{"success":true,"count":4,"items":[{"videoId":"OaRhpwz_TGM","chunkIndex":0,"totalChunks":4,"title":"The Future of AI Agents with Andrew Ng | Interrupt 26 — Part 1 of 4","thumbnail":"https://i.ytimg.com/vi/OaRhpwz_TGM/maxresdefault.jpg","duration":1899,"uploader":"LangChain","youtubeUrl":"https://www.youtube.com/watch?v=OaRhpwz_TGM","keywords":["ai-agents","software-engineering","coding-agents","enterprise-ai","productivity","developer-tools","education","automation"],"normalizedKeywords":["엔지니어링","프로덕트","교육"],"targetAudience":[{"who":"엔지니어","why":"코딩 에이전트와 AI 빌딩 블록을 어떻게 조합할지 직접적인 힌트를 얻을 수 있음"},{"who":"프로덕트 매니저","why":"소프트웨어가 빨라질수록 PM과 주변 기능의 역할이 어떻게 바뀌는지 이해할 수 있음"},{"who":"창업자","why":"작은 팀으로 더 빠르게 제품을 만들고 조직을 설계하는 방식에 도움이 됨"}],"normalizedAudience":["엔지니어·개발자","프로덕트 매니저·기획자","창업자·스타트업"],"summary":"이 영상은 Andrew Ng가 최근 1년간 AI 분야에서 가장 크게 달라진 지점을 짚으며, 특히 코딩 에이전트와 에이전틱 워크플로가 얼마나 빠르게 실무를 바꾸고 있는지 이야기한다. 그는 과장은 있었지만 실제로는 코딩 에이전트의 발전이 유독 빠르며, 기업에서도 이런 워크플로가 본격적으로 들어오고 있다고 본다.\n\n핵심 주장은 '코드 작성이 10배~100배 빨라질수록 병목은 코딩이 아니라 주변 모든 것'이라는 점이다. PM, 마케팅, 법무, 디자인 등 여러 기능이 한꺼번에 병목이 되므로, 미래의 팀은 1~10명 규모의 고맥락 일반주의자들이 AI를 활용해 여러 역할을 겸하며 움직이는 구조로 재편될 가능성이 크다고 말한다. 또한 개발자에게 필요한 것은 새로운 AI 빌딩 블록을 충분히 이해하고, 코딩 에이전트를 이용해 이를 빠르게 조립하는 능력이며, 최신 문서를 에이전트가 바로 활용할 수 있게 하는 Context Hub 같은 도구와 교육이 중요하다고 강조한다.","insights":["코딩이 빨라질수록 병목은 개발 밖으로 이동한다.","작은 팀이 여러 역할을 겸할수록 실행 속도가 극대화된다.","미래의 경쟁력은 빌딩 블록을 얼마나 잘 조립하느냐다.","에이전트는 최신 도구 문서를 모르면 실전에서 둔해진다.","AI 시대의 학습은 도구 사용법보다 조합 능력에 가깝다."],"keyClips":[{"clipId":"OaRhpwz_TGM:c0:4-14","startSegmentIndex":4,"endSegmentIndex":14,"startTime":41.32,"endTime":132.5335714285714,"durationSeconds":91.2,"preview":"코딩 에이전트의 가속","mustSee":false},{"clipId":"OaRhpwz_TGM:c0:15-32","startSegmentIndex":15,"endSegmentIndex":32,"startTime":132.38,"endTime":292.93333333333334,"durationSeconds":160.6,"preview":"병목은 코딩 밖에 있다","mustSee":false},{"clipId":"OaRhpwz_TGM:c0:33-40","startSegmentIndex":33,"endSegmentIndex":40,"startTime":292.59999999999997,"endTime":364.31928571428574,"durationSeconds":71.7,"preview":"누가 이 역할을 하나","mustSee":false},{"clipId":"OaRhpwz_TGM:c0:41-63","startSegmentIndex":41,"endSegmentIndex":63,"startTime":364.21999999999997,"endTime":573.0316666666668,"durationSeconds":208.8,"preview":"조립 능력이 실력이다","mustSee":true},{"clipId":"OaRhpwz_TGM:c0:64-66","startSegmentIndex":64,"endSegmentIndex":66,"startTime":572.84,"endTime":600.676,"durationSeconds":27.8,"preview":"교육의 재설계","mustSee":false}],"curatedSegments":[{"segmentIndex":68,"text":"Sometimes driving incremental gains is harder than driving transformative gains.","startTime":1183.2800000000002,"endTime":1189.62,"durationSeconds":6,"level":"B1","overallScore":6.8,"rationale":""},{"segmentIndex":16,"text":"that if building becomes much faster, then deciding what to build or the product management work of scoping the project, we're getting customer feedback,","startTime":140.38,"endTime":148.73749999999998,"durationSeconds":8,"level":"B2","overallScore":6,"rationale":""},{"segmentIndex":19,"text":"product management bottleneck, pretty much everything else becomes a bottleneck. So some of my teams have seen the marketing bottleneck, because we can build so many features,","startTime":166.22,"endTime":175.85642857142855,"durationSeconds":10,"level":"B2","overallScore":5.8,"rationale":""},{"segmentIndex":24,"text":"anywhere from one to 10 engineers in a team, I've often generalist, high context, highly empowered generalists,","startTime":211.66,"endTime":220.36666666666665,"durationSeconds":9,"level":"C1","overallScore":6.4,"rationale":"새 팀 구성 원칙을 구체적으로 제시함."},{"segmentIndex":42,"text":"let's rethink the entire workflow and market a get approved in 10-minute loan product. Because rather than waiting around for a week for a human","startTime":943.48,"endTime":953.2786363636365,"durationSeconds":10,"level":"B2","overallScore":6,"rationale":""},{"segmentIndex":50,"text":"Because we can only save so much money, but growth has almost no practical ceiling. So the more exciting ideas I find usually relate to driving business growth rather than","startTime":1019.16,"endTime":1029.2095000000002,"durationSeconds":10,"level":"B2","overallScore":6,"rationale":""},{"segmentIndex":65,"text":"But there are some things