{"success":true,"count":5,"items":[{"videoId":"v1wZwxY3CMg","chunkIndex":0,"totalChunks":5,"title":"Inside Anthropic, the $965 Billion AI Juggernaut | The Circuit — Part 1 of 5","thumbnail":"https://i.ytimg.com/vi/v1wZwxY3CMg/maxresdefault.jpg","duration":2860,"uploader":"Bloomberg Originals","youtubeUrl":"https://www.youtube.com/watch?v=v1wZwxY3CMg","keywords":["ai","anthropic","large-language-models","startup","openai","machine-learning","artificial-intelligence","enterprise-software","ai-safety","founder-story"],"normalizedKeywords":["비즈니스·전략","엔지니어링","기술 트렌드"],"targetAudience":[{"who":"AI 창업자","why":"안전·제품·성장 사이의 균형과 조직 운영 방식을 배울 수 있음"},{"who":"엔지니어","why":"Claude Code 같은 개발 생산성 도구와 모델 고도화 방향을 이해할 수 있음"},{"who":"투자자","why":"AI 시장의 승자 조건, 수익화, 밸류에이션 급등 배경을 읽을 수 있음"}],"normalizedAudience":["창업자·스타트업","엔지니어·개발자","투자자·VC"],"summary":"이 영상은 Anthropic과 Dario Amodei를 중심으로, 안전을 내세운 AI 회사가 어떻게 초대형 플레이어로 성장했는지를 보여준다. OpenAI 출신 팀이 왜 독립했고, Claude를 어떤 철학과 원칙으로 설계했는지, 그리고 그 선택이 실제 제품과 수익화로 어떻게 이어졌는지를 따라간다.\n\n핵심은 '좋은 AI'를 만들겠다는 가치 지향이 단순한 명분이 아니라 제품 전략, 고객 신뢰, 기업 성장의 엔진이 될 수 있다는 점이다. 동시에 AI가 강력해질수록 정부 사용, 보안, 거짓말·환각, 가치 주입 같은 난제가 더 커지며, Anthropic은 그 긴장 위에서 스스로를 정의하고 있다.","insights":["안전은 부가 기능이 아니라 제품 전략이 될 수 있다.","신뢰가 무너지면 기술적 협업보다 가치 충돌이 더 치명적이다.","AI 경쟁의 승부는 모델 성능뿐 아니라 배포 방식과 수익화에 있다.","Claude의 차별점은 '친근함'이 아니라 통제된 전문성이다.","좋은 모델은 정확성뿐 아니라 거짓말하지 않는 성질이 중요하다."],"keyClips":[{"clipId":"v1wZwxY3CMg:c0:2-14","startSegmentIndex":2,"endSegmentIndex":14,"startTime":12.89,"endTime":134.964,"durationSeconds":122.1,"preview":"안전이 곧 전략","mustSee":true},{"clipId":"v1wZwxY3CMg:c0:15-19","startSegmentIndex":15,"endSegmentIndex":19,"startTime":140.21,"endTime":184.73749999999998,"durationSeconds":44.5,"preview":"폭발적 성장의 논리","mustSee":false},{"clipId":"v1wZwxY3CMg:c0:20-25","startSegmentIndex":20,"endSegmentIndex":25,"startTime":184.44,"endTime":235.755,"durationSeconds":51.3,"preview":"형제의 서로 다른 재능","mustSee":false},{"clipId":"v1wZwxY3CMg:c0:26-41","startSegmentIndex":26,"endSegmentIndex":41,"startTime":236.16,"endTime":365.3728571428572,"durationSeconds":129.2,"preview":"OpenAI 탈출의 이유","mustSee":false},{"clipId":"v1wZwxY3CMg:c0:42-64","startSegmentIndex":42,"endSegmentIndex":64,"startTime":368.46999999999997,"endTime":576.193,"durationSeconds":207.7,"preview":"Claude를 길들이는 법","mustSee":true},{"clipId":"v1wZwxY3CMg:c0:65-68","startSegmentIndex":65,"endSegmentIndex":68,"startTime":575.9,"endTime":602.9275,"durationSeconds":27,"preview":"기업 고객에 베팅","mustSee":false}],"curatedSegments":[{"segmentIndex":31,"text":"predicting that large language models would improve simply by adding more data and computing power, even if the underlying algorithm stayed the same At that point in time,","startTime":275.64,"endTime":285.21299999999997,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"스케일링 법칙 핵심을 설명함."},{"segmentIndex":38,"text":"that alone is not sufficient to leave. When you feel that you can't trust someone, when you feel that their values are not what they say they are,","startTime":334.28999999999996,"endTime":343.1689999999999,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"신뢰 붕괴의 기준을 명확히 말함."},{"segmentIndex":39,"text":"when you feel that they're not honest, that makes it very hard to, you know, to continue to work with a company, to continue to trust the company.","startTime":342.98999999999995,"endTime":352.12500000000006,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"신뢰 문제의 실질적 결과를 설명."},{"segmentIndex":40,"text":"And look, at the end of the day, why argue with someone when you don't have the same vision and you don't trust them?","startTime":351.75,"endTime":359.19000000000005,"durationSeconds":7,"level":"B2","overallScore":7.8,"rationale":"관계 단절의 원칙을 직설적으로 제시."},{"segmentIndex":51,"text":"So the goal is not for it to be your best friend, but it's not for it to be sort of cold, rote, calculating. It should feel approachable, but distant, right? Professional.","startTime":448.23999999999995,"endTime":458.01124999999996,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"이상적 AI 톤을 정교하게 규정함."},{"segmentIndex":53,"text":"accidentally or intentionally, right? Lying, we call hallucinations. It makes something up. The models are just trained to predict the next word. So sometimes they don't know and they just invent something.","startTime":466.41999999999996,"endTime":476.34049999999996,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"환각 원인을 학습자 친화적으로 설명."},{"segmentIndex":54,"text":"Models sometimes, as we've shown in our research, can purposely try to deceive you. We have to make sure that doesn't happen in production models that we expose to customers. And then there's a lot of work around harmlessness,","startTime":475.96,"endTime":486.01,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"모델 기만 가능성과 대응을 제시."},{"segmentIndex":56,"text":"like is there a universal good? There's so many religions, so many different beliefs, like how do you even decide that? So of course there's no universal standard for what makes something helpful or harmless,","startTime":495.46,"endTime":505.4525,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"가치 주입의 난제를 날카롭게 제기."},{"segmentIndex":57,"text":"but there are founding documents in human history, like the UN Declaration of Human Rights that we can use to train Claude's character.","startTime":505.15,"endTime":513.8225,"durationSeconds":9,"level":"C1","overallScore":7.8,"rationale":"보편 기준의 대안을 제시함."