiring whole new offices of people,โ he notes.
Intercomโs current strategy is to prioritize speed over cost optimization. Their attitude is, โeveryone just turn on Opus for everythingโฆ going as fast as possible and caring about the bill later.โ This reflects a strategic investment mindset, believing that the significant benefits gained from rapid innovation and increased throughput outweigh the immediate costs. However, Scanlan acknowledges this might not be feasible for every business and that future optimization will be necessary if costs continue at the current rate.
The Future: Agent-First and Higher-Level Concerns
Looking ahead, Intercom envisions a future where โall technical work will become agent first.โ This means AI agents will handle the foundational, repetitive tasks, freeing human engineers to operate at a higher level, focusing on complex problem-solving, strategic design, and truly innovative features.
The shift is profound: instead of engineers spending time on boilerplate code or debugging minor issues, agents will take on this โbasic work,โ allowing humans to โmove up to higher level to be able to like work on higher level concerns or just getting more stuff built more stuff out there or higher quality.โ
Intercomโs journey with Claude Code is more than just a case study in productivity; itโs a blueprint for reimagining the entire software development lifecycle. By embracing AI comprehensively, treating the organization as a product, meticulously measuring outcomes, and proactively engineering for quality, Intercom has not only doubled its engineering velocity but has also cultivated a more engaging, productive, and ultimately more fun environment for its engineers. For any organization looking to leverage AI in engineering, Intercomโs experience offers invaluable lessons and a compelling vision for the future.
Based on โInside the Five Days That Remade the Supreme Courtโ from New York Times Podcasts Watch the original video
The Secret Birth of the Supreme Courtโs Shadow Power
For the better part of a decade, the U.S. Supreme Court has operated with a parallel, often perplexing system for issuing major rulings โ one that bypasses the traditional, deliberate processes that define American justice. This increasingly powerful mechanism, dubbed the โshadow docket,โ has been responsible for consequential decisions on everything from immigration policy to presidential authority, often with little to no public explanation. Now, a groundbreaking New York Times investigation has pierced this veil of secrecy, revealing the precise, dramatic five days in 2016 when this expedited system was born, effectively remaking the nationโs highest court.
At the heart of this revelation are 16 pages of confidential correspondence among the justices themselves, offering an unprecedented look into their private deliberations. These documents, unearthed by Times journalists Jodie Caner and Adam Liptac, allow us to โeavesdrop on the justices at the exact moment that they are abandoning time-tested norms of judicial procedure and backing themselves into a new way of doing business.โ
Understanding the Shadow Docket
To truly grasp the significance of this shift, one must understand the stark contrast between the Supreme Courtโs traditional โmerits docketโ and its shadowy counterpart. The merits docket is the familiar image of the court: justices meticulously select cases, receive multiple rounds of detailed briefs, hear extensive oral arguments, engage in in-person deliberations, and then exchange numerous drafts of lengthy, reasoned opinions, concurrences, and dissents. This painstaking process, often stretching over a year, culminates in decisions that can be hundreds of pages long, providing binding law and clear guidance to lower courts and the nation.
The shadow docket, however, short-circuits all of this. It operates with extreme speed, relying on thin briefs, no oral arguments, and no in-person deliberations. Its rulings are typically brief, often consisting of scant or even no reasoning at all. Yet, over the last ten years, it has become a major part of the courtโs business, particularly during the Trump administration, where it was used to grant the president enormous leeway on issues like immigration, government spending, and agency power. Critics have consistently warned that this exponential increase in use, combined with the lack of transparent reasoning, raises critical questions: Is the court rushing to rule based on gut instinct, personal pique, or partisan impulse โ precisely the elements a slow, deliberate, and judicious system is designed to avoid?
Defenders of the shadow docket argue that emergency orders have a long history, typically reserved for truly time-sensitive matters like death penalty cases or election disputes. They also contend that these orders are merely temporary, designed to maintain the status quo while cases proceed through lower courts. However, as Caner and Liptac point out, this โtemporaryโ label often masks a very different reality. If the court allows the deportation of thousands of people, or the withholding of aid money, or the firing of employees, those actions, though nominally temporary, are often irreversible and conclusive in their practical effect. They are, in essence, final decisions made without the benefit of the courtโs usual rigorous process.
The Unprecedented Request: Obamaโs Clean Power Plan
The crucible for this seismic shift was a case in 2016 concerning President Barack Obamaโs Clean Power Plan. Facing legislative gridlock on climate change, Obama had instructed the Environmental Protection Agency (EPA) to issue regulations that would fundamentally move the American power system away from coal. This initiative, predictably, enraged industry groups and โred states,โ who challenged it in the DC Circuit Court. They asked the court not only to declare the plan unlawful but also to immediately halt its implementation. The DC Circuit agreed to fast-track arguments on the planโs legality but refused to pause its rollout.
In an โunprecedented request,โ the challengers then turned directly to the Supreme Court. They asked the justices to freeze the Clean Power Plan before any lower court had even ruled on its lawfulness โ a move that everyone involved recognized as extraordinary. At this moment, the Supreme Court was evenly split, a 5-4 court leaning conservative, but with Justice Anthony Kennedy, a Republican appointee, serving as the crucial swing vote. Kennedy, famously โpersuadable,โ had recently authored the majority opinion legalizing gay marriage, making the courtโs direction unpredictable.
Five Days That Shook the Court
The emergency request landed in the chambers of Chief Justice John Roberts, who oversees the DC Circuit. The ordinary expectation was a swift denial, or at most, a collective denial by his colleagues. What transpired instead was a five-day sprint of internal memos that would forever alter the courtโs operational landscape.
Day One: Robertsโs Salvo Chief Justice Roberts initiated the debate with a three-page, single-spaced memo. He argued forcefully that Obamaโs plan must be halted due to the โenormous burdensโ it would impose on states and the coal industry. He claimed there was โno time to wasteโ and questioned the EPAโs authority under the Clean Air Act, invoking the โmajor questions doctrineโ โ a principle suggesting that unless Congress clearly grants an agency vast power, that power doesnโt exist.
But beyond legal arguments, Robertsโs memo betrayed a deep-seated grievance. He felt the EPA had โtrickedโ the court just months earlier in a mercury emissions case. In that instance, the court had ruled against the EPA after a three-year litigation, but by then, the regulation had effectively gone into effect, rendering the ruling somewhat moot. Roberts was โpeeved,โ โirked,โ and determined โnot to let it happen again.โ He viewed the traditional slow legal process as enabling the Obama administration to be โsneakyโ in implementing its regulations.
Day One: Breyerโs Counterpoint Justice Stephen Breyer, a Democratic appointee, immediately responded, laying out a contrasting chronology. He pointed out that the Clean Power Plan didnโt require industry action for six years, with total compliance not due until 2030. There was, in his view, โplenty of time to do this in the usual course.โ Breyer found the courtโs intervention โquite unusualโ and saw no urgent need to bypass the DC Circuit, advocating for the lower court to be allowed to finish its work.
Day Two: Robertsโs Insistence Robertsโs reply the next day revealed a growing irritation and an unyielding insistence on blocking the plan. He dismissed Breyerโs procedural concerns: โI recognize that the posture of this stay request is not typical, but review is sought of what has been described as the most expensive regulation ever imposed on the power sector.โ The Chief Justiceโs impatience was palpable; he was โready to rule now.โ He believed the court would ultimately strike down the Clean Power Plan, stating it was โhighly unlikely to surviveโ a full review. He saw no reason to let the process play out when he already knew the answer.
This memo made it clear that Roberts was engaged in a power struggle with the Obama administration. He explicitly stated, โI am of the mind that a rule designed to transform a substantial swath of the nationโs economy should be tested by this court before it is presented as a fait accompli.โ As Liptac observes, this was not the โeven-handed, magisterial toneโ typically seen in public; here, Roberts was โacting as a bulldozer.โ
Day Three: Kagan Sounds the Alarm Justice Elena Kagan, another Democratic appointee, filed an even more direct memo, opposing the Chief Justiceโs stance. She called the relief sought โuniqueโ and used the word โunprecedentedโ to describe what Roberts wanted to do. Kagan stressed the complexity of the case, arguing that it involved โa complex statutory and regulatory regimeโ that demanded more time and consideration. โThis is weird. Are we sure we want to do this?โ she seemed to ask, highlighting the gravity of the situation.
Day Four: Alitoโs Existential Threat Justice Samuel Alito, a Republican appointee, then weighed in, echoing Robertsโs sense of insult and grievance. He declared that a โfailure to stay this rule threatens to render our ability to provide meaningful judicial review and by extension our institutional legitimacy a nullity.โ Alito framed the issue as an โalmost existential threatโ to the court itself, portraying the Obama administration as attempting to sideline the justices.
Whatโs extraordinary about these private exchanges, the journalists emphasize, is how frankly the justices reveal their โactual agendas.โ Unlike their public decisions, which often present a โmaskโ of confining themselves to facts and legal materials, these memos expose the human element: grievances with an administration, fears about the courtโs institutional place, and concerns about legitimacy โ factors not typically associated with purely legal reasoning.
Day Five: Kennedyโs Decisive Whisper The entire debate, as predicted, came down to Justice Anthony Kennedy. On February 9th, the fifth day, Kennedy sent a terse, three-sentence note: he was voting with the Chief. The debate was over. Within hours, the Supreme Court issued its order, a single paragraph of โlegal boilerplateโ with no explanation, blocking President Obamaโs signature environmental initiative.
The Enduring Legacy of the Shadow Docket
The insights gleaned from these confidential memos profoundly vindicate the long-standing criticisms of the shadow docket. This was โnot the court doing A+ work,โ Caner and Liptac conclude. Instead, it was a court โthrowing ideas around, seeming to be motivated by grievances against a presidentโฆgetting snippy with each other and just in general not doing the kind of work we associate with the nationโs highest court.โ The justices were disregarding time-tested procedures without seemingly considering the long-term implications. As Kagan noted, it was โunprecedented,โ but no one seemed to ask: โwhere is it going to lead?โ
Where it led was to an explosion of emergency applications under the Trump administration, with the court often quickly ruling in favor of presidential initiatives on major questions, a stark contrast to the Obama era. Political scientists have observed more partisan voting on the shadow docket than on the merits docket, suggesting that when acting fast, justices may rely more heavily on partisan impulses. For instance, in the Biden years, the court initially voted against Biden on three emergency applications, only to rule for him when those same cases returned to the merits docket for full consideration. This pattern suggests that deliberation truly โdampens partisan impulses.โ
Beyond partisan concerns, the rise of the shadow docket poses a fundamental risk to the courtโs legitimacy. Unlike elected officials, Supreme Court justices serve for life and derive their legitimacy not from votes, but from public trust. The act of writing a reasoned opinion is a judgeโs way of saying, โHereโs why you should trust me; hereโs my work.โ When the court increasingly abandons this practice, issuing barely explained rulings on major national issues, it erodes that trust. As the Supreme Courtโs public approval ratings plumb historic lows, its increasing reliance on the shadow docket to make pivotal decisions only exacerbates this critical problem, threatening the very foundation of its authority.
