{"success":true,"count":3,"items":[{"videoId":"qhnJDDX2hhU","chunkIndex":0,"totalChunks":3,"title":"Sam Altman Talks AGI Timeline & Next-Gen AI Capabilities | Snowflake Summit 2025 Fireside Chat — Part 1 of 3","thumbnail":"https://i.ytimg.com/vi/qhnJDDX2hhU/maxresdefault.jpg","duration":1254,"uploader":"Maginative","youtubeUrl":"https://www.youtube.com/watch?v=qhnJDDX2hhU","keywords":["artificial-intelligence","generative-ai","enterprise-software","agents","agi","machine-learning","retrieval","memory","openai","technology-trends"],"normalizedKeywords":["기술 트렌드","비즈니스·전략","엔지니어링"],"targetAudience":[{"who":"기업 리더","why":"AI 도입을 언제 어떻게 시작해야 하는지 실전 기준을 얻을 수 있음"},{"who":"창업자","why":"빠른 실험과 반복이 AI 시대의 경쟁우위라는 점을 배울 수 있음"},{"who":"개발자","why":"코딩 에이전트와 미래 업무 자동화의 방향을 이해하는 데 유용함"},{"who":"기획자","why":"메모리·검색·문맥이 제품 경험을 어떻게 바꾸는지 참고할 수 있음"}],"normalizedAudience":["창업자·스타트업","엔지니어·개발자","지식노동자 일반"],"summary":"이 대담은 2025년 기준 AI가 이미 실험 단계를 넘어 기업의 본격적인 도입 단계에 들어섰다는 판단을 중심으로 전개된다. Sam Altman은 모델이 빠르게 좋아지는 지금은 완벽한 타이밍을 기다리기보다, 작은 실험을 빠르게 반복하며 실패 비용을 낮추는 조직이 이긴다고 강조한다. 또한 검색·메모리·문맥 확장 같은 요소가 점점 중요해지고, 코딩 에이전트 같은 도구가 단순 자동화를 넘어 실제 업무를 맡는 수준으로 발전하고 있다고 본다.\n\n대화 후반부로 갈수록 핵심은 '에이전트가 무엇을 할 수 있나'에서 'AGI에 더 가까워지는 징후가 무엇인가'로 옮겨간다. Altman은 Codex 같은 코딩 에이전트를 AGI를 체감한 순간으로 언급하며, 앞으로는 복잡한 비즈니스 문제 해결이나 신지식 발견까지 일부 가능해질 것이라고 전망한다. 전반적으로 이 영상은 AI를 둘러싼 낙관론이 아니라, 지금 당장 기업이 어떻게 준비해야 하는지에 대한 실무적 조언과 다음 1년의 변화 방향을 제시한다.","insights":["AI 시대의 승자는 기다리는 회사가 아니라 빨리 실험하는 회사다.","모델 성능보다 중요한 건 조직의 반복 속도와 학습 속도다.","지금의 AI는 파일럿을 넘어 본격적인 생산성 도구로 성숙했다.","에이전트는 반복 업무를 넘어서 점점 긴 호흡의 일을 맡게 된다.","문맥·메모리·검색이 늘수록 AI는 더 유용하고 더 신뢰 가능해진다."],"keyClips":[{"clipId":"qhnJDDX2hhU:c0:14-22","startSegmentIndex":14,"endSegmentIndex":22,"startTime":124.88,"endTime":215.84,"durationSeconds":91,"preview":"빨리 실험하는 조직","mustSee":false},{"clipId":"qhnJDDX2hhU:c0:26-33","startSegmentIndex":26,"endSegmentIndex":33,"startTime":233.519,"endTime":309.759,"durationSeconds":76.2,"preview":"AI는 이제 실전이다","mustSee":true},{"clipId":"qhnJDDX2hhU:c0:35-39","startSegmentIndex":35,"endSegmentIndex":39,"startTime":314.44,"endTime":360.16,"durationSeconds":45.7,"preview":"다음 단계의 업무","mustSee":false},{"clipId":"qhnJDDX2hhU:c0:42-48","startSegmentIndex":42,"endSegmentIndex":48,"startTime":377.919,"endTime":434.72,"durationSeconds":56.8,"preview":"메모리와 검색의 역할","mustSee":false},{"clipId":"qhnJDDX2hhU:c0:50-62","startSegmentIndex":50,"endSegmentIndex":62,"startTime":448.87,"endTime":563.44,"durationSeconds":114.6,"preview":"에이전트의 미래","mustSee":true},{"clipId":"qhnJDDX2hhU:c0:63-67","startSegmentIndex":63,"endSegmentIndex":67,"startTime":563.44,"endTime":605.04,"durationSeconds":41.6,"preview":"AGI를 체감한 순간","mustSee":false}],"curatedSegments":[{"segmentIndex":14,"text":"I think just do it. Like there's still a lot of hesitancy and the models are changing so fast and there's all this reason to wait for the next model or you're going to sort of like wait and see if this is going to shake out this way or that or if you should build, you know, with thing that thing A or thing B. I as a general principle of technology when things are changing quickly that companies that have the quickest iteration speed um and sort of make the cost of making mistakes the lowest and the learning rate the highest win.","startTime":124.88,"endTime":154,"durationSeconds":29,"level":"C1","overallScore":9,"rationale":"실행 조언과 고급 표현이 매우 풍부."