where we swing for the fences and create so much value. We're not debating the disk drive 2% growth","startTime":1155.68,"endTime":1164.91375,"durationSeconds":9,"level":"B2","overallScore":6.6,"rationale":""},{"segmentIndex":23,"text":"work with the stakeholders to figure out what workflow to automate, help with the change management. So this is a very valuable role that takes deep technical judgment.","startTime":1401.1200000000001,"endTime":1409.8225,"durationSeconds":9,"level":"C1","overallScore":6.4,"rationale":""},{"segmentIndex":28,"text":"the leading coding agent a year from now. And so in moments of uncertainty like this, optionality is very valuable.","startTime":1442.76,"endTime":1450.67,"durationSeconds":8,"level":"C1","overallScore":6.4,"rationale":""},{"segmentIndex":53,"text":"organizing that to get it to AI or agents in the right time, in the right place, for it to create value, is suddenly much more valuable than it used to before.","startTime":1648.68,"endTime":1657.253,"durationSeconds":9,"level":"C1","overallScore":6.2,"rationale":""},{"segmentIndex":6,"text":"And then also the doomsaying narratives also got more traction, unfortunately, that would have guessed, including the job apocalypse, which I don't think is going to be a thing.","startTime":56.3,"endTime":65.4446111111111,"durationSeconds":9,"level":"B2","overallScore":5.2,"rationale":""},{"segmentIndex":20,"text":"our marketers are struggling to figure out what a new feature did in order to figure out how to communicate it. Previously, if you had a product that needed legal compliance,","startTime":175.72,"endTime":185.279,"durationSeconds":10,"level":"C1","overallScore":5.6,"rationale":"실무 예시와 인과관계가 분명함."},{"segmentIndex":21,"text":"if you took three months to build it, waiting a week for legal to sign off is maybe OK. But now, if you're building a day, then wait a week for legal to sign off, that's a legal compliance bottleneck.","startTime":185.22,"endTime":195.183,"durationSeconds":10,"level":"B2","overallScore":5,"rationale":""},{"segmentIndex":30,"text":"to what I had to do with my AI assistant. So I find these small teams of high context generalists that are deeply technical but able to use frontier technology","startTime":266,"endTime":275.0058333333333,"durationSeconds":9,"level":"C1","overallScore":5.6,"rationale":"역할 조합에 대한 인재상 설명이 좋음."},{"segmentIndex":32,"text":"to finally polish it before it goes out. But I find that these processes allow teams to move much faster. So that's been really exciting. - And what's the right background for these folks?","startTime":283.82,"endTime":292.93333333333334,"durationSeconds":9,"level":"B2","overallScore":5.2,"rationale":"효율 향상 결론은 있으나 길지 않음."},{"segmentIndex":2,"text":"So in terms of training, the thing that's clear is that what people have to learn has changed significantly. I think for developers, with the learning coding agents,","startTime":609.4,"endTime":619.2505555555556,"durationSeconds":10,"level":"B2","overallScore":5.6,"rationale":""},{"segmentIndex":21,"text":"But it feels like I think the transformation of education has been over-hyped. I think something is coming. But today, taking online courses,","startTime":761.4399999999999,"endTime":771.2040000000001,"durationSeconds":10,"level":"B2","overallScore":5.4,"rationale":""},{"segmentIndex":31,"text":"All of us have invested in the bottom-up innovation, that a thousand flowers bloom strategy, and for the most part is not paying off.","startTime":853.88,"endTime":861.7522222222223,"durationSeconds":8,"level":"C1","overallScore":5.2,"rationale":""},{"segmentIndex":32,"text":"So CEOs and boards are asking, where is the ROI for AI? I think we should keep on investing in bottom-up innovation. Let's keep on doing that.","startTime":861.56,"endTime":870.609,"durationSeconds":9,"level":"B2","overallScore":5.4,"rationale":""},{"segmentIndex":34,"text":"incremental efficiency gains, which are actually a good thing, but not the transmission, not the broader transformation that AI has been promising us and that I think we should work to deliver.","startTime":875.98,"endTime":886.2,"durationSeconds":10,"level":"C1","overallScore":5.2,"rationale":""}],"generatedAt":"2026-06-26T14:59:43.977Z","keyClipsTotalSec":811},{"videoId":"OaRhpwz_TGM","chunkIndex":1,"totalChunks":4,"title":"The