ํ๊ตญ์ด
โHow Intercom 2Xโd engineering velocity with Claude Code | Brian Scanlanโ โ How I AI ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
AI๊ฐ ์์ง๋์ด๋ง์ โํ์ดํ ํ๊ณโ๋ฅผ ํ๋ฌผ๋ค: ์ธํฐ์ฝค, ํด๋ก๋ ์ฝ๋๋ก ๊ฐ๋ฐ ์๋ 2๋ฐฐ ๋ฌ์ฑ ๋น๊ฒฐ
[์์ธ, ๋ํ๋ฏผ๊ตญ] ์์ง๋์ด๋ง ํ์ ์์ฐ์ฑ์ ๋ ๋ฐฐ๋ก ๋์ด์ฌ๋ฆฐ๋ค๋ ๊ฒ์ ๋๋ค์ ๊ธฐ์ ๊ธฐ์ ์ ๊ฟ์ผ ๊ฒ์ ๋๋ค. ํ์ง๋ง ๊ณ ๊ฐ ๊ด๊ณ ๊ด๋ฆฌ(CRM) ๋ฐ ๋ฉ์์ง ํ๋ซํผ ๋ถ์ผ์ ์ ๋ ์ฃผ์์ธ ์ธํฐ์ฝค(Intercom)์ ์ด ๊ฟ์ ํ์ค๋ก ๋ง๋ค์์ต๋๋ค. AI ๋๊ตฌ, ํนํ ์ค์ค๋กํฝ(Anthropic)์ ํด๋ก๋ ์ฝ๋(Claude Code)๋ฅผ ๋์ ํ ์ง ๋ถ๊ณผ ๋ช ๋ฌ ๋ง์ ์ฐ๊ตฌ ๊ฐ๋ฐ(R&D) ๋ถ์์ ํ ๋ฆฌํ์คํธ(PR, Pull Request) ์ฒ๋ฆฌ๋์ 2๋ฐฐ๋ก ๋๋ฆฌ๋ ๋๋ผ์ด ์ฑ๊ณผ๋ฅผ ๋ฌ์ฑํ ๊ฒ์ ๋๋ค.
โHow I AIโ ํ์บ์คํธ์ ์ถ์ฐํ ์ธํฐ์ฝค์ ์ ์ ์์ ์์ง๋์ด ๋ธ๋ผ์ด์ธ ์ค์บ๋ฐ(Brian Scanlan)์ AI๊ฐ ์ด๋ป๊ฒ ์์ง๋์ด๋ง์ ๊ทผ๋ณธ์ ์ธ ๋ณํ๋ฅผ ์ด๋๊ณ ์์ผ๋ฉฐ, ์ธํฐ์ฝค์ด ์ด๋ฅผ ํตํด ์ด๋ค ํ์ ์ ์ด๋ฃจ์๋์ง ์์ธํ ๊ณต์ ํ์ต๋๋ค. ์ด๋ฒ ๊ธฐ์ฌ์์๋ ์ธํฐ์ฝค์ ์ฑ๊ณต ์ฌ๋ก๋ฅผ ํตํด AI ์๋ ์์ง๋์ด๋ง ์กฐ์ง์ ๋ฏธ๋์ ๊ทธ ์ค์ง์ ์ธ ๊ตฌํ ์ ๋ต์ ์ฌ์ธต์ ์ผ๋ก ํ๊ตฌํฉ๋๋ค.
1. AI๊ฐ ๊ฐ์ ธ์จ โ์์๋ ฅ์ ์๋โ: ์์ง๋์ด๋ง์ ์๋ก์ด ์งํ
๋ธ๋ผ์ด์ธ ์ค์บ๋ฐ์ AI๊ฐ ์์ง๋์ด๋ง์ ๊ฐ์ ธ์จ ๊ฐ์ฅ ํฐ ๋ณํ ์ค ํ๋๋ก โ์์๋ ฅ์ด ์ฅ๋ฒฝ์ด์ง, ๋๊ตฌ๊ฐ ์๋๋ค(imagination is now the barrier, not the tool)โ๋ผ๋ ์ ์ ๊ผฝ์์ต๋๋ค. ๊ทธ๋ โ๋ง ๊ทธ๋๋ก ์ฝ๋๋ฅผ ํ์ดํํ๋ ๋ฅ๋ ฅ์ ๋ฌผ๋ฆฌ์ ํ๊ณ๋ฅผ AI๊ฐ ํด์ ํ๋คโ๊ณ ๊ฐ์กฐํ๋ฉฐ, ๊ฐ๋ฐ์๋ค์ด ๋ ์ด์ ๋๊ตฌ์ ๋ง์ถฐ ์์ ํ๋ ๋์ ์์ด๋์ด ๊ตฌํ์ ์ง์คํ ์ ์๊ฒ ๋์๋ค๊ณ ์ค๋ช ํ์ต๋๋ค.
์ธํฐ์ฝค์ ์ด๋ฏธ ์ ํ ์ธก๋ฉด์์ AI๋ฅผ ์ ๊ทน์ ์ผ๋ก ์์ฉํ๋ฉฐ โAI ์ฐ์ (AI-first)โ ์ ๋ต์ ํผ์ณ์์ต๋๋ค. ์ด๋ฌํ ๋ฐฐ๊ฒฝ ์์์ ์์ง๋์ด๋ง ํ ์ญ์ AI์ ์ ์ฌ๋ ฅ์ ๋ช ํํ ์ธ์งํ๊ณ ์์์ต๋๋ค. ํนํ ์ง๋ 2023๋ 11์๊ณผ 12์๊ฒฝ, ์ค์ค๋กํฝ์ ์คํธ์ค(Opus) 46 ๋ชจ๋ธ(GPT-4U ๋ชจ๋ธ์ ์ง์นญํ๋ ๊ฒ์ผ๋ก ๋ณด์)์ ์ถ์์ ํจ๊ป โ๋ชจ๋ ๊ฒ์ด ๋ณํ๋คโ๊ณ ์ค์บ๋ฐ์ ํ๊ณ ํฉ๋๋ค. ์ด ์์ ์ ๊ธฐ์ ์ผ๋ก AI ๋ชจ๋ธ์ ์ฑ๋ฅ์ด ๊ธ๊ฒฉํ ํฅ์๋๋ฉด์ ๊ฐ๋ฐ์๋ค์ AI๋ฅผ ๋จ์ํ ์๋ ์์ฑ ๋๊ตฌ๊ฐ ์๋, ์์ด๋์ด๋ฅผ ํ์ค๋ก ๋ง๋๋ ๊ฐ๋ ฅํ ํ๋ ฅ์๋ก ์ธ์ํ๊ธฐ ์์ํ์ต๋๋ค.
ํฅ๋ฏธ๋กญ๊ฒ๋, ์ค์บ๋ฐ์ ์ฐ๋ง ํด๊ฐ ๊ธฐ๊ฐ ๋์ ๋ง์ ๊ฐ๋ฐ์๋ค์ด ํด๋ก๋ ์ฝ๋๋ฅผ ํ์ฉํด ๊ฐ์ธ ํ๋ก์ ํธ๋ฅผ ์งํํ๋ฉฐ ์์ฒญ๋ ์๋๋ก ์๋ก์ด ๊ธฐ์ ์ ์ต๋ํ๋ ํ์์ ๋ชฉ๊ฒฉํ๋ค๊ณ ์ธ๊ธํ์ต๋๋ค. ์ด๋ ๊ธฐ์ ๋ค์ด ์ง์๋ค์๊ฒ ์ถฉ๋ถํ ์ ๊ธ ํด๊ฐ(PTO, Paid Time Off)๋ฅผ ์ ๊ณตํจ์ผ๋ก์จ AI ํ์ฉ ๋ฅ๋ ฅ์ ์์ฐ์ค๋ฝ๊ฒ ํฅ์์ํฌ ์ ์๋ โAI ๋ฆฌํฐ๋ฌ์ ํดํนโ ์ ๋ต์ด ๋ ์ ์์์ ์์ฌํฉ๋๋ค.
2. 2๋ฐฐ ๊ฐ์, ๊ทธ ์ด์์ ๋ชฉํ: ๋ฐ์ดํฐ๋ก ์ฆ๋ช ํ๋ค
์ธํฐ์ฝค์ CTO ๋ค๋ผ(Darra)๋ R&D ์ฒ๋ฆฌ๋ 2๋ฐฐ ์ฆ๊ฐ๋ผ๋ ๋ช ํํ ๋ชฉํ๋ฅผ ์ ์ํ๊ณ , ์ธํฐ์ฝค ํ์ ์ด๋ฅผ ๋ฌ์ฑํ๊ธฐ ์ํด AI ๋์ ์ โ์ ํโ์ฒ๋ผ ๋ค๋ฃจ์์ต๋๋ค. ์ฌ์ฉ์ ํผ๋๋ฐฑ์ ์์งํ๊ณ , ๋๊ตฌ ์ฌ์ฉ ๋ฐฉ์์ ๋ถ์ํ๋ฉฐ, ํ ๋ ๋ฉํธ๋ฆฌ(telemetry) ๋ฐ์ดํฐ๋ฅผ ํ์ฉํด ์ฑ๊ณผ๋ฅผ ์ธก์ ํ์ต๋๋ค.
ํต์ฌ ์ธก์ ์งํ: ์ธํฐ์ฝค์ โR&D ์ธ์๋น ๋ณํฉ๋ PR(merged PRs per R&D head)โ์ ์ฃผ์ ์งํ๋ก ์ผ์์ต๋๋ค. ์ด๋ ๋จ์ํ ์ํํธ์จ์ด ์์ง๋์ด๋ฟ๋ง ์๋๋ผ ์ ํ ๊ด๋ฆฌ์(PM, Product Manager), ๋์์ด๋, ๊ธฐ์ ํ๋ก์ ํธ ๊ด๋ฆฌ์(TPM, Technical Project Manager) ๋ฑ ๋ชจ๋ R&D ๊ตฌ์ฑ์์ ๊ธฐ์ฌ๋ฅผ ํฌํจํ๋ ํฌ๊ด์ ์ธ ์งํ์ ๋๋ค.
์ฑ๊ณผ ๋ฐ ๋์ :
- 2๋ฐฐ์ ์ฒ๋ฆฌ๋: 9๊ฐ์ ์ ๊ณผ ๋น๊ตํ์ฌ ์์ง๋์ด๋ง ํ์ ์ฒ๋ฆฌ๋์ด 2๋ฐฐ ์ฆ๊ฐํ์ต๋๋ค.
- CI ์์คํ ์ ๋ณํ: PR ์๊ฐ ๊ธ์ฆํ์ ๊ธฐ์กด CI(Continuous Integration) ์์คํ ์ด ๊ณผ๋ถํ๋ก โ๋ น์๋ด๋ฆดโ ์ ๋์์ต๋๋ค. ์ธํฐ์ฝค์ ๋ณ๋ชฉ ํ์์ ํด๊ฒฐํ๊ธฐ ์ํด CI ์์คํ ์ฑ๋ฅ์ ๊ฐ์ ํ๊ณ , ์ด์ ๋ ์ฝ๋ ๋ฆฌ๋ทฐ๊ฐ ์๋ก์ด ๋ณ๋ชฉ ์ง์ ์ด ๋์์ต๋๋ค.