},{"segmentIndex":59,"text":"and how can we bake in some of the kind of core values that are consistent across religions types of belief","startTime":522.6999999999999,"endTime":530.26125,"durationSeconds":8,"level":"C1","overallScore":7.8,"rationale":"초월적 가치 설계 원리를 설명함."},{"segmentIndex":1,"text":"Why did you make that bet? Was it a, was it a values decision or a business decision? Look, if you pick a business model that fundamentally conflicts with your values,","startTime":602.06,"endTime":612.075,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":""},{"segmentIndex":2,"text":"you're gonna have a hard time, right? Either you betray your own values or you become irrelevant. And so when we thought about it, we said, look, you know,","startTime":611.87,"endTime":620.6912500000001,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":""},{"segmentIndex":24,"text":"for engineers here? It's writing almost all of it on my team. So for me personally, it's been writing a hundred percent of my code for at least six months.","startTime":804.75,"endTime":813.5331249999999,"durationSeconds":9,"level":"B1","overallScore":7.4,"rationale":"AI가 실제 업무를 거의 대체함을 보여줌."},{"segmentIndex":51,"text":"A lot of low wage jobs, high inequality. You've been really direct about job loss. AI could eliminate half of all entry level white collar jobs in the next one to five years.","startTime":1039.18,"endTime":1049.0025,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"구체적이고 강한 일자리 전망을 제시함."},{"segmentIndex":56,"text":"But we're already starting to see the beginning of like, you know, there may be some people that it's, it's not making more productive, that it's better for the AI to just, to just do the thing.","startTime":1085.45,"endTime":1095.2857142857142,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"일부 업무에서 인간이 밀릴 조짐을 말함."},{"segmentIndex":1,"text":"because we haven't allowed it, but someone else's AI model, the AI model just makes the decision and the human never sees it. That's what we were standing up for. That's what we were fighting against.","startTime":1802.43,"endTime":1811.4325000000001,"durationSeconds":9,"level":"B2","overallScore":7.2,"rationale":""},{"segmentIndex":3,"text":"technology as a shortcut in war. The principle that was obeyed here is a human makes the final decision.","startTime":1820.1000000000001,"endTime":1828.6650000000002,"durationSeconds":9,"level":"B2","overallScore":7.4,"rationale":""},{"segmentIndex":39,"text":"So everything is a trade off now that we're in, you know, what I would describe as a commercially leading position, we can afford to move the dial even further","startTime":2146.52,"endTime":2156.693333333333,"durationSeconds":10,"level":"B2","overallScore":7.8,"rationale":"시장 지위와 신중함의 관계 설명."},{"segmentIndex":42,"text":"So I think, I actually think that's a very, that's a very serious question and I share those concerns. I don't think the government should outright take us over.","startTime":2172.4100000000003,"endTime":2181.5860000000002,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"우려를 인정하는 균형감이 크다."},{"segmentIndex":45,"text":"that's been built in the private sector and where government has not really had a serious role and is coming in late to the game.","startTime":2198.87,"endTime":2206.84375,"durationSeconds":8,"level":"B2","overallScore":7.8,"rationale":"정부의 늦은 개입을 날카롭게 짚음."}],"generatedAt":"2026-06-24T23:24:37.892Z","keyClipsTotalSec":1124},{"videoId":"v1wZwxY3CMg","chunkIndex":1,"totalChunks":5,"title":"Inside Anthropic, the $965 Billion AI Juggernaut | The Circuit — Part 2 of 5","thumbnail":"https://i.ytimg.com/vi/v1wZwxY3CMg/maxresdefault.jpg","duration":2860,"uploader":"Bloomberg Originals","youtubeUrl":"https://www.youtube.com/watch?v=v1wZwxY3CMg","keywords":["ai","anthropic","enterprise-ai","coding-agents","software-industry","jobs","productivity","ai-safety","business-model","future-of-work"],"normalizedKeywords":["기술 트렌드","비즈니스·전략","엔지니어링"],"targetAudience":[{"who":"엔지니어","why":"코딩 에이전트가 개발 업무를 어떻게 바꾸는지 직접 볼 수 있음"},{"who":"창업자","why":"AI 비즈니스 모델과 성장 전략, 시장 충격을 함께 이해할 수 있음"},{"who":"직장인","why":"AI가 내 일과 고용에 어떤 변화를 만들지 가늠하는 데 유용함"}],"normalizedAudience":["엔지니어·개발자","창업자·스타트업","지식노동자 일반"],"summary":"이 영상은 앤트로픽이 왜 소비자 대상보다 기업용 AI에 집중했는지, 그리고 그 선택이 어떻게 Claude Code/Cowork의 폭발적 성장으로 이어졌는지를 보여준다. 핵심 메시지는 AI의 가치를 '주의를 빼앗는 도구'가 아니라 '사람의 생산성을 높이는 도구'에 두어야 하며, 그 방향이 회사의 가치와도 맞아야 한다는 것이다.\n\n동시에 대규모 AI 도입이 소프트웨어 산업과 일자리에 미칠 충격도 정면으로 다룬다. 엔지니어는 더 생산적으로 변하지만, 결국 AI가 업무 대부분을 대신하게 되면 일부 역할은 사라질 수 있고, 그 부담은 기업만이 아니라 사회 전체가 논의해야 한다는 점을 강조한다. 마지막으로 앤트로픽이 이러한 위험을 과장 마케팅이 아니라 책임 있는 경고로 다루려 한다는 입장을 분명히 한다.","insights":["비즈니스 모델이 가치관과 충돌하면 결국 둘 중 하나를 잃는다.","기업용 AI는 주의력 착취보다 실질적 효용에 더 잘 맞는다.","코딩 에이전트는 개발자를 대체하기보다 개발 방식을 재정의한다.","AI의 생산성 향상은 장기적으로 일자리 재배치 문제를 만든다.","일자리 충격은 한 회사가 아니라 사회가 풀어야 할 문제다."],"keyClips":[{"clipId":"v1wZwxY3CMg:c1:1-8","startSegmentIndex":1,"endSegmentIndex":8,"startTime":602.06,"endTime":672.865,"durationSeconds":70.8,"preview":"가치와 사업의 정렬","mustSee":false},{"clipId":"v1wZwxY3CMg:c1:9-14","startSegmentIndex":9,"endSegmentIndex":14,"startTime":672.56,"endTime":725.5150000000001,"durationSeconds":53,"preview":"소프트웨어 재편","mustSee":false},{"clipId":"v1wZwxY3CMg:c1:19-25","startSegmentIndex":19,"endSegmentIndex":25,"startTime":758.88,"endTime":821.719,"durationSeconds":62.8,"preview":"코딩의 새 기본값","mustSee":true},{"clipId":"v1wZwxY3CMg:c1:26-32","startSegmentIndex":26,"endSegmentIndex":32,"startTime":821.43,"endTime":882.2645454545454,"durationSeconds":60.8,"preview":"몇 분 만의 앱 시연","mustSee":false},{"clipId":"v1wZwxY3CMg:c1:34-46","startSegmentIndex":34,"endSegmentIndex":46,"startTime":892.09,"endTime":1007.9285714285713,"durationSeconds":115.8,"preview":"폭발적 성장과 역할 변화","mustSee":false},{"clipId":"v1wZwxY3CMg:c1:47-56","startSegmentIndex":47,"endSegmentIndex":56,"startTime":1008.16,"endTime":1095.2857142857142,"durationSeconds":87.1,"preview":"일자리 충격의 현실","mustSee":false},{"clipId":"v1wZwxY3CMg:c1:57-69","startSegmentIndex":57,"endSegmentIndex":69,"startTime":1094.96,"endTime":1201.3149999999998,"durationSeconds":106.4,"preview":"책임 있는 경고","mustSee":false}],"curatedSegments":[{"segmentIndex":31,"text":"predicting that large language models would improve simply by adding more data and computing power, even if the underlying algorithm stayed the same At that point in time,","startTime":275.64,"endTime":285.21299999999997,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"스케일링 법칙 핵심을 설명함."