Future of AI Agents with Andrew Ng | Interrupt 26 — Part 2 of 4","thumbnail":"https://i.ytimg.com/vi/OaRhpwz_TGM/maxresdefault.jpg","duration":1899,"uploader":"LangChain","youtubeUrl":"https://www.youtube.com/watch?v=OaRhpwz_TGM","keywords":["ai-agents","enterprise-ai","education-tech","online-learning","workflow-automation","customer-service","business-transformation","roi","generative-ai"],"normalizedKeywords":["비즈니스·전략","프로덕트","교육"],"targetAudience":[{"who":"기업 의사결정자","why":"AI를 비용절감이 아닌 성장 레버로 바꾸는 관점을 얻을 수 있음"},{"who":"프로덕트 매니저","why":"기능 단위 자동화보다 업무 흐름 재설계를 보는 시야가 필요함"},{"who":"교육 기획자","why":"AI 시대 학습 경험을 어떻게 바꿀지 감을 잡을 수 있음"},{"who":"엔지니어","why":"인터랙티브한 학습/제품 경험을 기술로 어떻게 구현하는지 참고됨"}],"normalizedAudience":["프로덕트 매니저·기획자","엔지니어·개발자","지식노동자 일반"],"summary":"이 영상은 AI 시대에 교육과 기업 도입이 어떻게 바뀌어야 하는지에 대한 앤드류 응의 시각을 중심으로 전개된다. 그는 온라인 강의 자체의 한계를 인정하면서도, 이제는 단순한 영상 강의보다 대화형·인터랙티브 학습 경험이 더 중요해지고 있다고 말한다. 특히 CodeDream.ai 같은 실험을 예로 들어, 학습자가 강의를 '듣는' 대신 질문하고 끼어들 수 있는 형태가 교육의 다음 단계라고 본다.\n\n후반부에서는 기업의 AI 도입이 왜 기대만큼의 ROI를 못 내는지 설명한다. 그는 대부분의 조직이 바텀업 실험을 통해 개별 업무를 조금씩 자동화하는 데 그치고 있으며, 진짜 변화는 전체 워크플로우를 다시 설계하는 탑다운 접근이 있어야 나온다고 강조한다. 은행의 대출 심사나 콜센터, 드라이브스루 같은 사례를 통해, AI는 비용 절감 도구를 넘어 성장과 고객 경험 개선의 수단이 되어야 한다는 메시지를 밀도 있게 전달한다.","insights":["AI 도입은 기능 자동화보다 업무 흐름 재설계가 핵심이다.","바텀업 실험만으로는 ROI가 작고, 전사적 설계가 필요하다.","AI의 가장 큰 가치는 비용 절감보다 성장 창출에 있다.","교육의 미래는 수동 시청보다 대화형 학습 경험에 가깝다.","측정 가능한 프로젝트도 좋지만, 변혁적 과제는 더 큰 임팩트를 낸다."],"keyClips":[{"clipId":"OaRhpwz_TGM:c1:2-16","startSegmentIndex":2,"endSegmentIndex":16,"startTime":609.4,"endTime":734.816,"durationSeconds":125.4,"preview":"교육 경험의 재설계","mustSee":false},{"clipId":"OaRhpwz_TGM:c1:21-25","startSegmentIndex":21,"endSegmentIndex":25,"startTime":761.4399999999999,"endTime":807.7875,"durationSeconds":46.3,"preview":"강의보다 대화","mustSee":false},{"clipId":"OaRhpwz_TGM:c1:31-49","startSegmentIndex":31,"endSegmentIndex":49,"startTime":853.88,"endTime":1019.235,"durationSeconds":165.4,"preview":"AI는 전체를 바꾼다","mustSee":true},{"clipId":"OaRhpwz_TGM:c1:52-57","startSegmentIndex":52,"endSegmentIndex":57,"startTime":1040.06,"endTime":1092.589,"durationSeconds":52.5,"preview":"성장으로 연결하기","mustSee":false},{"clipId":"OaRhpwz_TGM:c1:59-70","startSegmentIndex":59,"endSegmentIndex":70,"startTime":1102.5400000000002,"endTime":1204.7314285714283,"durationSeconds":102.2,"preview":"ROI보다 큰 그림","mustSee":false}],"curatedSegments":[{"segmentIndex":68,"text":"Sometimes driving incremental gains is harder than driving transformative gains.","startTime":1183.2800000000002,"endTime":1189.62,"durationSeconds":6,"level":"B1","overallScore":6.8,"rationale":""},{"segmentIndex":16,"text":"that if building becomes much faster, then deciding what to build or the product management work of scoping the project, we're getting customer feedback,","startTime":140.38,"endTime":148.73749999999998,"durationSeconds":8,"level":"B2","overallScore":6,"rationale":""},{"segmentIndex":19,"text":"product management bottleneck, pretty much everything else becomes a bottleneck. So some of my teams have seen the marketing bottleneck, because we can build so many features,","startTime":166.22,"endTime":175.85642857142855,"durationSeconds":10,"level":"B2","overallScore":5.8,"rationale":""},{"segmentIndex":24,"text":"anywhere from one to 10 engineers in a team, I've often generalist, high context, highly empowered generalists,","startTime":211.66,"endTime":220.36666666666665,"durationSeconds":9,"level":"C1","overallScore":6.4,"rationale":"새 팀 구성 원칙을 구체적으로 제시함."