- โ์ฌ์ธ(All-in)โ ์ ๋ต์ ํจ๊ณผ: ์ค์บ๋ฐ์ โ๋ง์ฝ ๋น์ ์ด ๋ชจ๋ ๊ฒ์ ๊ฑธ๊ณ (go all-in), ํ๊ณผ ์ฝ๋๋ฒ ์ด์ค๋ฅผ ์ค๋นํ๋ค๋ฉด, ์ ๋ฐ์ ์ธ ์ ํ ํ์ง๊ณผ ๊ฐ๋ฐ์ ๊ฒฝํ์ด ํฅ์๋ ๊ฒโ์ด๋ผ๊ณ ๊ฐ์กฐํ์ต๋๋ค. ๋์ ์ ๋ขฐ ๋ฌธํ๋ฅผ ๋ฐํ์ผ๋ก ํ์ด ์์คํ ์ ์ ์ฉํ์ง ์๊ณ ๋ชฉํ๋ฅผ ํฅํด ๋์๊ฐ๋๋ก ๋ ๋ คํ์ต๋๋ค.
- 10๋ฐฐ์ ๋น์ : 2๋ฐฐ ์ฑ๊ณผ์ ๋ง์กฑํ์ง ์๊ณ , โ์ 10๋ฐฐ๋ ์ ๋๋๊ฐ?โ๋ผ๋ ์ง๋ฌธ์ ๋์ง๋ฉฐ ์ง์์ ์ธ ๊ฐ์ ์ ์ถ๊ตฌํ๊ณ ์์ต๋๋ค.
3. ์๋์ ํจ๊ป ํฅ์๋ ํ์ง: ๊ธฐ์ ๋ถ์ฑ ํด๊ฒฐ์ ๊ธฐํ
์ผ๊ฐ์์๋ AI๋ก ์ฝ๋๋ฅผ ๋น ๋ฅด๊ฒ ์์ฑํ๋ฉด ํ์ง์ด ์ ํ๋ ๊ฒ์ด๋ผ๋ ์ฐ๋ ค๋ฅผ ์ ๊ธฐํฉ๋๋ค. ํ์ง๋ง ์ธํฐ์ฝค์ ๊ฒฝํ์ ์ ๋ฐ๋์์ต๋๋ค.
ํ์ง ์ธก์ ๋ฐ ๊ฒ์ฆ:
- ์ถ์ ์๊ฐ ๋จ์ถ: ๊ธฐ๋ฅ ๊ฐ๋ฐ ์ฒซ ์ฝ๋ ์์ฑ๋ถํฐ ๋ด์ค ์ฑ๋์ ์ ๋ฐ์ดํธ๊ฐ ๊ฒ์๋๊ธฐ๊น์ง ๊ฑธ๋ฆฌ๋ ์๊ฐ์ด ๊พธ์คํ ๋จ์ถ๋์์ต๋๋ค.
- ์คํ ํฌ๋ ์ฐ๊ตฌ ํ๋ ฅ: ์ธํฐ์ฝค์ ์คํ ํฌ๋ ์ฐ๊ตฌ ๊ทธ๋ฃน๊ณผ ํ๋ ฅํ์ฌ ์์ฒด ๋ฐ์ดํฐ๋ฅผ ์ ๊ณตํ์ต๋๋ค. ์ด ์ฐ๊ตฌ ๊ฒฐ๊ณผ, ์ฝ๋ ํ์ง์ด ์คํ๋ ค ํฅ์๋๊ณ ์์์ด ๋ฐํ์ก์ต๋๋ค.
- ์ฌ๊ณ ๊ฐ์: ์ค์บ๋ฐ์ ๋ชจ๋ ์ธ์๋ํธ(incident)์ ์ฐธ์ฌํ์ง๋ง, AI ๋์ ์ดํ ํ๋ก๋์ ๊ด๋ จ ์ฌ๊ณ ๋ ๊ณ ๊ฐ ๋ถ๋ง์ด ์ฆ๊ฐํ์ง ์์๋ค๊ณ ๋ณด๊ณ ํ์ต๋๋ค.
๊ธฐ์ ๋ถ์ฑ ํด๊ฒฐ์ ๊ธฐํ: AI๋ ๊ธฐ์ ๋ถ์ฑ(tech debt)๋ฅผ ํด๊ฒฐํ๊ณ ์ฝ๋๋ฒ ์ด์ค์ ์ ๋ฐ์ ์ธ ํ์ง์ ํฅ์์ํฌ ์ ์๋ ์ ํธ์ ๊ธฐํ๋ฅผ ์ ๊ณตํฉ๋๋ค. ๊ณผ๊ฑฐ์๋ ์ฝ๋ ํ์ง ๊ฐ์ , ๋ณด์ ๊ฐํ, ๊ฐ๋ฐ์ ๊ฒฝํ ํฅ์ ๋ฑ ๋ด๋ถ ํ๋ก์ ํธ์ ํ ๋นํ ์ ์๋ R&D ์ญ๋์ด ์ ํ์ ์ด์์ต๋๋ค. ์ด๋ฌํ ํ๋ก์ ํธ๋ ์ง์ ์ ์ธ ๋งค์ถ(ARR, Annual Recurring Revenue)์ ์ฐฝ์ถํ์ง ์๊ธฐ ๋๋ฌธ์ ๋น์ฆ๋์ค์ ์ฐ์ ์์์์ ๋ฐ๋ ค๋๊ธฐ ์ฌ์ ์ต๋๋ค.
ํ์ง๋ง AI๋ฅผ ํตํด ์ด๋ฌํ ์์ ์ โ๋น์ฉโ์ด ํฌ๊ฒ ์์ถ๋๋ฉด์, ์ด์ ๊ธฐ์ ์ ๋ด๋ถ ํฌ์์ ๋ ์ ๊ทน์ ์ผ๋ก ๋์ค ์ ์๊ฒ ๋์์ต๋๋ค. ์ค์บ๋ฐ์ CTO๋ ์์ง๋์ด๋ง VP(Vice President)๋ค์๊ฒ โ์ฝ๋๋ฒ ์ด์ค์์ ์ซ์ดํ๋ ๋ชจ๋ ๊ฒ์ ํ ๋ฌ ๋์ ์์ ํ๋ ๋ฐ ํฌ์ํ๊ณ , ์ผ๋ง๋ ๋น ๋ฅด๊ฒ ํด๊ฒฐํ ์ ์๋์ง ๋ณด๋ผโ๊ณ ์กฐ์ธํ๋ฉฐ, ์ด๊ฒ์ด ํ์ ์ฌ๊ธฐ์ ์์ฐ์ฑ ํฅ์์ ํฐ ๋์์ด ๋ ๊ฒ์ด๋ผ๊ณ ๋ง๋ถ์์ต๋๋ค.
4. AI ์์ด์ ํธ ์ค์ฌ์ ์ํํธ์จ์ด ๊ณต์ฅ
์ธํฐ์ฝค์ ๋ชจ๋ ๊ธฐ์ ์์ ์ด โ์์ด์ ํธ ์ค์ฌ(agent-first)โ์ผ๋ก ์ ํ๋ ๊ฒ์ด๋ผ๊ณ ๋ฏฟ์ต๋๋ค. ์ด๋ AI ์์ด์ ํธ๊ฐ ๊ธฐ๋ณธ์ ์ธ ์์ ์ ์ํํ๊ณ , ์ธ๊ฐ ์์ง๋์ด๋ ๋ ๋์ ์์ค์ ๋ฌธ์ ํด๊ฒฐ๊ณผ ์ฐฝ์์ ์ธ ์์ ์ ์ง์คํ๋ ๋ฏธ๋๋ฅผ ์๋ฏธํฉ๋๋ค.
โCreate PRโ ์คํฌ์ ์ฌ๋ก: ์ธํฐ์ฝค์ ํด๋ก๋ ์ฝ๋๋ฅผ ์ฌ์ฉํ๋ฉด์ ํ ๊ฐ์ง ๋ฌธ์ ์ ์ง๋ฉดํ์ต๋๋ค. AI๊ฐ ์์ฑํ๋ ํ ๋ฆฌํ์คํธ ์ค๋ช (PR description)์ด ์ฝ๋ ์์ฒด๋ง ์ค๋ช ํ ๋ฟ, PR์ โ์๋(intent)โ๋ ๊ด๋ จ ๋งฅ๋ฝ์ ์ ๋๋ก ์ ๋ฌํ์ง ๋ชปํด ํ์ง์ด ์ ํ๋๋ ๊ฒฝํฅ์ด ์์์ต๋๋ค.
- ๋ฌธ์ ํด๊ฒฐ: ์ธํฐ์ฝค์ LLM ์ฌ์ฌ๊ด(LLM judge)์ ํ์ฉํ์ฌ ์ข์ PR ์ค๋ช ์ ๊ธฐ์ค์ ์ ์ํ๊ณ , ๊ธฐ์กด PR ์ค๋ช ์ ํ์ง ์ ํ ์ถ์ธ๋ฅผ ํ์ธํ์ต๋๋ค.
- โCreate PRโ ์คํฌ ๊ฐ๋ฐ: ์ด๋ฅผ ํด๊ฒฐํ๊ธฐ ์ํด โcreate PRโ์ด๋ผ๋ ๋ด๋ถ ์คํฌ(skill)์ ๊ฐ๋ฐํ์ต๋๋ค. ์ด ์คํฌ์ ์ธ์ ์ ๋งฅ๋ฝ์ ํ์ฉํ์ฌ PR์ ์๋๋ฅผ ๋ช ํํ๊ฒ ์ค๋ช ํ๋๋ก ์์ด์ ํธ๋ฅผ ์๋ดํฉ๋๋ค.
- ํ์ค ๊ฐ์ : ์ด๊ธฐ์๋ ๊ฐ๋ฐ์๋ค์ด ์ด ์คํฌ์ ์๋ฐ์ ์ผ๋ก ์ฌ์ฉํ์ง ์์๊ธฐ์, ์ธํฐ์ฝค์ GitHub CLI๋ฅผ ํตํ PR ์์ฑ์ ์ฐจ๋จํ๊ณ ๋ฐ๋์ โcreate PRโ ์คํฌ์ ์ฌ์ฉํ๋๋ก ๊ฐ์ ํ์ต๋๋ค.
- โ์ํํธ์จ์ด ๊ณต์ฅโ ๋น์ : ์ค์บ๋ฐ์ ์ด๋ฅผ โ์ํํธ์จ์ด ๊ณต์ฅ(software factory)โ์ ๋น์ ํ๋ฉฐ, CI/CD ํ์ดํ๋ผ์ธ์ฒ๋ผ ์ฝ๋ ์์ฑ ๋จ๊ณ๋ถํฐ ํ์คํ๋ ๊ฐ๋๋ ์ผ(guardrails)์ ์ ์ฉํ ์ ์๊ฒ ๋์๋ค๊ณ ์ค๋ช ํ์ต๋๋ค. ์ด๋ ๊ฐ๋ณ ์์ ์ ์ฌ์ํด ๋ณด์ด๋ ํ์ง ๊ฐ์ ์ด ๋ชจ์ฌ ์ ์ฒด ์์คํ ์ ์ ๋ขฐ์ฑ๊ณผ ์์ธก ๊ฐ๋ฅ์ฑ์ ๋์ด๋ ๊ฒฐ๊ณผ๋ฅผ ๊ฐ์ ธ์ต๋๋ค.