},{"segmentIndex":38,"text":"that alone is not sufficient to leave. When you feel that you can't trust someone, when you feel that their values are not what they say they are,","startTime":334.28999999999996,"endTime":343.1689999999999,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"신뢰 붕괴의 기준을 명확히 말함."},{"segmentIndex":39,"text":"when you feel that they're not honest, that makes it very hard to, you know, to continue to work with a company, to continue to trust the company.","startTime":342.98999999999995,"endTime":352.12500000000006,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"신뢰 문제의 실질적 결과를 설명."},{"segmentIndex":40,"text":"And look, at the end of the day, why argue with someone when you don't have the same vision and you don't trust them?","startTime":351.75,"endTime":359.19000000000005,"durationSeconds":7,"level":"B2","overallScore":7.8,"rationale":"관계 단절의 원칙을 직설적으로 제시."},{"segmentIndex":51,"text":"So the goal is not for it to be your best friend, but it's not for it to be sort of cold, rote, calculating. It should feel approachable, but distant, right? Professional.","startTime":448.23999999999995,"endTime":458.01124999999996,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"이상적 AI 톤을 정교하게 규정함."},{"segmentIndex":53,"text":"accidentally or intentionally, right? Lying, we call hallucinations. It makes something up. The models are just trained to predict the next word. So sometimes they don't know and they just invent something.","startTime":466.41999999999996,"endTime":476.34049999999996,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"환각 원인을 학습자 친화적으로 설명."},{"segmentIndex":54,"text":"Models sometimes, as we've shown in our research, can purposely try to deceive you. We have to make sure that doesn't happen in production models that we expose to customers. And then there's a lot of work around harmlessness,","startTime":475.96,"endTime":486.01,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"모델 기만 가능성과 대응을 제시."},{"segmentIndex":56,"text":"like is there a universal good? There's so many religions, so many different beliefs, like how do you even decide that? So of course there's no universal standard for what makes something helpful or harmless,","startTime":495.46,"endTime":505.4525,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"가치 주입의 난제를 날카롭게 제기."},{"segmentIndex":57,"text":"but there are founding documents in human history, like the UN Declaration of Human Rights that we can use to train Claude's character.","startTime":505.15,"endTime":513.8225,"durationSeconds":9,"level":"C1","overallScore":7.8,"rationale":"보편 기준의 대안을 제시함."},{"segmentIndex":59,"text":"and how can we bake in some of the kind of core values that are consistent across religions types of belief","startTime":522.6999999999999,"endTime":530.26125,"durationSeconds":8,"level":"C1","overallScore":7.8,"rationale":"초월적 가치 설계 원리를 설명함."},{"segmentIndex":1,"text":"Why did you make that bet? Was it a, was it a values decision or a business decision? Look, if you pick a business model that fundamentally conflicts with your values,","startTime":602.06,"endTime":612.075,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":""},{"segmentIndex":2,"text":"you're gonna have a hard time, right? Either you betray your own values or you become irrelevant. And so when we thought about it, we said, look, you know,","startTime":611.87,"endTime":620.6912500000001,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":""},{"segmentIndex":24,"text":"for engineers here? It's writing almost all of it on my team. So for me personally, it's been writing a hundred percent of my code for at least six months.","startTime":804.75,"endTime":813.5331249999999,"durationSeconds":9,"level":"B1","overallScore":7.4,"rationale":"AI가 실제 업무를 거의 대체함을 보여줌."},{"segmentIndex":51,"text":"A lot of low wage jobs, high inequality. You've been really direct about job loss. AI could eliminate half of all entry level white collar jobs in the next one to five years.","startTime":1039.18,"endTime":1049.0025,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"구체적이고 강한 일자리 전망을 제시함."},{"segmentIndex":56,"text":"But we're already starting to see the beginning of like, you know, there may be some people that it's, it's not making more productive, that it's better for the AI to just, to just do the thing.","startTime":1085.45,"endTime":1095.2857142857142,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"일부 업무에서 인간이 밀릴 조짐을 말함."