},{"segmentIndex":28,"text":"And so there are lots of application like chat bots whether it's structured or unstructured data the technology is mature you can adopt it yes you can always push the boundary on what else you can do with it there are edgier agentic applications but far away from the frontier I think this technology is actually ready for mainstream use interestingly I wouldn't have quite said the same thing last year um I would have said the same thing to a startup last year but to like a big enterprise I would have say like I would say like uh you can experiment a little bit, but this is maybe not totally ready for production use in most cases.","startTime":245.12,"endTime":281.199,"durationSeconds":36,"level":"C1","overallScore":9.2,"rationale":"기술 성숙도 판단과 표현이 매우 풍부."},{"segmentIndex":62,"text":"And as that expands to longer time horizons and higher levels, you know, at some point you get an AI scientist uh an AI agent that can go discover new science and that will be kind of a significant moment in the world.","startTime":549.12,"endTime":563.44,"durationSeconds":14,"level":"C1","overallScore":8.8,"rationale":"미래 AI의 전환점을 전망하는 고통찰 문장."},{"segmentIndex":9,"text":"And whether you declare the AGI victory in 24 or 26 or 28, um, and whether you declare the super intelligence victory in 28 or 30 or 32 is way less important than this one long beautiful, shockingly smooth exponential.","startTime":655.04,"endTime":674.2,"durationSeconds":19,"level":"C1","overallScore":8.6,"rationale":"라벨보다 성장곡선이 중요하단 통찰."},{"segmentIndex":32,"text":"I think it becomes a matter of debate like Sam is saying in uh I think sometimes it's also a philosophical question that I would liken to I don't know does a submarine swim um at one level it's absurd but of course it does and so I see these models as having incredible capabilities that we will like any person looking at what things are going to be like in 2030 would just declare that's AGI but remember you and I would to Sam's point would say the same thing in 2020 about what we are seeing in 25 to me it's the rate of progress that is truly astonishing and I sincerely believe that many great things are going to come out of it and similar to again how do we feel about the fact that a pretty decent computer can beat every person in the world that can play chess doesn't matter we still have people that play chess that are very good at it so I the definition matters.","startTime":847.04,"endTime":903.04,"durationSeconds":56,"level":"C2","overallScore":8.8,"rationale":"비유와 예시로 핵심 관점을 풍부히 전개."},{"segmentIndex":38,"text":"Um I have to ask you because we have you you're training more models you know you see the next capabilities before anybody else does uh what emergent behaviors are you seeing in the next set of models that change you know how you operate what you want to build from a product perspective how you're running open AI yeah the models over the next year or two years are going to be quite breathtaking um really there's a lot of progress ahead of us a lot of improvement to come and like we have seen in the previous big jumps you know from GPT3 to GPT4 businesses can just do things that totally were impossible with the previous generation of models and so what an enterprise will be able to do we talked about this a little bit but just like give it your hardest problem if you're a chip design company say go design me a better chip than I could have possibly had before um if you're a biotech company trying to cure some disease say just go work on this for me like that's not so far away.","startTime":931.279,"endTime":994.639,"durationSeconds":63,"level":"C2","overallScore":9,"rationale":"미래 역량과 활용상을 강하게 제시."