},{"segmentIndex":42,"text":"let's rethink the entire workflow and market a get approved in 10-minute loan product. Because rather than waiting around for a week for a human","startTime":943.48,"endTime":953.2786363636365,"durationSeconds":10,"level":"B2","overallScore":6,"rationale":""},{"segmentIndex":50,"text":"Because we can only save so much money, but growth has almost no practical ceiling. So the more exciting ideas I find usually relate to driving business growth rather than","startTime":1019.16,"endTime":1029.2095000000002,"durationSeconds":10,"level":"B2","overallScore":6,"rationale":""},{"segmentIndex":65,"text":"But there are some things where we swing for the fences and create so much value. We're not debating the disk drive 2% growth","startTime":1155.68,"endTime":1164.91375,"durationSeconds":9,"level":"B2","overallScore":6.6,"rationale":""},{"segmentIndex":23,"text":"work with the stakeholders to figure out what workflow to automate, help with the change management. So this is a very valuable role that takes deep technical judgment.","startTime":1401.1200000000001,"endTime":1409.8225,"durationSeconds":9,"level":"C1","overallScore":6.4,"rationale":""},{"segmentIndex":28,"text":"the leading coding agent a year from now. And so in moments of uncertainty like this, optionality is very valuable.","startTime":1442.76,"endTime":1450.67,"durationSeconds":8,"level":"C1","overallScore":6.4,"rationale":""},{"segmentIndex":53,"text":"organizing that to get it to AI or agents in the right time, in the right place, for it to create value, is suddenly much more valuable than it used to before.","startTime":1648.68,"endTime":1657.253,"durationSeconds":9,"level":"C1","overallScore":6.2,"rationale":""},{"segmentIndex":6,"text":"And then also the doomsaying narratives also got more traction, unfortunately, that would have guessed, including the job apocalypse, which I don't think is going to be a thing.","startTime":56.3,"endTime":65.4446111111111,"durationSeconds":9,"level":"B2","overallScore":5.2,"rationale":""},{"segmentIndex":20,"text":"our marketers are struggling to figure out what a new feature did in order to figure out how to communicate it. Previously, if you had a product that needed legal compliance,","startTime":175.72,"endTime":185.279,"durationSeconds":10,"level":"C1","overallScore":5.6,"rationale":"실무 예시와 인과관계가 분명함."},{"segmentIndex":21,"text":"if you took three months to build it, waiting a week for legal to sign off is maybe OK. But now, if you're building a day, then wait a week for legal to sign off, that's a legal compliance bottleneck.","startTime":185.22,"endTime":195.183,"durationSeconds":10,"level":"B2","overallScore":5,"rationale":""},{"segmentIndex":30,"text":"to what I had to do with my AI assistant. So I find these small teams of high context generalists that are deeply technical but able to use frontier technology","startTime":266,"endTime":275.0058333333333,"durationSeconds":9,"level":"C1","overallScore":5.6,"rationale":"역할 조합에 대한 인재상 설명이 좋음."},{"segmentIndex":32,"text":"to finally polish it before it goes out. But I find that these processes allow teams to move much faster. So that's been really exciting. - And what's the right background for these folks?","startTime":283.82,"endTime":292.93333333333334,"durationSeconds":9,"level":"B2","overallScore":5.2,"rationale":"효율 향상 결론은 있으나 길지 않음."},{"segmentIndex":2,"text":"So in terms of training, the thing that's clear is that what people have to learn has changed significantly. I think for developers, with the learning coding agents,","startTime":609.4,"endTime":619.2505555555556,"durationSeconds":10,"level":"B2","overallScore":5.6,"rationale":""},{"segmentIndex":21,"text":"But it feels like I think the transformation of education has been over-hyped. I think something is coming. But today, taking online courses,","startTime":761.4399999999999,"endTime":771.2040000000001,"durationSeconds":10,"level":"B2","overallScore":5.4,"rationale":""},{"segmentIndex":31,"text":"All of us have invested in the bottom-up innovation, that a thousand flowers bloom strategy, and for the most part is not paying off.","startTime":853.88,"endTime":861.7522222222223,"durationSeconds":8,"level":"C1","overallScore":5.2,"rationale":""},{"segmentIndex":32,"text":"So CEOs and boards are asking, where is the ROI for AI? I think we should keep on investing in bottom-up innovation. Let's keep on doing that.","startTime":861.56,"endTime":870.609,"durationSeconds":9,"level":"B2","overallScore":5.4,"rationale":""},{"segmentIndex":34,"text":"incremental efficiency gains, which are actually a good thing, but not the transmission, not the broader transformation that AI has been promising us and that I think we should work to