},{"segmentIndex":1,"text":"because we haven't allowed it, but someone else's AI model, the AI model just makes the decision and the human never sees it. That's what we were standing up for. That's what we were fighting against.","startTime":1802.43,"endTime":1811.4325000000001,"durationSeconds":9,"level":"B2","overallScore":7.2,"rationale":""},{"segmentIndex":3,"text":"technology as a shortcut in war. The principle that was obeyed here is a human makes the final decision.","startTime":1820.1000000000001,"endTime":1828.6650000000002,"durationSeconds":9,"level":"B2","overallScore":7.4,"rationale":""},{"segmentIndex":39,"text":"So everything is a trade off now that we're in, you know, what I would describe as a commercially leading position, we can afford to move the dial even further","startTime":2146.52,"endTime":2156.693333333333,"durationSeconds":10,"level":"B2","overallScore":7.8,"rationale":"시장 지위와 신중함의 관계 설명."},{"segmentIndex":42,"text":"So I think, I actually think that's a very, that's a very serious question and I share those concerns. I don't think the government should outright take us over.","startTime":2172.4100000000003,"endTime":2181.5860000000002,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"우려를 인정하는 균형감이 크다."},{"segmentIndex":45,"text":"that's been built in the private sector and where government has not really had a serious role and is coming in late to the game.","startTime":2198.87,"endTime":2206.84375,"durationSeconds":8,"level":"B2","overallScore":7.8,"rationale":"정부의 늦은 개입을 날카롭게 짚음."}],"generatedAt":"2026-06-24T23:25:09.194Z","keyClipsTotalSec":1124},{"videoId":"v1wZwxY3CMg","chunkIndex":2,"totalChunks":5,"title":"Inside Anthropic, the $965 Billion AI Juggernaut | The Circuit — Part 3 of 5","thumbnail":"https://i.ytimg.com/vi/v1wZwxY3CMg/maxresdefault.jpg","duration":2860,"uploader":"Bloomberg Originals","youtubeUrl":"https://www.youtube.com/watch?v=v1wZwxY3CMg","keywords":["artificial-intelligence","ai-safety","government-policy","defense-tech","geopolitics","labor-market","healthcare","anthropic","llm","ethics"],"normalizedKeywords":["기술 트렌드","비즈니스·전략","리더십·매니지먼트"],"targetAudience":[{"who":"AI 업계 실무자","why":"AI의 사회적 영향, 규제, 군사 활용 경계를 이해하는 데 유용함"},{"who":"정책·전략 담당자","why":"칩 수출통제와 군사 AI 사용을 둘러싼 논점을 읽을 수 있음"},{"who":"미래 일자리 관심자","why":"AI가 직업 구조와 의료·사무직을 어떻게 바꿀지 살펴볼 수 있음"}],"normalizedAudience":["지식노동자 일반","학생·주니어"],"summary":"이 영상은 Anthropic의 Dario Amodei를 통해 AI가 일자리, 사회, 국가 안보에 미칠 영향을 둘러싼 긴장을 다룬다. 한편으로는 AI가 생산성을 크게 높이고 의료·관리·법률 같은 지식노동을 재편할 것이라 보지만, 다른 한편으로는 대규모 실업과 사회 불안, 더 나아가 군사적 오용과 권위주의 국가와의 기술 경쟁을 경고한다.\n\n특히 이 대목에서는 'AI는 위험하지만 긍정적으로 대응해야 한다'는 메시지와 함께, Anthropic이 왜 정부·국방과 협력하면서도 무인살상무기·대규모 감시에는 선을 긋는지, 그리고 그 경계가 실제 전쟁 상황에서 얼마나 흔들리는지가 집중적으로 드러난다. 결과적으로 영상은 AI의 가능성을 낙관하면서도, 누가 어떤 기준으로 사용권을 결정할지에 대한 정치·윤리적 संघर्ष을 핵심 주제로 제시한다.","insights":["AI 충격의 핵심은 '대체'보다 '재배치'다.","기술이 좋아질수록 인간의 역할은 판단·관계·책임으로 이동한다.","의료처럼 진단은 자동화되고, 공감과 맥락은 더 중요해진다.","AI 경쟁은 제품 경쟁이 아니라 국가 안보 경쟁으로 번진다.","강한 기술 기업일수록 '어디까지 허용할지'의 경계 설정이 중요하다."],"keyClips":[{"clipId":"v1wZwxY3CMg:c2:1-18","startSegmentIndex":1,"endSegmentIndex":18,"startTime":1200.5,"endTime":1365.8099999999997,"durationSeconds":165.3,"preview":"AI와 일자리 전환","mustSee":true},{"clipId":"v1wZwxY3CMg:c2:22-30","startSegmentIndex":22,"endSegmentIndex":30,"startTime":1399.3200000000002,"endTime":1477.7074999999998,"durationSeconds":78.4,"preview":"AI와 칩 전쟁","mustSee":false},{"clipId":"v1wZwxY3CMg:c2:31-39","startSegmentIndex":31,"endSegmentIndex":39,"startTime":1478.95,"endTime":1554.72875,"durationSeconds":75.8,"preview":"반전주의와 현실","mustSee":false},{"clipId":"v1wZwxY3CMg:c2:40-68","startSegmentIndex":40,"endSegmentIndex":68,"startTime":1554.46,"endTime":1803.047,"durationSeconds":248.6,"preview":"군사용 AI의 딜레마","mustSee":true}],"curatedSegments":[{"segmentIndex":31,"text":"predicting that large language models would improve simply by adding more data and computing power, even if the underlying algorithm stayed the same At that point in time,","startTime":275.64,"endTime":285.21299999999997,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"스케일링 법칙 핵심을 설명함."},{"segmentIndex":38,"text":"that alone is not sufficient to leave. When you feel that you can't trust someone, when you feel that their values are not what they say they are,","startTime":334.28999999999996,"endTime":343.1689999999999,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"신뢰 붕괴의 기준을 명확히 말함."},{"segmentIndex":39,"text":"when you feel that they're not honest, that makes it very hard to, you know, to continue to work with a company, to continue to trust the company.","startTime":342.98999999999995,"endTime":352.12500000000006,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"신뢰 문제의 실질적 결과를 설명."},{"segmentIndex":40,"text":"And look, at the end of the day, why argue with someone when you don't have the same vision and you don't trust them?","startTime":351.75,"endTime":359.19000000000005,"durationSeconds":7,"level":"B2","overallScore":7.8,"rationale":"관계 단절의 원칙을 직설적으로 제시."},{"segmentIndex":51,"text":"So the goal is not for it to be your best friend, but it's not for it to be sort of cold, rote, calculating. It should feel approachable, but distant, right? Professional.","startTime":448.23999999999995,"endTime":458.01124999999996,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"이상적 AI 톤을 정교하게 규정함."},{"segmentIndex":53,"text":"accidentally or intentionally, right? Lying, we call hallucinations. It makes something up. The models are just trained to predict the next word. So sometimes they don't know and they just invent something.","startTime":466.41999999999996,"endTime":476.34049999999996,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"환각 원인을 학습자 친화적으로 설명."