},{"segmentIndex":40,"text":"Uh and these models ability to understand all the context you want to possibly give them, connect to every tool, every system, whatever, and then go think really hard like really brilliant reasoning and come back with an answer and have enough robustness that you can trust them to go off and do some work autonomously.","startTime":997.279,"endTime":1025.76,"durationSeconds":28,"level":"C1","overallScore":8.8,"rationale":"에이전트형 모델의 조건을 잘 묘사."},{"segmentIndex":48,"text":"But the amazing thing is they can reason. And if you think of it as this reasoning engine that we can then throw like all of the possible context of a business or a person's life into and any tool they need for that physics simulator or whatever else.","startTime":1076.16,"endTime":1089.48,"durationSeconds":13,"level":"C2","overallScore":9,"rationale":"모델 역할을 재정의하는 핵심 통찰."},{"segmentIndex":15,"text":"And certainly what we're seeing with enterprises and AI is the people that are making the early bets and iterating very quickly are doing much better than the people that are waiting to see how it's all going to shake out. Straight, what would you say?","startTime":154,"endTime":166.16,"durationSeconds":12,"level":"C1","overallScore":7.8,"rationale":"빠른 실행의 이점이 분명한 조언."},{"segmentIndex":35,"text":"Um and I think we'll be at the point next year where you can not only use a system to sort of automate some business processes or build these new products and services but you can really say I have this hugely important problem in my business.","startTime":314.44,"endTime":329.199,"durationSeconds":15,"level":"C1","overallScore":8,"rationale":"내년 변화상을 구체적으로 전망함."},{"segmentIndex":37,"text":"And the models will be able to go figure out things that teams of people on their own can't do.","startTime":332.8,"endTime":337.68,"durationSeconds":5,"level":"C1","overallScore":7.6,"rationale":"AI의 문제 해결 잠재력을 잘 드러냄."},{"segmentIndex":38,"text":"And the companies that are have gotten experience with these models are well positioned for a world where they can say okay you know AI system whatever go you know like redo my most critical project and here's a ton of compute think really hard just figure out the answer.","startTime":337.68,"endTime":354.8,"durationSeconds":17,"level":"C1","overallScore":8.2,"rationale":"미래 업무 방식 변화가 선명하게 제시됨."},{"segmentIndex":54,"text":"And you know maybe today it is like a sort of intern that can work for a couple of hours but at some point it'll be like an experienced software engineer that can work for days.","startTime":478.72,"endTime":488.24,"durationSeconds":10,"level":"C1","overallScore":8.2,"rationale":"에이전트 성장 비유가 매우 선명함."},{"segmentIndex":60,"text":"Um, I would bet next year that in some limited cases, at least in some small ways, we start to see agents that can help us discover new knowledge or can figure out solutions to business problems that are kind of very non-trivial.","startTime":521.8,"endTime":537.44,"durationSeconds":16,"level":"C1","overallScore":8,"rationale":"에이전트의 다음 단계 전망이 선명함."},{"segmentIndex":8,"text":"Um the thing that matters is the rate of progress that we have seen year over year for the last 5 years should continue for at least the next five probably well beyond that but hard to say.","startTime":643.24,"endTime":655.04,"durationSeconds":12,"level":"C1","overallScore":7.8,"rationale":"핵심은 지속적 진보라는 주장."},{"segmentIndex":10,"text":"Um, all of that said, to me, a system that can either autonomously discover new science or be such an incredible tool to people that our rate of scientific discovery in the world like quadruples or something.","startTime":674.2,"endTime":690.24,"durationSeconds":16,"level":"C1","overallScore":7.8,"rationale":"AGI 기준을 과학 발전으로 제시."