๊ฒฐ๊ณผ์ ์ผ๋ก AI๋ ๋ฆฌ๋ค์ด๋ ํธ(redirect) ์ถ๊ฐ์ ๊ฐ์ ๋ฐ๋ณต์ ์ด๊ณ ์ง๋ฃจํ ์์ ์ ์์ด์ ํธ์๊ฒ ๋งก๊ธฐ๊ณ , ์ธ๊ฐ ์์ง๋์ด๋ ๊ณ ๊ฐ์ ์ํ ํ์ ์ ์ธ ๋ฌธ์ ํด๊ฒฐ์ ์ง์คํ ์ ์๋๋ก ํด์ค๋๋ค.
5. AI ํ์ฉ์ ํฌ๋ช ์ฑ๊ณผ ์ต์ ํ: ํ ๋ ๋ฉํธ๋ฆฌ
AI ๋๊ตฌ์ ์ฑ๊ณต์ ์ธ ๋์ ๊ณผ ํ์ฐ์ ์ํด์๋ โ๋์ ๊ฐ๋ฆฐ ์ฑ ๋นํํ์ง ์๋ ๊ฒ(not flying blind)โ์ด ์ค์ํฉ๋๋ค. ์ธํฐ์ฝค์ AI ํ์ฉ ํํฉ์ ํ์ ํ๊ณ ๊ฐ์ ํ๊ธฐ ์ํด ๊ฐ๋ ฅํ ํ ๋ ๋ฉํธ๋ฆฌ ์์คํ ์ ๊ตฌ์ถํ์ต๋๋ค.
- ์คํฌ ์ฌ์ฉ๋ ์ถ์ (Honeycomb): ๊ฐ ๋ด๋ถ ์คํฌ์ ์ฌ์ฉ๋ ๋ฐ์ดํฐ๋ฅผ ํ๋์ฝค(Honeycomb)์ผ๋ก ์ ์กํ์ฌ, ์ด๋ค ์คํฌ์ด ์ผ๋ง๋ ์์ฃผ ์ฌ์ฉ๋๋์ง ์ค์๊ฐ์ผ๋ก ๋ชจ๋ํฐ๋งํฉ๋๋ค. ์ด๋ฅผ ํตํด ๊ฐ๋ฐ์๋ค์ ์์ ์ด ๋ง๋ ์คํฌ์ ์ค์ ์ํฅ์ ํ์ ํ๊ณ , ํ ์ ์ฒด์ AI ํ์ฉ ํจํด์ ๋ถ์ํ ์ ์์ต๋๋ค.
- ์ธ์ ๋ฐ์ดํฐ ๋ถ์ (S3 ๋ฐ ๋ด๋ถ ๋๊ตฌ): ๋ชจ๋ ํด๋ก๋ ์ฝ๋ ์ธ์ ๋ฐ์ดํฐ(๋ํ ๊ธฐ๋ก)๋ฅผ S3์ ์ต๋ช ํํ์ฌ ์ ์ฅํฉ๋๋ค. ์ด ๋ฐ์ดํฐ๋ฅผ ๊ธฐ๋ฐ์ผ๋ก ์ธ์ ์ดํ๋ฅ , ์ ์ฉํ ๊ฒฐ๊ณผ(์: PR ์์ฑ) ๋๋ฌ ์๊ฐ ๋ฑ์ ๋ถ์ํ์ฌ AI ํ์ฉ์ ํจ๊ณผ์ฑ์ ํ๊ฐํฉ๋๋ค.
- ๊ฐ์ธ ๋ง์ถคํ ํผ๋๋ฐฑ ๋๊ตฌ: ์ธ์ ๋ฐ์ดํฐ๋ฅผ ํ์ฉํ์ฌ ๊ฐ์ธ๋ณ AI ํ์ฉ ํจํด์ ๋ํ ํผ๋๋ฐฑ์ ์ ๊ณตํ๋ ๋ด๋ถ ๋๊ตฌ๋ฅผ ๊ฐ๋ฐํ์ต๋๋ค. ์๋ฅผ ๋ค์ด, ํน์ ์คํฌ ์ฌ์ฉ์ด ๋นํจ์จ์ ์ด์๊ฑฐ๋ ๊ฐ์ ์ด ํ์ํ ๋ถ๋ถ์ ์ฌ์ฉ์์๊ฒ ์๋ ค์ค๋๋ค. ์ด๋ ์ง์๋ค์ด AI ๋๊ตฌ๋ฅผ ๋ ํจ๊ณผ์ ์ผ๋ก ํ์ฉํ๊ณ ์ค์ค๋ก ๊ฐ์ ํด ๋๊ฐ ์ ์๋๋ก ์ง์ํฉ๋๋ค.
์ด๋ฌํ ํ ๋ ๋ฉํธ๋ฆฌ ์์คํ ์ ์ด๋ค ์คํฌ์ด ๊ฐ์ฅ ํจ๊ณผ์ ์ธ์ง, ์ด๋ค ์คํฌ์ด ๊ฐ์ ์ด ํ์ํ์ง ๋ฑ ์คํ ๊ฐ๋ฅํ ํต์ฐฐ๋ ฅ์ ์ ๊ณตํ์ฌ AI ํ๋ซํผ์ ์ง์์ ์ธ ์ฑ์ฅ์ ๋์ต๋๋ค.
6. ๋ด๋ถ ์คํฌ ์ ์ฅ์ ๊ตฌ์ถ ๋ฐ ๋ฐฐํฌ
์ธํฐ์ฝค์ ๋ด๋ถ AI ์คํฌ๋ค์ ํจ์จ์ ์ผ๋ก ๊ด๋ฆฌํ๊ณ ๋ฐฐํฌํ๊ธฐ ์ํด GitHub ์ ์ฅ์(repo)๋ฅผ ํ์ฉํ๊ณ ์์ต๋๋ค.
- ์ค์ ์ง์ค์ ์ ์ฅ์: ๋ชจ๋ ํ์ด ๊ฐ์์ ํ์์ ๋ง๋ ํ๋ฌ๊ทธ์ธ๊ณผ ์คํฌ์ ๊ฐ๋ฐํ์ฌ ์ด ์ค์ ์ ์ฅ์์ ๊ธฐ์ฌํฉ๋๋ค. ํต์ฌ ํ๋ฌ๊ทธ์ธ(์: ์์ ํํฌ, PR ๋ณํฉ, ํ ๋ ๋ฉํธ๋ฆฌ ๊ด๋ จ ๊ธฐ๋ฅ)์ ์๊ฒฉํ ํ์ง ๊ด๋ฆฌ๋ฅผ ๊ฑฐ์ณ ๋ชจ๋ ์ง์์๊ฒ ๋ฐฐํฌ๋ฉ๋๋ค.
- IT ํ๊ณผ์ ํ๋ ฅ: ํด๋ก๋ ์ฝ๋์ ๊ธฐ๋ณธ ํ๋ฌ๊ทธ์ธ ๋ฐฐํฌ ๋ฉ์ปค๋์ฆ์ด ๋ถ์์ ํ๋ค๊ณ ํ๋จํ์ฌ, ์ธํฐ์ฝค์ ๋ด๋ถ IT ์์คํ ์ ํ์ฉํด ๋ชจ๋ ํ๋ฌ๊ทธ์ธ์ ์ง์๋ค์ ๋ ธํธ๋ถ ๋์คํฌ์ ์ง์ ๋๊ธฐํํฉ๋๋ค. ์ด๋ ์์ ์ ์ธ ๋ฐฐํฌ๋ฅผ ๋ณด์ฅํ๊ณ ๋๋ฒ๊น ์ ๋ณต์ก์ฑ์ ์ค์ฌ์ค๋๋ค.
7. ๋น์ฉ๊ณผ ํฌ์: ์ํ๋ฅผ ์ํ ๊ณผ๊ฐํ ์ ํ
AI ๋ชจ๋ธ ์ฌ์ฉ ๋น์ฉ์ ์ธํฐ์ฝค์ ๊ฐ๋ฐ ์๋ ๊ทธ๋ํ์ฒ๋ผ ๊ฐํ๋ฅด๊ฒ ์์นํ๊ณ ์์ต๋๋ค. ์ด๋ ๋ง์น ์๋ก์ด ์ฌ๋ฌด์ค์ ์ฌ๋ฌ ๊ฐ ์ฌ๋ ๊ฒ๊ณผ ๊ฐ์ ๋ง๋ํ ๋น์ฉ์ ๋๋ค.
- โ๋น์ฉ์ ๋์ค์ ๊ณ ๋ คโ ์ ๋ต: ์ธํฐ์ฝค์ ํ์ฌ โ๋ชจ๋ ๊ฒ์ ์คํธ์ค ๋ชจ๋ธ์ ์ฌ์ฉํ๊ณ , ๋น์ฉ์ ๋์ค์ ์ ๊ฒฝ ์ฐ์(care about the bill later)โ๋ ๊ณผ๊ฐํ ์ ๋ต์ ์ทจํ๊ณ ์์ต๋๋ค. ๊ทธ ์ด์ ๋ ํ์ฌ AI๋ฅผ ํตํด ์ป๋ โ์ํ(alpha)โ ๋๋ โ์ด์ โ์ด ๋น์ฉ์ ํจ์ฌ ์ํํ๋ค๊ณ ํ๋จํ๊ธฐ ๋๋ฌธ์ ๋๋ค.
- ์ฅ๊ธฐ์ ํฌ์ ๊ด์ : ์ธํฐ์ฝค์ AI ํ๋ซํผ์ ๋ํ ํฌ์๋ฅผ ์ฅ๊ธฐ์ ์ธ ๊ด์ ์์ ์ ๊ทผํ๊ณ ์์ต๋๋ค. ๋น์ฅ์ ๋น์ฉ ์ต์ ํ๋ณด๋ค๋ ์ต๋ํ ๋น ๋ฅด๊ฒ AI์ ์ ์ฌ๋ ฅ์ ํ์ฉํ์ฌ ํ์ ์ ์ธ ์ฑ๊ณผ๋ฅผ ๋ด๋ ๋ฐ ์ง์คํ๊ณ , ์ถฉ๋ถํ ์ด์ ์ ํ๋ณดํ ํ์ ๋น์ฉ ํจ์จํ ๋จ๊ณ๋ฅผ ๋ฐ์ ๊ณํ์ ๋๋ค.