},{"segmentIndex":54,"text":"Models sometimes, as we've shown in our research, can purposely try to deceive you. We have to make sure that doesn't happen in production models that we expose to customers. And then there's a lot of work around harmlessness,","startTime":475.96,"endTime":486.01,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"모델 기만 가능성과 대응을 제시."},{"segmentIndex":56,"text":"like is there a universal good? There's so many religions, so many different beliefs, like how do you even decide that? So of course there's no universal standard for what makes something helpful or harmless,","startTime":495.46,"endTime":505.4525,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"가치 주입의 난제를 날카롭게 제기."},{"segmentIndex":57,"text":"but there are founding documents in human history, like the UN Declaration of Human Rights that we can use to train Claude's character.","startTime":505.15,"endTime":513.8225,"durationSeconds":9,"level":"C1","overallScore":7.8,"rationale":"보편 기준의 대안을 제시함."},{"segmentIndex":59,"text":"and how can we bake in some of the kind of core values that are consistent across religions types of belief","startTime":522.6999999999999,"endTime":530.26125,"durationSeconds":8,"level":"C1","overallScore":7.8,"rationale":"초월적 가치 설계 원리를 설명함."},{"segmentIndex":1,"text":"Why did you make that bet? Was it a, was it a values decision or a business decision? Look, if you pick a business model that fundamentally conflicts with your values,","startTime":602.06,"endTime":612.075,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":""},{"segmentIndex":2,"text":"you're gonna have a hard time, right? Either you betray your own values or you become irrelevant. And so when we thought about it, we said, look, you know,","startTime":611.87,"endTime":620.6912500000001,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":""},{"segmentIndex":24,"text":"for engineers here? It's writing almost all of it on my team. So for me personally, it's been writing a hundred percent of my code for at least six months.","startTime":804.75,"endTime":813.5331249999999,"durationSeconds":9,"level":"B1","overallScore":7.4,"rationale":"AI가 실제 업무를 거의 대체함을 보여줌."},{"segmentIndex":51,"text":"A lot of low wage jobs, high inequality. You've been really direct about job loss. AI could eliminate half of all entry level white collar jobs in the next one to five years.","startTime":1039.18,"endTime":1049.0025,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"구체적이고 강한 일자리 전망을 제시함."},{"segmentIndex":56,"text":"But we're already starting to see the beginning of like, you know, there may be some people that it's, it's not making more productive, that it's better for the AI to just, to just do the thing.","startTime":1085.45,"endTime":1095.2857142857142,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"일부 업무에서 인간이 밀릴 조짐을 말함."},{"segmentIndex":1,"text":"because we haven't allowed it, but someone else's AI model, the AI model just makes the decision and the human never sees it. That's what we were standing up for. That's what we were fighting against.","startTime":1802.43,"endTime":1811.4325000000001,"durationSeconds":9,"level":"B2","overallScore":7.2,"rationale":""},{"segmentIndex":3,"text":"technology as a shortcut in war. The principle that was obeyed here is a human makes the final decision.","startTime":1820.1000000000001,"endTime":1828.6650000000002,"durationSeconds":9,"level":"B2","overallScore":7.4,"rationale":""},{"segmentIndex":39,"text":"So everything is a trade off now that we're in, you know, what I would describe as a commercially leading position, we can afford to move the dial even further","startTime":2146.52,"endTime":2156.693333333333,"durationSeconds":10,"level":"B2","overallScore":7.8,"rationale":"시장 지위와 신중함의 관계 설명."},{"segmentIndex":42,"text":"So I think, I actually think that's a very, that's a very serious question and I share those concerns. I don't think the government should outright take us over.","startTime":2172.4100000000003,"endTime":2181.5860000000002,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"우려를 인정하는 균형감이 크다."},{"segmentIndex":45,"text":"that's been built in the private sector and where government has not really had a serious role and is coming in late to the game.","startTime":2198.87,"endTime":2206.84375,"durationSeconds":8,"level":"B2","overallScore":7.8,"rationale":"정부의 늦은 개입을 날카롭게 짚음."}],"generatedAt":"2026-06-24T23:25:17.188Z","keyClipsTotalSec":1124},{"videoId":"v1wZwxY3CMg","chunkIndex":3,"totalChunks":5,"title":"Inside Anthropic, the $965 Billion AI Juggernaut | The Circuit — Part 4 of 5","thumbnail":"https://i.ytimg.com/vi/v1wZwxY3CMg/maxresdefault.jpg","duration":2860,"uploader":"Bloomberg Originals","youtubeUrl":"https://www.youtube.com/watch?v=v1wZwxY3CMg","keywords":["artificial-intelligence","ai-safety","cybersecurity","national-security","governance","ethics","regulation","leadership","geopolitics"],"normalizedKeywords":["기술 트렌드","비즈니스·전략","리더십·매니지먼트"],"targetAudience":[{"who":"AI 업계 실무자","why":"모델 출시, 안전성, 규제 대응의 현실적 딜레마를 볼 수 있음"},{"who":"정책·규제 관심자","why":"민간 AI 통제와 정부 개입의 경계 문제를 이해하는 데 유용함"},{"who":"리더십 고민자","why":"불확실한 기술 환경에서 원칙을 세우고 조직을 정렬하는 법을 배울 수 있음"}],"normalizedAudience":["지식노동자 일반","창업자·스타트업","리서처·학자"],"summary":"이 영상은 Anthropic과 공동창업자 다리오 아모데이가 AI의 위험과 통제 문제를 어떻게 바라보는지 집중적으로 보여준다. 특히 Claude 계열보다 더 강력한 내부 모델 Mythos를 둘러싼 보안·국가안보·배포 제한 논쟁을 통해, AI가 왜 단순한 제품이 아니라 사회적 인프라이자 잠재적 무기인지 드러낸다.\n\n핵심 메시지는 '무조건 공개'도, '무조건 금지'도 아니라는 것이다. 아모데이는 인간 최종결정 원칙, 사전 테스트·감사, 제한적 배포 같은 안전장치를 강조하면서도, 기술을 독점하거나 정부가 전면 장악하는 극단 모두에 비판적이다. 