},{"segmentIndex":20,"text":"That was like a bit of an aha moment for weight. If you could do this on the entirety of the web car corpus, you of course have search which can figure out which 10 pages to look at.","startTime":755.6,"endTime":765.519,"durationSeconds":10,"level":"C1","overallScore":8,"rationale":"요약과 검색의 연결 통찰이 큼."},{"segmentIndex":26,"text":"Even if you look at many of the post- training techniques that have come, it's a little bit of okay, take this incredibly powerful model, give it context for what's worked, what's not work, and use it to improve what it is producing.","startTime":805.68,"endTime":817.6,"durationSeconds":12,"level":"C1","overallScore":8,"rationale":"후처리의 본질을 맥락으로 설명함."},{"segmentIndex":29,"text":"There's always an infinity of context. Humans solve it by what we call attention where we focus on something.","startTime":826.079,"endTime":832.639,"durationSeconds":7,"level":"C1","overallScore":7.6,"rationale":"주의를 맥락 선택으로 비유한 통찰."},{"segmentIndex":30,"text":"I think of search as a tool for setting attention for a model.","startTime":832.639,"endTime":836.56,"durationSeconds":4,"level":"B2","overallScore":7.6,"rationale":"검색의 역할을 새 틀로 설명함."}],"generatedAt":"2026-06-24T23:33:30.011Z","keyClipsTotalSec":596},{"videoId":"qhnJDDX2hhU","chunkIndex":1,"totalChunks":3,"title":"Sam Altman Talks AGI Timeline & Next-Gen AI Capabilities | Snowflake Summit 2025 Fireside Chat — Part 2 of 3","thumbnail":"https://i.ytimg.com/vi/qhnJDDX2hhU/maxresdefault.jpg","duration":1254,"uploader":"Maginative","youtubeUrl":"https://www.youtube.com/watch?v=qhnJDDX2hhU","keywords":["ai","agi","language-models","reasoning","enterprise-ai","search","context-window","compute","openai","technology-trends"],"normalizedKeywords":["비즈니스·전략","프로덕트","기술 트렌드"],"targetAudience":[{"who":"창업자","why":"AI의 제품화 방향과 기업 고객이 원하는 가치가 무엇인지 읽을 수 있음"},{"who":"프로덕트 매니저","why":"모델 능력을 제품 기능으로 번역하는 관점과 컨텍스트 설계가 중요함을 보여줌"},{"who":"지식노동자","why":"AI를 단순 검색이 아니라 고난도 문제를 푸는 도구로 쓰는 법을 배울 수 있음"}],"normalizedAudience":["창업자·스타트업","프로덕트 매니저·기획자","지식노동자 일반"],"summary":"이 영상은 Sam Altman과 Snowflake 측 대담을 통해 AGI를 어떻게 정의하느냐보다, AI 능력이 매년 얼마나 매끄럽게 확장되고 있는지가 더 중요하다고 말한다. 화자는 ChatGPT 수준의 모델이 5년 전 기준으로 보면 이미 AGI로 보였을 것이라며, 실제 핵심은 '정의'가 아니라 과학 발견과 기업 생산성을 얼마나 크게 끌어올리느냐라고 주장한다.\n\n또한 검색은 모델의 attention을 설정하는 도구이며, 앞으로의 시스템은 방대한 컨텍스트와 외부 도구를 붙여 스스로 생각하고 행동하는 reasoning engine에 가까워질 것이라고 전망한다. 기업은 모델에게 가장 어려운 문제를 맡기는 방식으로 이미 큰 수익을 얻고 있으며, 더 많은 compute를 어디에 쓰느냐보다 hardest problem에 집중적으로 투입하는 습관이 중요하다는 메시지로 이어진다.","insights":["AGI의 경계보다 중요한 건 능력의 증가 속도다.","좋은 AI는 정답 저장소보다 추론 엔진에 가깝다.","검색의 본질은 모델의 주의를 좁혀 주는 것이다.","가장 큰 가치는 모델을 가장 어려운 문제에 쓰는 데서 나온다.","compute는 많이 쓸수록 아니라 가장 가치 있는 데 쓸수록 크다."],"keyClips":[{"clipId":"qhnJDDX2hhU:c1:1-13","startSegmentIndex":1,"endSegmentIndex":13,"startTime":600.949,"endTime":709.92,"durationSeconds":109,"preview":"AGI 정의보다 속도","mustSee":true},{"clipId":"qhnJDDX2hhU:c1:14-21","startSegmentIndex":14,"endSegmentIndex":21,"startTime":709.92,"endTime":772.8,"durationSeconds":62.9,"preview":"GPT3가 준 첫 신호","mustSee":false},{"clipId":"qhnJDDX2hhU:c1:22-30","startSegmentIndex":22,"endSegmentIndex":30,"startTime":772.8,"endTime":836.56,"durationSeconds":63.8,"preview":"검색은 주의 조절","mustSee":false},{"clipId":"qhnJDDX2hhU:c1:31-37","startSegmentIndex":31,"endSegmentIndex":37,"startTime":836.56,"endTime":931.279,"durationSeconds":94.7,"preview":"AGI보다 중요한 