deliver.","startTime":875.98,"endTime":886.2,"durationSeconds":10,"level":"C1","overallScore":5.2,"rationale":""}],"generatedAt":"2026-06-26T16:09:28.806Z","keyClipsTotalSec":811},{"videoId":"OaRhpwz_TGM","chunkIndex":2,"totalChunks":4,"title":"The Future of AI Agents with Andrew Ng | Interrupt 26 — Part 3 of 4","thumbnail":"https://i.ytimg.com/vi/OaRhpwz_TGM/maxresdefault.jpg","duration":1899,"uploader":"LangChain","youtubeUrl":"https://www.youtube.com/watch?v=OaRhpwz_TGM","keywords":["ai-agents","enterprise-ai","vendor-neutral","open-source-ai","data-architecture","unstructured-data","forward-deployed-engineers","llm-infrastructure","business-strategy","prototype"],"normalizedKeywords":["비즈니스·전략","엔지니어링","기술 트렌드"],"targetAudience":[{"who":"기업 AI 담당자","why":"에이전트 도입 시 인력, 벤더, 데이터 전략의 우선순위를 잡는 데 도움된다"},{"who":"창업자","why":"AI 제품/서비스를 만들 때 옵션 유지와 자원 배분의 원칙을 배울 수 있다"},{"who":"엔지니어","why":"에이전트 워크플로, 관측성, 데이터 아키텍처 문제를 실무적으로 이해할 수 있다"},{"who":"PM","why":"무엇을 자동화할지, 어떤 데이터를 연결할지 결정하는 기준을 얻을 수 있다"}],"normalizedAudience":["창업자·스타트업","엔지니어·개발자","프로덕트 매니저·기획자"],"summary":"이 대화는 기업이 AI 에이전트를 도입할 때 무엇부터 준비해야 하는지에 대한 현실적인 조언을 중심으로 전개된다. Andrew Ng는 아이디어가 수백 개 쏟아질 때는 기술 가능성과 비즈니스 임팩트를 함께 따져 소수의 베팅만 집중해야 하며, 에이전트 프로젝트는 단순 개발이 아니라 고객 대응, 관측성, 변경관리까지 포함하는 복합 업무라고 강조한다.\n\n또한 FDE(Forward Deployed Engineer)는 유용하지만 과대평가될 수 있고, 특히 빠르게 바뀌는 모델 시장에서는 벤더 종속을 줄이는 옵셔널리티가 중요하다고 말한다. 오픈웨이트 모델의 활용, 오픈소스 보호, 그리고 무엇보다도 비정형 데이터와 권한/거버넌스를 AI-준비 상태로 재구성하는 데이터 아키텍처가 앞으로 큰 투자 영역이 될 것이라고 전망한다.","insights":["AI 도입은 아이디어 수가 아니라 선택의 질이 성패를 가른다.","에이전트는 기술보다 비즈니스 이해와 변경관리 역량이 더 중요하다.","모델 시장이 빠르게 바뀌니 벤더 종속보다 옵션 유지가 유리하다.","비정형 데이터 정리가 다음 세대의 큰 AI 투자 과제다.","프로토타이핑은 싸졌지만, 자원 집중은 여전히 전략 문제다."],"keyClips":[{"clipId":"OaRhpwz_TGM:c2:2-14","startSegmentIndex":2,"endSegmentIndex":14,"startTime":1211.64,"endTime":1324.6933333333332,"durationSeconds":113.1,"preview":"아이디어 선별의 현실","mustSee":true},{"clipId":"OaRhpwz_TGM:c2:15-24","startSegmentIndex":15,"endSegmentIndex":24,"startTime":1325,"endTime":1417.732857142857,"durationSeconds":92.7,"preview":"FDE의 진짜 역할","mustSee":false},{"clipId":"OaRhpwz_TGM:c2:25-37","startSegmentIndex":25,"endSegmentIndex":37,"startTime":1417.7800000000002,"endTime":1531.164,"durationSeconds":113.4,"preview":"옵션을 남기는 선택","mustSee":true},{"clipId":"OaRhpwz_TGM:c2:38-46","startSegmentIndex":38,"endSegmentIndex":46,"startTime":1531.64,"endTime":1604.8333333333333,"durationSeconds":73.2,"preview":"오픈모델의 가치","mustSee":false},{"clipId":"OaRhpwz_TGM:c2:47-67","startSegmentIndex":47,"endSegmentIndex":67,"startTime":1604.72,"endTime":1773.95,"durationSeconds":169.2,"preview":"AI-ready 데이터","mustSee":true},{"clipId":"OaRhpwz_TGM:c2:68-71","startSegmentIndex":68,"endSegmentIndex":71,"startTime":1773.9,"endTime":1802.6,"durationSeconds":28.7,"preview":"빠른 반복의 선택","mustSee":false}],"curatedSegments":[{"segmentIndex":68,"text":"Sometimes driving incremental gains is harder than driving transformative gains.","startTime":1183.2800000000002,"endTime":1189.62,"durationSeconds":6,"level":"B1","overallScore":6.8,"rationale":""},{"segmentIndex":16,"text":"that if building becomes much faster, then deciding what to build or the product management work of scoping the project, we're getting customer feedback,","startTime":140.38,"endTime":148.73749999999998,"durationSeconds":8,"level":"B2","overallScore":6,"rationale":""},{"segmentIndex":19,"text":"product management bottleneck, pretty much everything else becomes a bottleneck. So some of my teams have seen the marketing bottleneck, because we can build so many features,","startTime":166.22,"endTime":175.85642857142855,"durationSeconds":10,"level":"B2","overallScore":5.8,"rationale":""},{"segmentIndex":24,"text":"anywhere from one to 10 engineers in a team, I've often generalist, high context, highly empowered generalists,","startTime":211.66,"endTime":220.36666666666665,"durationSeconds":9,"level":"C1","overallScore":6.4,"rationale":"새 팀 구성 원칙을 구체적으로 제시함."