동시에 그는 경쟁 압박 속에서도 조직 내부 정렬과 공개적 커뮤니케이션이 리더에게 중요한 방어막이 된다고 말한다.","insights":["강력한 AI일수록 속도보다 통제가 더 중요하다.","인간 최종결정 원칙은 자동화 사고를 막는 최소선이다.","안전 배포는 마케팅이 아니라 실제로는 경쟁 우위가 될 수 있다.","정부·기업 어느 한쪽의 독점도 고위험 기술엔 불안정하다.","불확실성이 큰 리더십일수록 내부 정렬과 투명성이 버팀목이다."],"keyClips":[{"clipId":"v1wZwxY3CMg:c3:1-13","startSegmentIndex":1,"endSegmentIndex":13,"startTime":1802.43,"endTime":1917.783,"durationSeconds":115.4,"preview":"전쟁에서 인간 결재","mustSee":false},{"clipId":"v1wZwxY3CMg:c3:14-19","startSegmentIndex":14,"endSegmentIndex":19,"startTime":1917.99,"endTime":1967.942857142857,"durationSeconds":50,"preview":"압박을 버티는 방식","mustSee":false},{"clipId":"v1wZwxY3CMg:c3:20-29","startSegmentIndex":20,"endSegmentIndex":29,"startTime":1973.92,"endTime":2067.0025,"durationSeconds":93.1,"preview":"Mythos와 배포제어","mustSee":true},{"clipId":"v1wZwxY3CMg:c3:30-40","startSegmentIndex":30,"endSegmentIndex":40,"startTime":2066.6800000000003,"endTime":2164.2749999999996,"durationSeconds":97.6,"preview":"안전이 곧 전략","mustSee":false},{"clipId":"v1wZwxY3CMg:c3:41-59","startSegmentIndex":41,"endSegmentIndex":59,"startTime":2163.98,"endTime":2325.4350000000004,"durationSeconds":161.5,"preview":"정부와 기업의 경계","mustSee":false},{"clipId":"v1wZwxY3CMg:c3:60-67","startSegmentIndex":60,"endSegmentIndex":67,"startTime":2331.7200000000003,"endTime":2400.9100000000003,"durationSeconds":69.2,"preview":"불안 속의 산업","mustSee":false}],"curatedSegments":[{"segmentIndex":31,"text":"predicting that large language models would improve simply by adding more data and computing power, even if the underlying algorithm stayed the same At that point in time,","startTime":275.64,"endTime":285.21299999999997,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"스케일링 법칙 핵심을 설명함."},{"segmentIndex":38,"text":"that alone is not sufficient to leave. When you feel that you can't trust someone, when you feel that their values are not what they say they are,","startTime":334.28999999999996,"endTime":343.1689999999999,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"신뢰 붕괴의 기준을 명확히 말함."},{"segmentIndex":39,"text":"when you feel that they're not honest, that makes it very hard to, you know, to continue to work with a company, to continue to trust the company.","startTime":342.98999999999995,"endTime":352.12500000000006,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"신뢰 문제의 실질적 결과를 설명."},{"segmentIndex":40,"text":"And look, at the end of the day, why argue with someone when you don't have the same vision and you don't trust them?","startTime":351.75,"endTime":359.19000000000005,"durationSeconds":7,"level":"B2","overallScore":7.8,"rationale":"관계 단절의 원칙을 직설적으로 제시."},{"segmentIndex":51,"text":"So the goal is not for it to be your best friend, but it's not for it to be sort of cold, rote, calculating. It should feel approachable, but distant, right? Professional.","startTime":448.23999999999995,"endTime":458.01124999999996,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"이상적 AI 톤을 정교하게 규정함."},{"segmentIndex":53,"text":"accidentally or intentionally, right? Lying, we call hallucinations. It makes something up. The models are just trained to predict the next word. So sometimes they don't know and they just invent something.","startTime":466.41999999999996,"endTime":476.34049999999996,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"환각 원인을 학습자 친화적으로 설명."},{"segmentIndex":54,"text":"Models sometimes, as we've shown in our research, can purposely try to deceive you. We have to make sure that doesn't happen in production models that we expose to customers. And then there's a lot of work around harmlessness,","startTime":475.96,"endTime":486.01,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"모델 기만 가능성과 대응을 제시."},{"segmentIndex":56,"text":"like is there a universal good? There's so many religions, so many different beliefs, like how do you even decide that? So of course there's no universal standard for what makes something helpful or harmless,","startTime":495.46,"endTime":505.4525,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"가치 주입의 난제를 날카롭게 제기."},{"segmentIndex":57,"text":"but there are founding documents in human history, like the UN Declaration of Human Rights that we can use to train Claude's character.","startTime":505.15,"endTime":513.8225,"durationSeconds":9,"level":"C1","overallScore":7.8,"rationale":"보편 기준의 대안을 제시함."},{"segmentIndex":59,"text":"and how can we bake in some of the kind of core values that are consistent across religions types of belief","startTime":522.6999999999999,"endTime":530.26125,"durationSeconds":8,"level":"C1","overallScore":7.8,"rationale":"초월적 가치 설계 원리를 설명함."},{"segmentIndex":1,"text":"Why did you make that bet? Was it a, was it a values decision or a business decision? Look, if you pick a business model that fundamentally conflicts with your values,","startTime":602.06,"endTime":612.075,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":""},{"segmentIndex":2,"text":"you're gonna have a hard time, right? Either you betray your own values or you become irrelevant. And so when we thought about it, we said, look, you know,","startTime":611.87,"endTime":620.6912500000001,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":""},{"segmentIndex":24,"text":"for engineers here? It's writing almost all of it on my team. So for me personally, it's been writing a hundred percent of my code for at least six months.","startTime":804.75,"endTime":813.5331249999999,"durationSeconds":9,"level":"B1","overallScore":7.4,"rationale":"AI가 실제 업무를 거의 대체함을 보여줌."