๋ฌผ๋ก ์ด๋ฌํ ์ ๊ทผ ๋ฐฉ์์ด ๋ชจ๋ ๊ธฐ์ ์ ์ ํฉํ ๊ฒ์ ์๋๋๋ค. ํ์ง๋ง ์ธํฐ์ฝค์ ์ ๋ ์ฃผ์๋ก์ AI๊ฐ ๊ฐ์ ธ์ฌ ๋ฏธ๋ ๊ฐ์น์ ๋ํ ํ๊ณ ํ ๋ฏฟ์์ ๋ฐํ์ผ๋ก ๊ณผ๊ฐํ ํฌ์๋ฅผ ๋จํํ๊ณ ์์ต๋๋ค.
๊ฒฐ๋ก : AI๊ฐ ์ด๋๋ ์์ง๋์ด๋ง์ ์๋ก์ด ์๋
์ธํฐ์ฝค์ ์ฌ๋ก๋ AI๊ฐ ๋จ์ํ ์์ฐ์ฑ ํฅ์์ ๋์ด, ์์ง๋์ด๋ง ์กฐ์ง์ ๋ณธ์ง๊ณผ ์๋ ๋ฐฉ์์ ๊ทผ๋ณธ์ ์ผ๋ก ๋ณํ์ํฌ ์ ์์์ ๋ช ํํ ๋ณด์ฌ์ค๋๋ค. 2๋ฐฐ์ ๊ฐ๋ฐ ์๋ ํฅ์, ์ฝ๋ ํ์ง ๊ฐ์ , ๊ธฐ์ ๋ถ์ฑ ํด๊ฒฐ, ๊ทธ๋ฆฌ๊ณ โ์์ด์ ํธ ์ค์ฌโ์ ์ํํธ์จ์ด ๊ณต์ฅ์ผ๋ก์ ์ ํ์ AI ์๋ ์์ง๋์ด๋ง์ ๋ฏธ๋๋ฅผ ์ฟ๋ณผ ์ ์๋ ์ค์ํ ์ฒญ์ฌ์ง์ ๋๋ค.
๋ธ๋ผ์ด์ธ ์ค์บ๋ฐ์ โAI์ ์๋ ์ฆ๊ฐ๋ ๋ชจ๋ ๊ฐ์ ๊ณผ ์ฝ์ ์ ์ฆํญ์ํจ๋คโ๊ณ ๋งํฉ๋๋ค. ์ธํฐ์ฝค์ ๊ฒฌ๊ณ ํ ์ํํธ์จ์ด ๊ฐ๋ฐ ์ญ๋๊ณผ ๋ณํ์ ๋ํ ๋์ ๊ฐ๋ฐฉ์ฑ์ ๋ฐํ์ผ๋ก AI์ ๊ฐ์ ์ ๊ทน๋ํํ ์ ์์์ต๋๋ค. ์ด์ ๊ธฐ์ ๋ค์ AI๋ฅผ ํตํด ์ธ๊ฐ ์์ง๋์ด๊ฐ ๋์ฑ ์ฐฝ์์ ์ด๊ณ ์ ๋ต์ ์ธ ๋ฌธ์ ํด๊ฒฐ์ ์ง์คํ ์ ์๋๋ก ์ง์ํ๋ฉฐ, ๊ถ๊ทน์ ์ผ๋ก ๋ ๋์ ํ์ง์ ์ ํ์ ๋ ๋น ๋ฅด๊ฒ ์์ฅ์ ์ ๋ณด์ผ ์ ์๋ ์๋ก์ด ์๋์ ์ง๋ฉดํด ์์ต๋๋ค. ์ธํฐ์ฝค์ ์ด์ผ๊ธฐ๋ ๋ณํ๋ฅผ ๋๋ ค์ํ์ง ์๊ณ ๊ณผ๊ฐํ๊ฒ ํฌ์ํ๋ฉฐ AI๋ฅผ ์กฐ์ง์ ํต์ฌ ๋๋ ฅ์ผ๋ก ์ผ๋ ๊ธฐ์ ๋ค์ด ์ป๊ฒ ๋ ๋๋ผ์ด ์ด์ ์ ์์ฌํฉ๋๋ค.
โInside the Five Days That Remade the Supreme Courtโ โ New York Times Podcasts ๊ธฐ๋ฐ ๊ธฐ์ฌ ์๋ณธ ์์ ๋ณด๊ธฐ
๋๋ฒ์ โ๊ทธ๋ฆผ์ ์ฌ์โ์ ์ด๋์ด ํ์: ๋น๋ฐ ๋ฌธ์๊ฐ ๋๋ฌ๋ธ 5์ผ๊ฐ์ ๊ฒฉ๋
๋ฏธ๊ตญ ๋๋ฒ์์ ์ง๋ 10๋ ๊ฐ ์ด๋ฏผ๋ถํฐ ๋ํต๋ น ๊ถํ์ ์ด๋ฅด๊ธฐ๊น์ง ์ฃผ์ ํ๊ฒฐ๋ค์ ์๋๋ฅด๊ณ ์๋ฐํ๊ฒ ์ฒ๋ฆฌํ๋ ์์คํ ์ ์์กดํด์์ต๋๋ค. ์ด ์์คํ ์ โ๊ทธ๋ฆผ์ ์ฌ์(shadow docket)โ๋ผ๊ณ ๋ถ๋ฆฌ๋ฉฐ, ๊ณต๊ฐ์ ์ธ ๋ ผ์์ ์ฌ๋ ์๋ ์ฌ์ ๊ณผ์ ์ ๊ฑด๋๋ฐ๊ณ ์ ์ํ๊ฒ ๊ฒฐ์ ์ ๋ด๋ฆฌ๋ ๋ฐฉ์์ผ๋ก ์๋ํฉ๋๋ค. ๋ด์ํ์์ค(New York Times)์ ์ฌ์ธต ๋ณด๋์ ๋ฐ๋ฅด๋ฉด, ์ด ๊ทธ๋ฆผ์ ์ฌ์๊ฐ ์์๋ ์ ํํ ์๊ฐ๊ณผ ๊ทธ ๋ฐฐ๊ฒฝ์ด ๊ธฐ๋ฐ ํด์ ๋ ๋๋ฒ๊ด๋ค์ ์์ ์ ํตํด ์ฒ์์ผ๋ก ์ธ์์ ๋๋ฌ๋ฌ์ต๋๋ค.
์กฐ๋ ์บ๋(Jodie Caner)์ ์ ๋ค ๋ฆฝํ(Adam Liptac) ๊ธฐ์๋ ์ด ์ถฉ๊ฒฉ์ ์ธ ๋ฌธ์๋ฅผ ํตํด ๋๋ฒ์์ ์ด์ ๋ฐฉ์์ ์ก๋๋ฆฌ์งธ ๋ฐ๊พผ 5์ผ๊ฐ์ ๋ด๋ถ ๋ ผ์์ ํํค์ณค์ต๋๋ค. ์ด๋ ๋จ์ํ ์ ์ฐจ์ ๋ณํ๊ฐ ์๋๋ผ, ์ฌ๋ฒ๋ถ์ ์ ๋น์ฑ๊ณผ ๊ณต์ ์ฑ์ ๋ํ ๊ทผ๋ณธ์ ์ธ ์ง๋ฌธ์ ๋์ง๋ ์ฌ๊ฑด์ผ๋ก ๊ธฐ๋ก๋ ๊ฒ์ ๋๋ค.
โ๊ทธ๋ฆผ์ ์ฌ์โ๋ ๋ฌด์์ธ๊ฐ? ๋๋ฒ์์ ๋ ์ผ๊ตด
๋๋ฒ์์ ํ๊ฒฐ์ ๊ตญ๊ฐ ์ ์ฒด๋ฅผ ๊ตฌ์ํ๋ ๋ฒ์ด๋ฉฐ ํ๊ธ ๋ฒ์์ ์ง์นจ์ ์ ๊ณตํ๊ณ ์์ก ๋น์ฌ์๋ค์๊ฒ ํ๋ ๋ฐฉ์นจ์ ์ ์ํฉ๋๋ค. ์ผ๋ฐ์ ์ผ๋ก ๋๋ฒ์์ด ์ฌ๊ฑด์ ์ฒ๋ฆฌํ๋ ๋ฐฉ์์ โ์ ์ ์ฌ์(merits docket)โ๋ผ๊ณ ๋ถ๋ฆฝ๋๋ค. ์ด๋ ์ฐ๋ฆฌ๊ฐ ํํ ์๋ ๋๋ฒ์์ ๋ชจ์ต์ ๋๋ค. ๋๋ฒ๊ด๋ค์ ์ด๋ค ์ฌ๊ฑด์ ์ฌ๋ฆฌํ ์ง ์ ์คํ๊ฒ ๊ฒฐ์ ํ๊ณ , ์๋ง์ ๋ฒ๋ฅ ์๋ฅ(briefs)๋ฅผ ๊ฒํ ํฉ๋๋ค. ์ฌ๊ฑด ์ฌ๋ฆฌ๋ฅผ ๊ฒฐ์ ํ๋ฉด ๋๋ค์ ์ฌ๋ฌ ์ฐจ๋ก ์๋ฅ๋ฅผ ์ ์ถ๋ฐ๊ณ , ๊ตฌ๋ ๋ณ๋ก (oral arguments)์ ๋ฃ์ต๋๋ค. ์ดํ ๋๋ฒ๊ด๋ค์ ํจ๊ป ์์ ํ ๋ก ํ๊ณ ํฌํํ๋ฉฐ, 5๋ฒ, 10๋ฒ, ์ฌ์ง์ด 15๋ฒ๊น์ง ํ๊ฒฐ๋ฌธ ์ด๊ณ ์ ๋ณด์ถฉ ์๊ฒฌ, ๋ฐ๋ ์๊ฒฌ์ ์ฃผ๊ณ ๋ฐ์ต๋๋ค. ์ด ๋ชจ๋ ๊ณผ์ ์ด 1๋ ์ฌ์ ๊ฑธ์ณ ์งํ๋ ํ, ์์ญ์์ ๋ฐฑ ํ์ด์ง์ ๋ฌํ๋ ๋ ผ๋ฆฌ์ ์ธ ํ๊ฒฐ๋ฌธ์ด ๋ฐํ๋ฉ๋๋ค. ์ด๋ ์ง๋ํ ์ฃผ์์ ์๊ณ ์ ์ฐ๋ฌผ์ ๋๋ค.
๋ฐ๋ฉด, โ๊ทธ๋ฆผ์ ์ฌ์โ๋ ์ด ๋ชจ๋ ๊ณผ์ ์ ์๋ตํฉ๋๋ค. ๋งค์ฐ ์งง์ ๊ธฐ๊ฐ ์์ ๊ฐ๋ตํ ์๋ฅ๋ง์ผ๋ก ์งํ๋๋ฉฐ, ๊ตฌ๋ ๋ณ๋ก ์ด๋ ๋๋ฉด ์ฌ์๋ ์์ต๋๋ค. ๊ทธ ๊ฒฐ๊ณผ๋ ๊ทผ๊ฑฐ๊ฐ ๊ฑฐ์ ์๊ฑฐ๋ ์์ ์๋ ํ๊ฒฐ๋ก ์ด์ด์ง๋๋ค. ์ง๋ 10๋ ๊ฐ ๊ทธ๋ฆผ์ ์ฌ์๋ ๋๋ฒ์ ์ ๋ฌด์ ์ฃผ์ ๋ถ๋ถ์ด ๋์์ผ๋ฉฐ, ํนํ ํธ๋ผํ ๋ํต๋ น์๊ฒ ๋ง๋ํ ์ฌ๋๊ถ๊ณผ ๊ถ๋ ฅ์ ๋ถ์ฌํ ๊ฒฐ์ ๋ค์ด ํฌํจ๋์ด ์์ต๋๋ค.