},{"segmentIndex":42,"text":"let's rethink the entire workflow and market a get approved in 10-minute loan product. Because rather than waiting around for a week for a human","startTime":943.48,"endTime":953.2786363636365,"durationSeconds":10,"level":"B2","overallScore":6,"rationale":""},{"segmentIndex":50,"text":"Because we can only save so much money, but growth has almost no practical ceiling. So the more exciting ideas I find usually relate to driving business growth rather than","startTime":1019.16,"endTime":1029.2095000000002,"durationSeconds":10,"level":"B2","overallScore":6,"rationale":""},{"segmentIndex":65,"text":"But there are some things where we swing for the fences and create so much value. We're not debating the disk drive 2% growth","startTime":1155.68,"endTime":1164.91375,"durationSeconds":9,"level":"B2","overallScore":6.6,"rationale":""},{"segmentIndex":23,"text":"work with the stakeholders to figure out what workflow to automate, help with the change management. So this is a very valuable role that takes deep technical judgment.","startTime":1401.1200000000001,"endTime":1409.8225,"durationSeconds":9,"level":"C1","overallScore":6.4,"rationale":""},{"segmentIndex":28,"text":"the leading coding agent a year from now. And so in moments of uncertainty like this, optionality is very valuable.","startTime":1442.76,"endTime":1450.67,"durationSeconds":8,"level":"C1","overallScore":6.4,"rationale":""},{"segmentIndex":53,"text":"organizing that to get it to AI or agents in the right time, in the right place, for it to create value, is suddenly much more valuable than it used to before.","startTime":1648.68,"endTime":1657.253,"durationSeconds":9,"level":"C1","overallScore":6.2,"rationale":""},{"segmentIndex":6,"text":"And then also the doomsaying narratives also got more traction, unfortunately, that would have guessed, including the job apocalypse, which I don't think is going to be a thing.","startTime":56.3,"endTime":65.4446111111111,"durationSeconds":9,"level":"B2","overallScore":5.2,"rationale":""},{"segmentIndex":20,"text":"our marketers are struggling to figure out what a new feature did in order to figure out how to communicate it. Previously, if you had a product that needed legal compliance,","startTime":175.72,"endTime":185.279,"durationSeconds":10,"level":"C1","overallScore":5.6,"rationale":"실무 예시와 인과관계가 분명함."},{"segmentIndex":21,"text":"if you took three months to build it, waiting a week for legal to sign off is maybe OK. But now, if you're building a day, then wait a week for legal to sign off, that's a legal compliance bottleneck.","startTime":185.22,"endTime":195.183,"durationSeconds":10,"level":"B2","overallScore":5,"rationale":""},{"segmentIndex":30,"text":"to what I had to do with my AI assistant. So I find these small teams of high context generalists that are deeply technical but able to use frontier technology","startTime":266,"endTime":275.0058333333333,"durationSeconds":9,"level":"C1","overallScore":5.6,"rationale":"역할 조합에 대한 인재상 설명이 좋음."},{"segmentIndex":32,"text":"to finally polish it before it goes out. But I find that these processes allow teams to move much faster. So that's been really exciting. - And what's the right background for these folks?","startTime":283.82,"endTime":292.93333333333334,"durationSeconds":9,"level":"B2","overallScore":5.2,"rationale":"효율 향상 결론은 있으나 길지 않음."},{"segmentIndex":2,"text":"So in terms of training, the thing that's clear is that what people have to learn has changed significantly. I think for developers, with the learning coding agents,","startTime":609.4,"endTime":619.2505555555556,"durationSeconds":10,"level":"B2","overallScore":5.6,"rationale":""},{"segmentIndex":21,"text":"But it feels like I think the transformation of education has been over-hyped. I think something is coming. But today, taking online courses,","startTime":761.4399999999999,"endTime":771.2040000000001,"durationSeconds":10,"level":"B2","overallScore":5.4,"rationale":""},{"segmentIndex":31,"text":"All of us have invested in the bottom-up innovation, that a thousand flowers bloom strategy, and for the most part is not paying off.","startTime":853.88,"endTime":861.7522222222223,"durationSeconds":8,"level":"C1","overallScore":5.2,"rationale":""},{"segmentIndex":32,"text":"So CEOs and boards are asking, where is the ROI for AI? I think we should keep on investing in bottom-up innovation. Let's keep on doing that.","startTime":861.56,"endTime":870.609,"durationSeconds":9,"level":"B2","overallScore":5.4,"rationale":""},{"segmentIndex":34,"text":"incremental efficiency gains, which are actually a good thing, but not