},{"segmentIndex":51,"text":"A lot of low wage jobs, high inequality. You've been really direct about job loss. AI could eliminate half of all entry level white collar jobs in the next one to five years.","startTime":1039.18,"endTime":1049.0025,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"구체적이고 강한 일자리 전망을 제시함."},{"segmentIndex":56,"text":"But we're already starting to see the beginning of like, you know, there may be some people that it's, it's not making more productive, that it's better for the AI to just, to just do the thing.","startTime":1085.45,"endTime":1095.2857142857142,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"일부 업무에서 인간이 밀릴 조짐을 말함."},{"segmentIndex":1,"text":"because we haven't allowed it, but someone else's AI model, the AI model just makes the decision and the human never sees it. That's what we were standing up for. That's what we were fighting against.","startTime":1802.43,"endTime":1811.4325000000001,"durationSeconds":9,"level":"B2","overallScore":7.2,"rationale":""},{"segmentIndex":3,"text":"technology as a shortcut in war. The principle that was obeyed here is a human makes the final decision.","startTime":1820.1000000000001,"endTime":1828.6650000000002,"durationSeconds":9,"level":"B2","overallScore":7.4,"rationale":""},{"segmentIndex":39,"text":"So everything is a trade off now that we're in, you know, what I would describe as a commercially leading position, we can afford to move the dial even further","startTime":2146.52,"endTime":2156.693333333333,"durationSeconds":10,"level":"B2","overallScore":7.8,"rationale":"시장 지위와 신중함의 관계 설명."},{"segmentIndex":42,"text":"So I think, I actually think that's a very, that's a very serious question and I share those concerns. I don't think the government should outright take us over.","startTime":2172.4100000000003,"endTime":2181.5860000000002,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"우려를 인정하는 균형감이 크다."},{"segmentIndex":45,"text":"that's been built in the private sector and where government has not really had a serious role and is coming in late to the game.","startTime":2198.87,"endTime":2206.84375,"durationSeconds":8,"level":"B2","overallScore":7.8,"rationale":"정부의 늦은 개입을 날카롭게 짚음."}],"generatedAt":"2026-06-24T23:25:46.634Z","keyClipsTotalSec":1124},{"videoId":"v1wZwxY3CMg","chunkIndex":4,"totalChunks":5,"title":"Inside Anthropic, the $965 Billion AI Juggernaut | The Circuit — Part 5 of 5","thumbnail":"https://i.ytimg.com/vi/v1wZwxY3CMg/maxresdefault.jpg","duration":2860,"uploader":"Bloomberg Originals","youtubeUrl":"https://www.youtube.com/watch?v=v1wZwxY3CMg","keywords":["artificial-intelligence","ai-safety","ethics","technology-policy","governance","future-of-work","societal-impact","leadership","anthropic","risk-management"],"normalizedKeywords":["비즈니스·전략","리더십·매니지먼트","기술 트렌드"],"targetAudience":[{"who":"창업자","why":"강력한 기술을 만들며 신뢰와 책임을 어떻게 설계할지 배울 수 있음"},{"who":"투자자","why":"AI 산업의 규제, 위험, 장기적 파급효과를 판단하는 관점을 얻을 수 있음"},{"who":"리서처","why":"AI 안전, 사회적 영향, 거버넌스 논리를 구조적으로 이해하는 데 유용함"}],"normalizedAudience":["창업자·스타트업","투자자·VC","리서처·학자"],"summary":"이 영상은 Anthropic 공동창업자 Dario Amodei가 AI의 거대한 잠재력과 그에 따른 사회적 위험을 어떻게 동시에 바라보는지에 초점을 맞춘다. 그는 AI가 소셜미디어처럼 후폭풍을 남길 수 있다고 보고, 그래서 “처음부터 제대로 만드는 것”과 선제적 안전 설계를 회사의 핵심 임무로 강조한다. 동시에 AI가 일자리, 권력, 정치, 규제에 미칠 충격을 인정하며, 업계가 그 피해에 책임을 져야 한다는 입장을 분명히 한다.\n\n후반부에서는 Anthropic이 왜 위험을 경고하면서도 기술 발전을 밀어붙이는지, 그리고 그 모순처럼 보이는 태도가 사실은 ‘최악을 대비하면서 최선의 미래를 만들려는 전략’이라는 점을 설명한다. Google의 “don’t be evil” 사례, 오펜하이머와 원자폭탄의 역사, 10~25% 수준의 문명 붕괴 가능성 같은 강한 표현을 통해, AI 경쟁이 단순한 제품 경쟁이 아니라 문명적 리스크 관리의 문제임을 강조한다.","insights":["AI는 늦게 대응할수록 피해가 커져 선제 설계가 핵심이다.","기술 회사의 신뢰는 말이 아니라 실제 행동으로만 회복된다.","소셜미디어의 실패는 AI의 미래를 미리 보여주는 경고다.","AI 안전은 '성장'과 반대가 아니라 장기 생존의 조건이다.","문명적 리스크는 0이 될 수 없으니, 계속 낮추는 게 목표다."],"keyClips":[{"clipId":"v1wZwxY3CMg:c4:1-13","startSegmentIndex":1,"endSegmentIndex":13,"startTime":2400.1800000000003,"endTime":2517.04,"durationSeconds":116.9,"preview":"AI의 역풍과 선제 대응","mustSee":true},{"clipId":"v1wZwxY3CMg:c4:14-24","startSegmentIndex":14,"endSegmentIndex":24,"startTime":2516.83,"endTime":2608.0764999999997,"durationSeconds":91.2,"preview":"책임지는 기술 조직","mustSee":false},{"clipId":"v1wZwxY3CMg:c4:25-33","startSegmentIndex":25,"endSegmentIndex":33,"startTime":2607.82,"endTime":2686.165,"durationSeconds":78.3,"preview":"미래와 신뢰의 조건","mustSee":false},{"clipId":"v1wZwxY3CMg:c4:34-46","startSegmentIndex":34,"endSegmentIndex":46,"startTime":2685.92,"endTime":2793.7471428571425,"durationSeconds":107.8,"preview":"체크와 균형의 전략","mustSee":true},{"clipId":"v1wZwxY3CMg:c4:47-49","startSegmentIndex":47,"endSegmentIndex":49,"startTime":2821.67,"endTime":2850.2833333333338,"durationSeconds":28.6,"preview":"위험 속의 일상 균형","mustSee":false}],"curatedSegments":[{"segmentIndex":31,"text":"predicting that large language models would improve simply by adding more data and computing power, even if the underlying algorithm stayed the same At that point in time,","startTime":275.64,"endTime":285.21299999999997,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"스케일링 법칙 핵심을 설명함."},{"segmentIndex":38,"text":"that alone is not sufficient to leave. When you feel that you can't trust someone, when you feel that their values are not what they say they are,","startTime":334.28999999999996,"endTime":343.1689999999999,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"신뢰 붕괴의 기준을 명확히 말함."},{"segmentIndex":39,"text":"when you feel that they're not honest, that makes it very hard to, you know, to continue to work with a company, to continue to trust the company.","startTime":342.98999999999995,"endTime":352.12500000000006,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"신뢰 문제의 실질적 결과를 설명."