질문","mustSee":false},{"clipId":"qhnJDDX2hhU:c1:38-49","startSegmentIndex":38,"endSegmentIndex":49,"startTime":931.279,"endTime":1095.679,"durationSeconds":164.4,"preview":"에이전트형 기업 AI","mustSee":true},{"clipId":"qhnJDDX2hhU:c1:50-58","startSegmentIndex":50,"endSegmentIndex":58,"startTime":1095.679,"endTime":1186.08,"durationSeconds":90.4,"preview":"compute를 쓰는 법","mustSee":true}],"curatedSegments":[{"segmentIndex":14,"text":"I think just do it. Like there's still a lot of hesitancy and the models are changing so fast and there's all this reason to wait for the next model or you're going to sort of like wait and see if this is going to shake out this way or that or if you should build, you know, with thing that thing A or thing B. I as a general principle of technology when things are changing quickly that companies that have the quickest iteration speed um and sort of make the cost of making mistakes the lowest and the learning rate the highest win.","startTime":124.88,"endTime":154,"durationSeconds":29,"level":"C1","overallScore":9,"rationale":"실행 조언과 고급 표현이 매우 풍부."},{"segmentIndex":28,"text":"And so there are lots of application like chat bots whether it's structured or unstructured data the technology is mature you can adopt it yes you can always push the boundary on what else you can do with it there are edgier agentic applications but far away from the frontier I think this technology is actually ready for mainstream use interestingly I wouldn't have quite said the same thing last year um I would have said the same thing to a startup last year but to like a big enterprise I would have say like I would say like uh you can experiment a little bit, but this is maybe not totally ready for production use in most cases.","startTime":245.12,"endTime":281.199,"durationSeconds":36,"level":"C1","overallScore":9.2,"rationale":"기술 성숙도 판단과 표현이 매우 풍부."},{"segmentIndex":62,"text":"And as that expands to longer time horizons and higher levels, you know, at some point you get an AI scientist uh an AI agent that can go discover new science and that will be kind of a significant moment in the world.","startTime":549.12,"endTime":563.44,"durationSeconds":14,"level":"C1","overallScore":8.8,"rationale":"미래 AI의 전환점을 전망하는 고통찰 문장."},{"segmentIndex":9,"text":"And whether you declare the AGI victory in 24 or 26 or 28, um, and whether you declare the super intelligence victory in 28 or 30 or 32 is way less important than this one long beautiful, shockingly smooth exponential.","startTime":655.04,"endTime":674.2,"durationSeconds":19,"level":"C1","overallScore":8.6,"rationale":"라벨보다 성장곡선이 중요하단 통찰."},{"segmentIndex":32,"text":"I think it becomes a matter of debate like Sam is saying in uh I think sometimes it's also a philosophical question that I would liken to I don't know does a submarine swim um at one level it's absurd but of course it does and so I see these models as having incredible capabilities that we will like any person looking at what things are going to be like in 2030 would just declare that's AGI but remember you and I would to Sam's point would say the same thing in 2020 about what we are seeing in 25 to me it's the rate of progress that is truly astonishing and I sincerely believe that many great things are going to come out of it and similar to again how do we feel about the fact that a pretty decent computer can beat every person in the world that can play chess doesn't matter we still have people that play chess that are very good at it so I the definition matters.","startTime":847.04,"endTime":903.04,"durationSeconds":56,"level":"C2","overallScore":8.8,"rationale":"비유와 예시로 핵심 관점을 풍부히 전개."