๊ทธ๋ฆผ์ ์ฌ์์ ๋ํ ๋นํ๊ณผ ์นํธ๋ก :
๋นํ๋ก ์๋ค์ ๊ทธ๋ฆผ์ ์ฌ์๊ฐ ๋๋ฌด ๋น ๋ฅด๊ฒ ์งํ๋์ด ์ถฉ๋ถํ ์ ๋ณด ์์ด ๊ฒฐ์ ์ด ๋ด๋ ค์ง๋ฉฐ, ํ๊ฒฐ์ ์ง์ด ๋จ์ด์ง ์ ์๋ค๊ณ ์ฐ๋ คํฉ๋๋ค. ๋ํ, ๋์ค๊ณผ์ ์ํต ๋ถ์กฑ์ผ๋ก ์ธํด ๋๋ฒ์์ด โ์ง๊ฐ, ๊ฐ์ธ์ ์ธ ๊ฐ์ , ๋นํ์ ๋ณธ๋ฅโ์ ๋ฐ๋ผ ์ฑ๊ธํ๊ฒ ํ๊ฒฐํ๋ ๊ฒ์ด ์๋์ง ์๋ฌธ์ ์ ๊ธฐํฉ๋๋ค. ๋๋ฆฌ๊ณ ์ ์คํ๋ฉฐ ์ฌ๋ ค ๊น์ ์ฌ๋ฒ ์์คํ ์ด ํผํ๊ณ ์ ํ๋ ๋ชจ๋ ์์๊ฐ ๊ทธ๋ฆผ์ ์ฌ์์์๋ ๊ทธ๋๋ก ๋๋ฌ๋ ์ ์๋ค๋ ๊ฒ์ ๋๋ค.
๋ฐ๋ฉด, ์นํธ๋ก ์๋ค์ ์ด๋ฌํ ์ข ๋ฅ์ ๋ช ๋ น์ด ์ค๋ ์ญ์ฌ๋ฅผ ๊ฐ์ง๊ณ ์๋ค๊ณ ์ฃผ์ฅํฉ๋๋ค. ์ฌํ ์งํ์ด ์๋ฐํ๊ฑฐ๋ ์ ๊ฑฐ ๋ถ์์ฒ๋ผ ์ ์ํ ๊ฒฐ์ ์ด ํ์ํ ๊ฒฝ์ฐ, ๋๋ฒ์์ ์ข ์ข ๊ธด๊ธํ๊ฒ ๊ฐ์ ํด์๋ค๋ ๊ฒ์ ๋๋ค. ๋ํ, ๊ทธ๋ฆผ์ ์ฌ์์ ๊ฒฐ์ ์ โ์ผ์์ ์ธ ๊ฒโ์ด๋ฉฐ, ์ฌ๊ฑด์ด ํ๊ธ ๋ฒ์์ ๊ฑฐ์ณ ๋๋ฒ์์ผ๋ก ๋ค์ ๋์์ฌ ๋๊น์ง ํ์ ์ ์ง๋ฅผ ์ํ ์กฐ์น์ผ ๋ฟ์ด๋ผ๊ณ ๋งํฉ๋๋ค.
ํ์ง๋ง ์ค์ ๋ก๋ ์ด๋ฌํ ๊ฒฐ์ ๋ค์ด ์ข ์ข โ์ฌ์ค์ ์ต์ข ์ ์ธ(conclusive)โ ๊ฒฐ๊ณผ๋ฅผ ๋ณ์ต๋๋ค. ์์ญ๋ง ๋ช ์ ์ถ๋ฐฉํ๊ฑฐ๋ ์์ฒ ๋ช ์ ํด๊ณ ํ๋ ๊ฒฐ์ ์ ๋์ค์ ๋๋ฒ์์ด ๋ฒ๋ณตํ๋๋ผ๋ ๋๋๋ฆฌ๊ธฐ ๋งค์ฐ ์ด๋ ต์ต๋๋ค. ์ผ์์ ์ด๋ผ๊ณ ํ์ง๋ง ๊ทธ ์ํฅ๋ ฅ์ ๋ง๋ํ๋ฉฐ, ๊ฒฐ๊ตญ ์ฌ๊ฑด์ ์ค์ง์ ์ธ ํด๊ฒฐ๋ก ์ด์ด์ง๋ ๊ฒฝ์ฐ๊ฐ ๋ง์ต๋๋ค.
2016๋ , ๋๋ฒ์์ ๋คํ๋ 5์ผ: โ์ฒญ์ ์ ๋ ฅ ๊ณํโ ์ฌ๊ฑด
๋ด์ํ์์ค๊ฐ ์ ์ํ ๋๋ฒ๊ด๋ค ๊ฐ์ 16ํ์ด์ง์ ๋ฌํ๋ ๊ธฐ๋ฐ ์์ ์ ๊ทธ๋ฆผ์ ์ฌ์์ ๊ธฐ์์ ์ ํํ ๋ณด์ฌ์ค๋๋ค. ์ด ์์ ์ ํตํด ์ฐ๋ฆฌ๋ ๋๋ฒ๊ด๋ค์ด ์ฌ๋ฒ ์ ์ฐจ์ ์ค๋ ๊ท๋ฒ์ ํฌ๊ธฐํ๊ณ ์๋ก์ด ์ ๋ฌด ๋ฐฉ์์ ์ฑํํ๋ ์๊ฐ์ ์ฟ๋ณผ ์ ์์ต๋๋ค.
์ฌ๊ฑด์ ๋ฐ๋จ์ 2016๋ ์ค๋ฐ๋ง ํ์ ๋ถ์ โ์ฒญ์ ์ ๋ ฅ ๊ณํ(Clean Power Plan)โ์ด์์ต๋๋ค. ์ค๋ฐ๋ง ๋ํต๋ น์ ๋ ๋ฒ์งธ ์๊ธฐ ๋ง์ ๊ธฐํ ๋ณํ์ ๋์ํ๊ธฐ ์ํด ์ํ๋ฅผ ํตํ ์ ๋ฒ์ ์คํจํ์ ํ๊ฒฝ๋ณดํธ์ฒญ(EPA)์ ๊ท์ ๋ฅผ ์ง์ํ์ต๋๋ค. ์ด ๊ท์ ๋ ๋ฏธ๊ตญ์ ์ ๋ ฅ ์์คํ ์ ์ํ์์ ๋ฒ์ด๋ ์ฒญ์ ์๋์ง์์ผ๋ก ์ ํํ๋ ๊ฒ์ ๋ชฉํ๋ก ํ์ต๋๋ค. ๋น์ฐํ ์ฐ์ ๊ณ์ ๊ณตํ๋น ์ฑํฅ์ ์ฃผ๋ค์ ์ด์ ๋ฐ๋ฐํ๊ณ , ๋ฒ์์ ๊ณํ์ ์ค๋จ์ ์์ฒญํ์ต๋๋ค.
์ด๋ค์ DC ํญ์๋ฒ์(DC Circuit)์ ๊ณํ์ ์๋ฒ์ฑ์ ํ๋จํด๋ฌ๋ผ๊ณ ์์ฒญํ๋ ๋์์, ์์ก์ด ์งํ๋๋ ๋์ ๊ณํ์ ์ํ์ โ์ค๋จ(halt)โํด๋ฌ๋ผ๊ณ ์๊ตฌํ์ต๋๋ค. ํญ์๋ฒ์์ ์๋ฒ์ฑ ์ฌ๋ฆฌ๋ ์ ์ํ๊ฒ ์งํํ๊ฒ ์ง๋ง, ๊ณํ์ ์ํ์ ์ค๋จํ์ง๋ ์๊ฒ ๋ค๊ณ ๊ฒฐ์ ํ์ต๋๋ค.
๊ทธ๋ฌ์ ์ฒญ์ ์ ๋ ฅ ๊ณํ์ ๋ฐ๋ํ๋ ์ธก์ ์ ๋ก ์๋ ํ๋์ ์ทจํ์ต๋๋ค. ํ๊ธ ๋ฒ์์ด ์์ง ๊ณํ์ ํฉ๋ฒ์ฑ์ ๊ฒํ ํ๊ธฐ๋ ์ ์ ๋๋ฒ์์ ์ง์ ๊ณํ์ ์ํ์ ์ค๋จํด๋ฌ๋ผ๊ณ ์์ฒญํ ๊ฒ์ ๋๋ค. ์ด ์์ฒญ์ ๋น์ ๋๋ฒ์ ๊ด๊ณ์๋ค ์ฌ์ด์์ ์ ๋ก ์๋ ์ผ๋ก ๋ฐ์๋ค์ฌ์ก์ต๋๋ค.
๋น์ ๋๋ฒ์์ 5๋ 4๋ก ๋๋์ด ๊ณตํ๋น์ด ์๋ช ํ ๋๋ฒ๊ด๋ค์ด ๋ค์๋ฅผ ์ฐจ์งํ์ง๋ง, ๋ถ์๊ธฐ๋ ์ง๊ธ๊ณผ๋ ์ฌ๋ญ ๋ฌ๋์ต๋๋ค. ์ค์๋ ์ผ๋ค๋ ๋๋ฒ๊ด(Justice Anthony Kennedy)์ ๊ณตํ๋น์ด ์๋ช ํ์ง๋ง, โ์ค๋ ๊ฐ๋ฅํ ์ธ๋ฌผโ๋ก ๋ถ๋ฆฌ๋ ์ค์ ๋ณดํฐ(swing vote)์์ต๋๋ค. ๊ทธ๋ ๋์ฑ ๊ฒฐํผ ํฉ๋ฒํ ํ๊ฒฐ์์ ๋ค์ ์๊ฒฌ์ ์ฐธ์ฌํ๋ ๋ฑ ์์ธก ๋ถ๊ฐ๋ฅํ ์ธ๋ฌผ์ด์์ต๋๋ค. ์ด๋ฐ ๋๋ฒ์์ ์ ๋ก ์๋ ๊ธด๊ธ ์์ฒญ์ด ๋์ฐฉํ ๊ฒ์ ๋๋ค.