},{"segmentIndex":40,"text":"And look, at the end of the day, why argue with someone when you don't have the same vision and you don't trust them?","startTime":351.75,"endTime":359.19000000000005,"durationSeconds":7,"level":"B2","overallScore":7.8,"rationale":"관계 단절의 원칙을 직설적으로 제시."},{"segmentIndex":51,"text":"So the goal is not for it to be your best friend, but it's not for it to be sort of cold, rote, calculating. It should feel approachable, but distant, right? Professional.","startTime":448.23999999999995,"endTime":458.01124999999996,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"이상적 AI 톤을 정교하게 규정함."},{"segmentIndex":53,"text":"accidentally or intentionally, right? Lying, we call hallucinations. It makes something up. The models are just trained to predict the next word. So sometimes they don't know and they just invent something.","startTime":466.41999999999996,"endTime":476.34049999999996,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"환각 원인을 학습자 친화적으로 설명."},{"segmentIndex":54,"text":"Models sometimes, as we've shown in our research, can purposely try to deceive you. We have to make sure that doesn't happen in production models that we expose to customers. And then there's a lot of work around harmlessness,","startTime":475.96,"endTime":486.01,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"모델 기만 가능성과 대응을 제시."},{"segmentIndex":56,"text":"like is there a universal good? There's so many religions, so many different beliefs, like how do you even decide that? So of course there's no universal standard for what makes something helpful or harmless,","startTime":495.46,"endTime":505.4525,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":"가치 주입의 난제를 날카롭게 제기."},{"segmentIndex":57,"text":"but there are founding documents in human history, like the UN Declaration of Human Rights that we can use to train Claude's character.","startTime":505.15,"endTime":513.8225,"durationSeconds":9,"level":"C1","overallScore":7.8,"rationale":"보편 기준의 대안을 제시함."},{"segmentIndex":59,"text":"and how can we bake in some of the kind of core values that are consistent across religions types of belief","startTime":522.6999999999999,"endTime":530.26125,"durationSeconds":8,"level":"C1","overallScore":7.8,"rationale":"초월적 가치 설계 원리를 설명함."},{"segmentIndex":1,"text":"Why did you make that bet? Was it a, was it a values decision or a business decision? Look, if you pick a business model that fundamentally conflicts with your values,","startTime":602.06,"endTime":612.075,"durationSeconds":10,"level":"B2","overallScore":8,"rationale":""},{"segmentIndex":2,"text":"you're gonna have a hard time, right? Either you betray your own values or you become irrelevant. And so when we thought about it, we said, look, you know,","startTime":611.87,"endTime":620.6912500000001,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":""},{"segmentIndex":24,"text":"for engineers here? It's writing almost all of it on my team. So for me personally, it's been writing a hundred percent of my code for at least six months.","startTime":804.75,"endTime":813.5331249999999,"durationSeconds":9,"level":"B1","overallScore":7.4,"rationale":"AI가 실제 업무를 거의 대체함을 보여줌."},{"segmentIndex":51,"text":"A lot of low wage jobs, high inequality. You've been really direct about job loss. AI could eliminate half of all entry level white collar jobs in the next one to five years.","startTime":1039.18,"endTime":1049.0025,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"구체적이고 강한 일자리 전망을 제시함."},{"segmentIndex":56,"text":"But we're already starting to see the beginning of like, you know, there may be some people that it's, it's not making more productive, that it's better for the AI to just, to just do the thing.","startTime":1085.45,"endTime":1095.2857142857142,"durationSeconds":10,"level":"C1","overallScore":7.8,"rationale":"일부 업무에서 인간이 밀릴 조짐을 말함."},{"segmentIndex":1,"text":"because we haven't allowed it, but someone else's AI model, the AI model just makes the decision and the human never sees it. That's what we were standing up for. That's what we were fighting against.","startTime":1802.43,"endTime":1811.4325000000001,"durationSeconds":9,"level":"B2","overallScore":7.2,"rationale":""},{"segmentIndex":3,"text":"technology as a shortcut in war. The principle that was obeyed here is a human makes the final decision.","startTime":1820.1000000000001,"endTime":1828.6650000000002,"durationSeconds":9,"level":"B2","overallScore":7.4,"rationale":""},{"segmentIndex":39,"text":"So everything is a trade off now that we're in, you know, what I would describe as a commercially leading position, we can afford to move the dial even further","startTime":2146.52,"endTime":2156.693333333333,"durationSeconds":10,"level":"B2","overallScore":7.8,"rationale":"시장 지위와 신중함의 관계 설명."},{"segmentIndex":42,"text":"So I think, I actually think that's a very, that's a very serious question and I share those concerns. I don't think the government should outright take us over.","startTime":2172.4100000000003,"endTime":2181.5860000000002,"durationSeconds":9,"level":"B2","overallScore":7.8,"rationale":"우려를 인정하는 균형감이 크다."},{"segmentIndex":45,"text":"that's been built in the private sector and where government has not really had a serious role and is coming in late to the game.","startTime":2198.87,"endTime":2206.84375,"durationSeconds":8,"level":"B2","overallScore":7.8,"rationale":"정부의 늦은 개입을 날카롭게 짚음."}],"generatedAt":"2026-06-24T23:26:26.629Z","keyClipsTotalSec":1124}]}