},{"segmentIndex":38,"text":"Um I have to ask you because we have you you're training more models you know you see the next capabilities before anybody else does uh what emergent behaviors are you seeing in the next set of models that change you know how you operate what you want to build from a product perspective how you're running open AI yeah the models over the next year or two years are going to be quite breathtaking um really there's a lot of progress ahead of us a lot of improvement to come and like we have seen in the previous big jumps you know from GPT3 to GPT4 businesses can just do things that totally were impossible with the previous generation of models and so what an enterprise will be able to do we talked about this a little bit but just like give it your hardest problem if you're a chip design company say go design me a better chip than I could have possibly had before um if you're a biotech company trying to cure some disease say just go work on this for me like that's not so far away.","startTime":931.279,"endTime":994.639,"durationSeconds":63,"level":"C2","overallScore":9,"rationale":"미래 역량과 활용상을 강하게 제시."},{"segmentIndex":40,"text":"Uh and these models ability to understand all the context you want to possibly give them, connect to every tool, every system, whatever, and then go think really hard like really brilliant reasoning and come back with an answer and have enough robustness that you can trust them to go off and do some work autonomously.","startTime":997.279,"endTime":1025.76,"durationSeconds":28,"level":"C1","overallScore":8.8,"rationale":"에이전트형 모델의 조건을 잘 묘사."},{"segmentIndex":48,"text":"But the amazing thing is they can reason. And if you think of it as this reasoning engine that we can then throw like all of the possible context of a business or a person's life into and any tool they need for that physics simulator or whatever else.","startTime":1076.16,"endTime":1089.48,"durationSeconds":13,"level":"C2","overallScore":9,"rationale":"모델 역할을 재정의하는 핵심 통찰."},{"segmentIndex":15,"text":"And certainly what we're seeing with enterprises and AI is the people that are making the early bets and iterating very quickly are doing much better than the people that are waiting to see how it's all going to shake out. Straight, what would you say?","startTime":154,"endTime":166.16,"durationSeconds":12,"level":"C1","overallScore":7.8,"rationale":"빠른 실행의 이점이 분명한 조언."},{"segmentIndex":35,"text":"Um and I think we'll be at the point next year where you can not only use a system to sort of automate some business processes or build these new products and services but you can really say I have this hugely important problem in my business.","startTime":314.44,"endTime":329.199,"durationSeconds":15,"level":"C1","overallScore":8,"rationale":"내년 변화상을 구체적으로 전망함."},{"segmentIndex":37,"text":"And the models will be able to go figure out things that teams of people on their own can't do.","startTime":332.8,"endTime":337.68,"durationSeconds":5,"level":"C1","overallScore":7.6,"rationale":"AI의 문제 해결 잠재력을 잘 드러냄."},{"segmentIndex":38,"text":"And the companies that are have gotten experience with these models are well positioned for a world where they can say okay you know AI system whatever go you know like redo my most critical project and here's a ton of compute think really hard just figure out the answer.","startTime":337.68,"endTime":354.8,"durationSeconds":17,"level":"C1","overallScore":8.2,"rationale":"미래 업무 방식 변화가 선명하게 제시됨."