๋น๋ฐ ๋ฉ๋ชจ๊ฐ ๋๋ฌ๋ธ 5์ผ๊ฐ์ ๊ฒฉ๋ก
์ด ๊ธด๊ธ ์์ฒญ์ DC ํญ์๋ฒ์์ ๊ดํ ํ๋ ์กด ๋ก๋ฒ์ธ ๋๋ฒ์์ฅ(Chief Justice John Roberts)์ ์ฌ๋ฌด์ค์ ๋์ฐฉํ์ต๋๋ค. ๋ณํธ์ฌ๋ค์ ๋ก๋ฒ์ธ ๋๋ฒ์์ฅ์ด ์ด ์์ฒญ์ ๊ธฐ๊ฐํ๊ฑฐ๋ ๋๋ฃ ๋๋ฒ๊ด๋ค๊ณผ ์์ ํ ๊ธฐ๊ฐํ ๊ฒ์ด๋ผ๊ณ ์์ํ์ต๋๋ค. ๊ทธ๋ฌ๋ ๋ก๋ฒ์ธ ๋๋ฒ์์ฅ์ ๋์ ๋ค๋ฅธ ๋๋ฒ๊ด๋ค์๊ฒ ์ด ์์ฒญ์ โ๋งค์ฐ ์ง์งํ๊ฒ ๋ฐ์๋ค์ฌ์ผ ํ๋คโ๋ ๋ฉ๋ชจ๋ฅผ ์์ฑํ๊ธฐ ์์ํ์ต๋๋ค. ์ด๊ฒ์ด ๋๋ฒ์์ ์ ์ธ๊ธฐ ๊ดํ์ ์ฌ์ ๋นํ๋ 5์ผ๊ฐ์ โ๋จ๊ฑฐ๋ฆฌ ๊ฒฝ์ฃผโ์ ์์์ด์์ต๋๋ค.
-
๋ก๋ฒ์ธ ๋๋ฒ์์ฅ์ ์ฒซ ํฌ๋ฌธ (2016๋ 2์ 4์ผ): ๋ก๋ฒ์ธ ๋๋ฒ์์ฅ์ ์ธ ํ์ด์ง์ง๋ฆฌ ๋ฉ๋ชจ์์ ์ค๋ฐ๋ง ํ์ ๋ถ์ ๊ณํ์ด ์ค๋จ๋์ด์ผ ํ๋ค๊ณ ์ฃผ์ฅํ์ต๋๋ค. ๊ทธ๋ ์ด ๊ณํ์ด ์ฃผ ์ ๋ถ์ ์ํ ์ฐ์ ์ ๋ง๋ํ ๋ถ๋ด์ ์ค ๊ฒ์ด๋ฉฐ, ๊ท์ ์ค์๋ฅผ ์ํ ์๊ฐ์ด ์ด๋ฐํ๋ค๊ณ ๊ฐ์กฐํ์ต๋๋ค. ๋ํ, ์ค๋ฐ๋ง ๋ํต๋ น์ด ์์งํ๋ โ์ฒญ์ ๋๊ธฐ๋ฒ(Clean Air Act)โ์ด ๊ทธ๋ฌํ ๊ถํ์ ๋ช ํํ๊ฒ ๋ถ์ฌํ์ง ์์ผ๋ฉฐ, ์ด๋ โ์ฃผ์ ์ง๋ฌธ ์์น(major questions doctrine)โ์ ์๋ฐฐ๋ ์ ์๋ค๊ณ ์ธ๊ธํ์ต๋๋ค. ๋ก๋ฒ์ธ ๋๋ฒ์์ฅ์ EPA๊ฐ ๋ถ๊ณผ ๋ช ๋ฌ ์ ์์ ๋ฐฐ์ถ๋ ๊ท์ ์ ๊ด๋ จํ์ฌ ๋๋ฒ์์ โ์์๋คโ๊ณ ๋๋ผ๋ฉฐ, โ์ด๋ฒ์๋ ๋ค์ ๊ทธ๋ฐ ์ผ์ด ์ผ์ด๋์ง ์๋๋ก ํ๊ฒ ๋คโ๋ ๊ฐํ ๋ถ์พ๊ฐ์ ๋๋ฌ๋์ต๋๋ค. ๊ทธ๋ ์ ํต์ ์ธ ๋๋ฆฐ ๋ฒ์ ์ ์ฐจ๊ฐ ์ค๋ฐ๋ง ํ์ ๋ถ์ โ๊ตํํโ ๊ท์ ๋ฅผ ํ์ฉํ๊ณ ์๋ค๊ณ ๋ณธ ๊ฒ์ ๋๋ค.
-
๋ธ๋ผ์ด์ด ๋๋ฒ๊ด์ ๋ฐ๋ฐ (2์ 4์ผ): ๊ฐ์ ๋ , ์คํฐ๋ธ ๋ธ๋ผ์ด์ด ๋๋ฒ๊ด(Justice Stephen Breyer)์ ๋ก๋ฒ์ธ ๋๋ฒ์์ฅ์ ๋ฉ๋ชจ์ ๋ฐ๋ฐํ๋ ์๊ฒฌ์ ์ ์ํ์ต๋๋ค. ๊ทธ๋ ์ฒญ์ ์ ๋ ฅ ๊ณํ์ด ๋น์ฅ ์ฐ์ ๊ณ์ ์๋ฌด๊ฒ๋ ์๊ตฌํ์ง ์์ผ๋ฉฐ, ์ ๋ฉด์ ์ธ ๊ท์ ์ค์๋ 2030๋ ๊น์ง ํ์ ์์ผ๋ฏ๋ก ์๋๋ฅผ ์ด์ ๊ฐ ์๋ค๊ณ ์ง์ ํ์ต๋๋ค. ๋ํ, ํญ์๋ฒ์์ด ์์ง ํฉ๋ฒ์ฑ์ ๊ฒํ ํ๋ ๋์ ๋๋ฒ์์ด ๊ฐ์ ํ๋ ๊ฒ์ โ๋งค์ฐ ์ด๋ก์ ์ธ ์ผโ์ด๋ผ๊ณ ๊ฐ์กฐํ๋ฉฐ, ํ๊ธ ๋ฒ์์ด ๋จผ์ ์ฒ๋ฆฌํ๋๋ก ๋์ด์ผ ํ๋ค๊ณ ์ฃผ์ฅํ์ต๋๋ค.
-
๋ก๋ฒ์ธ ๋๋ฒ์์ฅ์ ์ฌ๋ฐ๋ฐ (2์ 5์ผ): ๋ก๋ฒ์ธ ๋๋ฒ์์ฅ์ ๋ค์ ๋ ๊ณง๋ฐ๋ก ์ฌ๋ฐ๋ฐ ๋ฉ๋ชจ๋ฅผ ๋ณด๋๊ณ , ๊ทธ์ ์ด์กฐ๋ ๋์ฑ ์ง์ฆ์ค๋ฝ๊ณ ๋จํธํด์ก์ต๋๋ค. ๊ทธ๋ ๋ธ๋ผ์ด์ด ๋๋ฒ๊ด์ ์ ์ฐจ์ ์ฐ๋ ค๋ฅผ ์ผ์ถํ๋ฉฐ, โ์ด ๊ท์ ๋ ์ ๋ ฅ ๋ถ๋ฌธ์ ๋ถ๊ณผ๋ ๊ฐ์ฅ ๋น์ผ ๊ท์ ๋ก ๋ฌ์ฌ๋๋คโ๋ฉฐ ์ฆ๊ฐ์ ์ธ ์กฐ์น์ ํ์์ฑ์ ์ฃผ์ฅํ์ต๋๋ค. ๊ทธ๋ ๋๋ฒ์์ด ๊ฒฐ๊ตญ ์ด ๊ณํ์ ๊ฒํ ํ์ฌ ๋ฌดํจํํ ๊ฒ์ด๋ผ๋ ํ์ ์ ๊ฐ์ง๊ณ ์์์ผ๋ฉฐ, โ๋ฒ์์ด ๊ถ๊ทน์ ์ผ๋ก ์ด ์ฌ๊ฑด์ ์ค์ง์ ์ผ๋ก ์ฌ๋ฆฌํ๋ค๋ฉด, ์ด ๋ฒ์ ์ด์๋จ์ ๊ฐ๋ฅ์ฑ์ด ๋งค์ฐ ๋ฎ๋ค๊ณ ๋ณธ๋คโ๊ณ ๋ช ์ํ์ต๋๋ค. ์ฆ, ๊ทธ๋ ๋๋ฒ์์ ์ ์์ ์ธ ์ฌ์ ์ ์ฐจ๋ฅผ ๊ฑฐ์น์ง ์๊ณ ๋ ์ด๋ฏธ ๋ง์์ ์ ํ ์ํ์์ต๋๋ค.
-
์ผ์ด๊ฑด ๋๋ฒ๊ด์ ๊ฒฝ๊ณ (2์ 8์ผ): ๋ฏผ์ฃผ๋น์ด ์๋ช ํ ์๋ ๋ ์ผ์ด๊ฑด ๋๋ฒ๊ด(Justice Elena Kagan)์ ๋ก๋ฒ์ธ ๋๋ฒ์์ฅ์ ์ ์ฅ์ ๋ฐ๋ํ๋ ๋ฉ๋ชจ๋ฅผ ์ ์ถํ๋ฉฐ ๊ฒฝ๊ณ ์์ ์ธ๋ ธ์ต๋๋ค. ๊ทธ๋ ๋ ๋๋ฒ์์ฅ์ด ํ๋ ค๋ ์ผ์ด โ์ ๋ก ์๋(unprecedented)โ ์ผ์ด๋ผ๋ฉฐ โ์ง์ ํ ์ฐ๋ ค(real pause)โ๋ฅผ ํํ์ต๋๋ค. ๋ํ, ์ด ์ฌ๊ฑด์ โ๋ณต์กํ ๋ฒ๊ท ๋ฐ ๊ท์ ์ฒด์ ์ ๊ด๋ จ๋ ์ด๋ ค์ด ์ฌ๊ฑดโ์ด๋ฏ๋ก ๋ ๋ง์ ์๊ฐ์ด ํ์ํ๋ค๊ณ ์ฃผ์ฅํ์ต๋๋ค.
-
์๋ฆฌํ ๋๋ฒ๊ด์ ์ง์ง (2์ 8์ผ): ๊ณตํ๋น์ด ์๋ช ํ ์๋ฎค์ผ ์๋ฆฌํ ๋๋ฒ๊ด(Justice Samuel Alito)์ ๋ก๋ฒ์ธ ๋๋ฒ์์ฅ์ ์ง์งํ๋ฉฐ, ๋ฒ์์ด ํ๋ํ์ง ์์ผ๋ฉด โ์ฌ๋ฒ๋ถ์ ์ ๋น์ฑ(institutional legitimacy)์ด ๋ฌดํจํ๋ ์ ์๋คโ๊ณ ์ฃผ์ฅํ์ต๋๋ค. ๊ทธ๋ ์ค๋ฐ๋ง ํ์ ๋ถ๊ฐ ๋๋ฒ์์ ๋ฌด๋ ฅํ์ํค๋ ค ํ๋ค๊ณ ๋๋ผ๋ ๋ก๋ฒ์ธ ๋๋ฒ์์ฅ์ โ๋ชจ์๊ฐโ์ ๊ณต๊ฐํ์ต๋๋ค.