the transmission, not the broader transformation that AI has been promising us and that I think we should work to deliver.","startTime":875.98,"endTime":886.2,"durationSeconds":10,"level":"C1","overallScore":5.2,"rationale":""}],"generatedAt":"2026-06-26T16:09:43.877Z","keyClipsTotalSec":811},{"videoId":"OaRhpwz_TGM","chunkIndex":3,"totalChunks":4,"title":"The Future of AI Agents with Andrew Ng | Interrupt 26 — Part 4 of 4","thumbnail":"https://i.ytimg.com/vi/OaRhpwz_TGM/maxresdefault.jpg","duration":1899,"uploader":"LangChain","youtubeUrl":"https://www.youtube.com/watch?v=OaRhpwz_TGM","keywords":["ai-agents","coding-agents","database","nosql","relational-database","software-development","iteration","engineering","productivity"],"normalizedKeywords":["엔지니어링","프로덕트","기술 트렌드"],"targetAudience":[{"who":"엔지니어","why":"AI 에이전트로 개발 속도가 빨라질 때 DB 선택과 설계 방식이 달라짐"},{"who":"프로덕트 매니저","why":"빠른 실험과 반복을 위해 기술 스택 선택 기준을 이해하는 데 도움됨"},{"who":"창업자","why":"초기 제품 개발에서 속도와 구조화의 균형을 잡는 관점을 얻을 수 있음"}],"normalizedAudience":["엔지니어·개발자","프로덕트 매니저·기획자","창업자·스타트업"],"summary":"이 구간에서는 AI 에이전트가 코드를 빠르게 작성하는 시대에, 오히려 데이터베이스와 스키마 선택도 '빠른 반복'을 중심으로 다시 생각해야 한다는 점을 강조한다. 화자는 AI가 실수할 가능성은 거의 없지만 완전히 0은 아니므로, 스키마를 미리 엄격하게 고정하는 것보다 NoSQL처럼 유연하게 데이터를 쌓고 읽을 때 해석하는 방식이 더 빠른 실험에 유리하다고 말한다.\n\n다만 NoSQL이 모든 대규모 생산 환경에 최적은 아니며, 아주 큰 엔터프라이즈급 워크로드에서는 더 확장성 있는 관계형 DB나 다른 솔루션이 필요할 수 있다고 인정한다. 핵심 메시지는 AI로 개발 속도가 빨라진 만큼, 기술 선택도 '안전성 최적화'보다 '반복 속도 최적화'에 더 맞춰야 한다는 것이다.","insights":["AI가 코딩을 빠르게 할수록 병목은 스키마 고정에서 생긴다.","읽을 때 해석하는 데이터 구조가 반복 속도를 높인다.","완벽한 안전보다 빠른 실험이 더 큰 경쟁력이 된다.","DB 설계는 규모가 아니라 변화 속도 기준으로도 봐야 한다.","기술 선택도 개발 방식 변화에 맞춰 다시 최적화해야 한다."],"keyClips":[{"clipId":"OaRhpwz_TGM:c3:1-3","startSegmentIndex":1,"endSegmentIndex":3,"startTime":1802.3200000000002,"endTime":1824.7183333333335,"durationSeconds":22.4,"preview":"실수보다 반복속도","mustSee":false},{"clipId":"OaRhpwz_TGM:c3:4-6","startSegmentIndex":4,"endSegmentIndex":6,"startTime":1824.5600000000002,"endTime":1851.1614285714286,"durationSeconds":26.6,"preview":"NoSQL의 유연성","mustSee":false},{"clipId":"OaRhpwz_TGM:c3:7-9","startSegmentIndex":7,"endSegmentIndex":9,"startTime":1850.88,"endTime":1875.7525,"durationSeconds":24.9,"preview":"스키마 고정의 함정","mustSee":true},{"clipId":"OaRhpwz_TGM:c3:10-11","startSegmentIndex":10,"endSegmentIndex":11,"startTime":1875.76,"endTime":1892.2299999999998,"durationSeconds":16.5,"preview":"기술선택의 재정의","mustSee":true}],"curatedSegments":[{"segmentIndex":68,"text":"Sometimes driving incremental gains is harder than driving transformative gains.","startTime":1183.2800000000002,"endTime":1189.62,"durationSeconds":6,"level":"B1","overallScore":6.8,"rationale":""},{"segmentIndex":16,"text":"that if building becomes much faster, then deciding what to build or the product management work of scoping the project, we're getting customer feedback,","startTime":140.38,"endTime":148.73749999999998,"durationSeconds":8,"level":"B2","overallScore":6,"rationale":""},{"segmentIndex":19,"text":"product management bottleneck, pretty much everything else becomes a bottleneck. So some of my teams have seen the marketing bottleneck, because we can build so many features,","startTime":166.22,"endTime":175.85642857142855,"durationSeconds":10,"level":"B2","overallScore":5.8,"rationale":""},{"segmentIndex":24,"text":"anywhere from one to 10 engineers in a team, I've often generalist, high context, highly empowered generalists,","startTime":211.66,"endTime":220.36666666666665,"durationSeconds":9,"level":"C1","overallScore":6.4,"rationale":"새 팀 구성 원칙을 구체적으로 제시함."},{"segmentIndex":42,"text":"let's rethink the entire workflow and market a get approved in 10-minute loan product. Because rather than waiting around for a week for a human","startTime":943.48,"endTime":953.2786363636365,"durationSeconds":10,"level":"B2","overallScore":6,"rationale":""},{"segmentIndex":50,"text":"Because we can only save so much money, but growth has almost no practical ceiling. So the more exciting ideas I find usually relate to driving business growth rather than","startTime":1019.16,"endTime":1029.2095000000002,"durationSeconds":10,"level":"B2","overallScore":6,"rationale":""},{"segmentIndex":65,"text":"But there are some things where we swing for the fences and create so