์ด๋ฌํ ๋ฉ๋ชจ๋ค์ ๋๋ฒ๊ด๋ค์ด ๋ฒ์ ์ธ ๋ ผ๋ฆฌ๋ณด๋ค๋ ๋นํ์ ๋ถ๋ง, ๋๋ฒ์์ ์์์ ๋ํ ์ฐ๋ ค ๋ฑ โ๋ฒ๊ณผ ๋ฌด๊ดํโ ์์๋ค์ ์ผ๋ง๋ ๊น์ด ์ํฅ์ ๋ฐ๋์ง ๋ณด์ฌ์ค๋๋ค. ๊ทธ๋ค์ ์ผ๋ฐ์ ์ธ ํ๊ฒฐ๋ฌธ์ด ์ฌ์ค๊ณผ ๊ด๋ จ ๋ฒ๋ฅ ์๋ฃ์ ๊ตญํ๋๋ โ๊ฐ๋ฉดโ์ด๋ผ๋ฉด, ์ด ๋น๋ฐ ์์ ๋ค์ ๋๋ฒ๊ด๋ค์ ์ค์ ์์ ๋ฅผ ์์งํ๊ฒ ๋๋ฌ๋ด๋ ์ฐฝ๋ฌธ์ด์์ต๋๋ค.
์ค์ ๋ณดํฐ์ ๊ฒฐ์ ๊ณผ ๊ทธ ์ดํ
๊ฒฐ๊ตญ, ๋ชจ๋ ๊ฒ์ ์ค์ ๋ณดํฐ์๋ ์ค์๋ ์ผ๋ค๋ ๋๋ฒ๊ด์๊ฒ ๋ฌ๋ ค ์์์ต๋๋ค. ๋ ผ์์ 5์ผ์งธ์ธ 2์ 9์ผ, ๊ทธ๋ ๋จ ์ธ ๋ฌธ์ฅ์ผ๋ก ๋ ๋งค์ฐ ์งง์ ๋ฉ๋ชจ๋ฅผ ๋ณด๋์ต๋๋ค. โ๋๋ฒ์์ฅ์๊ฒ ์ฐฌ์ฑํ๋ค.โ ๊ทธ๊ฒ์ผ๋ก ๋ ผ์์ ๋๋ฌ์ต๋๋ค. ๋๋ฒ์์ 5๋ 4๋ก ์ค๋ฐ๋ง ๋ํต๋ น์ ํต์ฌ ํ๊ฒฝ ์ ์ฑ ์ ์ฐจ๋จํ๋ ๋ช ๋ น์ ๋ด๋ ธ์ต๋๋ค. ์ฐ๋ฆฌ๊ฐ ๋ฐฉ๊ธ ์ดํด๋ณธ ์ด ๋ชจ๋ ๊ฐ์ธ์ ์ธ ๋ ผ์๊ณผ๋ ๋์กฐ์ ์ผ๋ก, ๋์ค์๊ฒ ๋ฐํ๋ ๋ช ๋ น์ ๋ฒ์ ์ํฌ์ด๋ง ๋ด๊ธด ํ ๋จ๋ฝ์ง๋ฆฌ ์ค๋ช ์๋ ํ๊ฒฐ๋ฌธ์ด์์ต๋๋ค.
โ๊ทธ๋ฆผ์ ์ฌ์โ์ ์ ์ฐ: ์ ๋ขฐ ์์ค๊ณผ ๋นํ์ ์ฌ๋ฒ๋ถ
์ด ๋น๋ฐ ๋ฌธ์๋ค์ ๊ทธ๋ฆผ์ ์ฌ์์ ๋ํ ๋นํ์ ์ ๋นํํฉ๋๋ค. ์ด๊ฒ์ ๋๋ฒ์์ด โ์ต๊ณ ์์ค์ ์์ (A+ work)โ์ ํ ๊ฒ์ด ์๋์์ต๋๋ค. ๋๋ฒ๊ด๋ค์ ๋ถ๋ง๊ณผ ์ง์ฆ์ ์ฌ๋ก์กํ ์์ด๋์ด๋ฅผ ๋์ง๊ณ , ์๋ก์๊ฒ ๋ ์นด๋กญ๊ฒ ๋ฐ์ํ๋ฉฐ, ๊ตญ๊ฐ ์ต๊ณ ๋ฒ์์ด ํด์ผ ํ ์ข ๋ฅ์ ์ผ์ ํ์ง ์๊ณ ์์์ต๋๋ค. ๊ทธ๋ค์ ์ถฉ๋ถํ ์ ๋ณด ์์ด ๋๋ฌด ๋น ๋ฅด๊ฒ ์์ง์์ผ๋ฉฐ, ์๋ฌด๋ โ์ด๊ฒ์ด ์ด๋๋ก ์ด์ด์ง๊น?โ๋ผ๊ณ ๋ฌป์ง ์์์ต๋๋ค.
์ด๋ฌํ ๋ณํ๋ ํนํ ํธ๋ผํ ํ์ ๋ถ ์์ ์ ๊ทน๋ช ํ๊ฒ ๋๋ฌ๋ฌ์ต๋๋ค. ๋๋ฒ์์ ์ด๋ฏผ, ๊ธฐ๊ด ๊ถํ, ์ ๋ถ ์ง์ถ, ํธ๋์ค์ ๋ ๊ตฐ์ธ ๋ฑ ์ฃผ์ ์ฌ์์ ๋ํ ์๋ง์ ๊ธด๊ธ ์์ฒญ์ ์๋ฌ๋ ธ๊ณ , ๋๋ถ๋ถ ํธ๋ผํ ํ์ ๋ถ์ ์์ ๋ค์ด์ฃผ์์ต๋๋ค. ์ค๋ฐ๋ง ํ์ ๋ถ ๋๋ ์ฃผ์ ๋ํต๋ น ์ด๋์ ํฐ๋ธ๋ฅผ ์ค๋จ์์ผฐ๋ ๊ฒ๊ณผ ๋ฌ๋ฆฌ, ๊ทธ๋ฆผ์ ์ฌ์๋ ํธ๋ผํ ๋ํต๋ น์ ์ ์ฑ ์ ์๋์ ์ผ๋ก ํ์ฉํ์ต๋๋ค.
์ ์นํ์๋ค์ ๊ทธ๋ฆผ์ ์ฌ์์์ โ์ ์ ์ฌ์โ๋ณด๋ค ํจ์ฌ ๋ ๋นํ์ ์ธ ํฌํ๊ฐ ์ผ์ด๋๋ค๊ณ ์ง์ ํฉ๋๋ค. ์ถฉ๋ถํ ์๊ณ ์ ์ฌ์๊ฐ ๋ถ์กฑํ ์๋ก ๋๋ฒ๊ด๋ค์ ๋นํ์ ์ถฉ๋์ด ๋ ๊ฐํ๊ฒ ์์ฉํ๋ค๋ ๊ฒ์ ๋๋ค. ์ค์ ๋ก ๋ฐ์ด๋ ํ์ ๋ถ ์์ , ๋๋ฒ์์ ์ธ ๊ฑด์ ๊ธด๊ธ ์์ฒญ์์ ๋ฐ์ด๋ ๋ํต๋ น์๊ฒ ๋ฐ๋ํ๋ ๊ฒฐ์ ์ ๋ด๋ ธ์ง๋ง, ์ด ์ฌ๊ฑด๋ค์ด ๋์ค์ ์ ์ ์ฌ์๋ฅผ ๊ฑฐ์ณค์ ๋๋ ๋ฐ์ด๋ ํ์ ๋ถ์ ์์ ๋ค์ด์ฃผ์์ต๋๋ค. ์ฆ, ์๊ณ ํ ์๊ฐ์ด ์ถฉ๋ถํ ์๋ก ๋นํ์ ์ถฉ๋์ ์ฝํด์ง๋ ๊ฒฝํฅ์ ๋ณด์ ๋๋ค.
์๊ณ ๋ ๋ํ ๋ฒ์์ ๋ํ โ๊ตญ๋ฏผ์ ์ ๋ขฐโ๋ฅผ ๊ตฌ์ถํฉ๋๋ค. ์ฐ๋ฆฌ๋ ์ ์ถ์ง ๊ณต๋ฌด์๋ค์ ์ ํํ์ง๋ง, ๋๋ฒ๊ด๋ค์ ์ ํํ์ง ์์ต๋๋ค. ๊ทธ๋ค์ ํ์ ๊ถ๋ ฅ์ ์ ์งํฉ๋๋ค. ๋ฐ๋ผ์ ํ์ฌ๊ฐ ์๊ฒฌ์ ์์ฑํ๋ ํ์๋ โ์ด ๊ฒฐ๊ณผ์ ๋์ํ์ง ์๋๋ผ๋ ๋ด๊ฐ ์ด๋ป๊ฒ ์ด๋ฐ ๊ฒฐ๋ก ์ ๋๋ฌํ๋์ง ๋ณด์ฌ์ค ํ ๋ ๋๋ฅผ ์ ๋ขฐํด์ผ ํ๋คโ๊ณ ๋งํ๋ ํ์์ ๋๋ค. ํ์ง๋ง ๋๋ฒ์์ด ๊ทธ๋ฆผ์ ์ฌ์๋ฅผ ํตํด ์ด๋ฌํ ๊ณผ์ ์ ๋ํญ ์ถ์ํ๋ฉด์, ์ฌ๋ฒ๋ถ์ ์ ๋น์ฑ์๋ ์ฌ๊ฐํ ์ํ์ด ๋ฐ๋ฆ ๋๋ค. ํ์ฌ ๋๋ฒ์์ ๋ํ ๊ตญ๋ฏผ์ ์ ๋ขฐ๋ ์ญ๋ ์ต์ ์์ค์ด๋ฉฐ, ์ฃผ์ ๊ฒฐ์ ์ ๋ด๋ฆฌ๋ ๋ฐ ๊ทธ๋ฆผ์ ์ฌ์์ ๋ํ ์์กด๋๊ฐ ๋์์ง์๋ก ์ด๋ฌํ ๋ฌธ์ ๋ ๋์ฑ ์ ํ๋ ๊ฒ์ ๋๋ค.
๋๋ฒ์์ โ๊ทธ๋ฆผ์ ์ฌ์โ๋ ๋จ์ํ ์ ์ฐจ์ ํธ์๋ฅผ ๋์ด, ์ฌ๋ฒ๋ถ์ ๋ณธ์ง๊ณผ ์ญํ ์ ๋ํ ๊น์ ์ฑ์ฐฐ์ ์๊ตฌํ๊ณ ์์ต๋๋ค. ๋น๋ฐ ์์ ์ด ๋๋ฌ๋ธ 5์ผ๊ฐ์ ๊ฒฉ๋์ ๋ฏธ๊ตญ ๋๋ฒ์์ด ์ด๋ค ๋ฐฉํฅ์ผ๋ก ๋์๊ฐ๊ณ ์๋์ง, ๊ทธ๋ฆฌ๊ณ ๊ทธ ๊ณผ์ ์์ ๋ฌด์์ ์์ด๊ฐ๊ณ ์๋์ง์ ๋ํ ์ค์ํ ๋จ์๋ฅผ ์ ๊ณตํฉ๋๋ค.