much value. We're not debating the disk drive 2% growth","startTime":1155.68,"endTime":1164.91375,"durationSeconds":9,"level":"B2","overallScore":6.6,"rationale":""},{"segmentIndex":23,"text":"work with the stakeholders to figure out what workflow to automate, help with the change management. So this is a very valuable role that takes deep technical judgment.","startTime":1401.1200000000001,"endTime":1409.8225,"durationSeconds":9,"level":"C1","overallScore":6.4,"rationale":""},{"segmentIndex":28,"text":"the leading coding agent a year from now. And so in moments of uncertainty like this, optionality is very valuable.","startTime":1442.76,"endTime":1450.67,"durationSeconds":8,"level":"C1","overallScore":6.4,"rationale":""},{"segmentIndex":53,"text":"organizing that to get it to AI or agents in the right time, in the right place, for it to create value, is suddenly much more valuable than it used to before.","startTime":1648.68,"endTime":1657.253,"durationSeconds":9,"level":"C1","overallScore":6.2,"rationale":""},{"segmentIndex":6,"text":"And then also the doomsaying narratives also got more traction, unfortunately, that would have guessed, including the job apocalypse, which I don't think is going to be a thing.","startTime":56.3,"endTime":65.4446111111111,"durationSeconds":9,"level":"B2","overallScore":5.2,"rationale":""},{"segmentIndex":20,"text":"our marketers are struggling to figure out what a new feature did in order to figure out how to communicate it. Previously, if you had a product that needed legal compliance,","startTime":175.72,"endTime":185.279,"durationSeconds":10,"level":"C1","overallScore":5.6,"rationale":"실무 예시와 인과관계가 분명함."},{"segmentIndex":21,"text":"if you took three months to build it, waiting a week for legal to sign off is maybe OK. But now, if you're building a day, then wait a week for legal to sign off, that's a legal compliance bottleneck.","startTime":185.22,"endTime":195.183,"durationSeconds":10,"level":"B2","overallScore":5,"rationale":""},{"segmentIndex":30,"text":"to what I had to do with my AI assistant. So I find these small teams of high context generalists that are deeply technical but able to use frontier technology","startTime":266,"endTime":275.0058333333333,"durationSeconds":9,"level":"C1","overallScore":5.6,"rationale":"역할 조합에 대한 인재상 설명이 좋음."},{"segmentIndex":32,"text":"to finally polish it before it goes out. But I find that these processes allow teams to move much faster. So that's been really exciting. - And what's the right background for these folks?","startTime":283.82,"endTime":292.93333333333334,"durationSeconds":9,"level":"B2","overallScore":5.2,"rationale":"효율 향상 결론은 있으나 길지 않음."},{"segmentIndex":2,"text":"So in terms of training, the thing that's clear is that what people have to learn has changed significantly. I think for developers, with the learning coding agents,","startTime":609.4,"endTime":619.2505555555556,"durationSeconds":10,"level":"B2","overallScore":5.6,"rationale":""},{"segmentIndex":21,"text":"But it feels like I think the transformation of education has been over-hyped. I think something is coming. But today, taking online courses,","startTime":761.4399999999999,"endTime":771.2040000000001,"durationSeconds":10,"level":"B2","overallScore":5.4,"rationale":""},{"segmentIndex":31,"text":"All of us have invested in the bottom-up innovation, that a thousand flowers bloom strategy, and for the most part is not paying off.","startTime":853.88,"endTime":861.7522222222223,"durationSeconds":8,"level":"C1","overallScore":5.2,"rationale":""},{"segmentIndex":32,"text":"So CEOs and boards are asking, where is the ROI for AI? I think we should keep on investing in bottom-up innovation. Let's keep on doing that.","startTime":861.56,"endTime":870.609,"durationSeconds":9,"level":"B2","overallScore":5.4,"rationale":""},{"segmentIndex":34,"text":"incremental efficiency gains, which are actually a good thing, but not the transmission, not the broader transformation that AI has been promising us and that I think we should work to deliver.","startTime":875.98,"endTime":886.2,"durationSeconds":10,"level":"C1","overallScore":5.2,"rationale":""}],"generatedAt":"2026-06-26T16:09:53.052Z","keyClipsTotalSec":811}]}