},{"segmentIndex":54,"text":"And you know maybe today it is like a sort of intern that can work for a couple of hours but at some point it'll be like an experienced software engineer that can work for days.","startTime":478.72,"endTime":488.24,"durationSeconds":10,"level":"C1","overallScore":8.2,"rationale":"에이전트 성장 비유가 매우 선명함."},{"segmentIndex":60,"text":"Um, I would bet next year that in some limited cases, at least in some small ways, we start to see agents that can help us discover new knowledge or can figure out solutions to business problems that are kind of very non-trivial.","startTime":521.8,"endTime":537.44,"durationSeconds":16,"level":"C1","overallScore":8,"rationale":"에이전트의 다음 단계 전망이 선명함."},{"segmentIndex":8,"text":"Um the thing that matters is the rate of progress that we have seen year over year for the last 5 years should continue for at least the next five probably well beyond that but hard to say.","startTime":643.24,"endTime":655.04,"durationSeconds":12,"level":"C1","overallScore":7.8,"rationale":"핵심은 지속적 진보라는 주장."},{"segmentIndex":10,"text":"Um, all of that said, to me, a system that can either autonomously discover new science or be such an incredible tool to people that our rate of scientific discovery in the world like quadruples or something.","startTime":674.2,"endTime":690.24,"durationSeconds":16,"level":"C1","overallScore":7.8,"rationale":"AGI 기준을 과학 발전으로 제시."},{"segmentIndex":20,"text":"That was like a bit of an aha moment for weight. If you could do this on the entirety of the web car corpus, you of course have search which can figure out which 10 pages to look at.","startTime":755.6,"endTime":765.519,"durationSeconds":10,"level":"C1","overallScore":8,"rationale":"요약과 검색의 연결 통찰이 큼."},{"segmentIndex":26,"text":"Even if you look at many of the post- training techniques that have come, it's a little bit of okay, take this incredibly powerful model, give it context for what's worked, what's not work, and use it to improve what it is producing.","startTime":805.68,"endTime":817.6,"durationSeconds":12,"level":"C1","overallScore":8,"rationale":"후처리의 본질을 맥락으로 설명함."},{"segmentIndex":29,"text":"There's always an infinity of context. Humans solve it by what we call attention where we focus on something.","startTime":826.079,"endTime":832.639,"durationSeconds":7,"level":"C1","overallScore":7.6,"rationale":"주의를 맥락 선택으로 비유한 통찰."},{"segmentIndex":30,"text":"I think of search as a tool for setting attention for a model.","startTime":832.639,"endTime":836.56,"durationSeconds":4,"level":"B2","overallScore":7.6,"rationale":"검색의 역할을 새 틀로 설명함."}],"generatedAt":"2026-06-24T23:33:51.868Z","keyClipsTotalSec":596},{"videoId":"qhnJDDX2hhU","chunkIndex":2,"totalChunks":3,"title":"Sam Altman Talks AGI Timeline & Next-Gen AI Capabilities | Snowflake Summit 2025 Fireside Chat — Part 3 of 3","thumbnail":"https://i.ytimg.com/vi/qhnJDDX2hhU/maxresdefault.jpg","duration":1254,"uploader":"Maginative","youtubeUrl":"https://www.youtube.com/watch?v=qhnJDDX2hhU","keywords":["ai","agi","language-models","compute","biology","rna","genomics","healthcare","science","research"],"normalizedKeywords":["기술 트렌드","교육","엔지니어링"],"targetAudience":[{"who":"리서처","why":"AI를 대규모 과학 연구에 적용하는 관점을 얻을 수 있음"},{"who":"엔지니어","why":"대규모 컴퓨트를 어디에 써야 하는지 응용 사례를 볼 수 있음"},{"who":"학생","why":"AI가 기술을 넘어 생명과학 문제를 푸는 방향을 이해할 수 있음"}],"normalizedAudience":["리서처·학자","엔지니어·개발자","학생·주니어"],"summary":"이 구간은 대규모 컴퓨트를 어디에 쓰면 가장 의미가 큰지에 대한 답으로, RNA expression과 단백질 조절을 해독하는 생명과학 연구를 제시한다. 단순히 기술 업계 내부의 효율화가 아니라, AI와 언어모델을 활용해 인간의 질병 이해와 치료에 직접 기여할 수 있다는 점을 강조한다.\n\n핵심 메시지는 '거대한 컴퓨트의 가치는 모델 성능 향상만이 아니라, 인류의 가장 큰 문제를 푸는 데 있다'는 것이다. RNA가 DNA 발현과 단백질 작동을 어떻게 조절하는지 정확히 이해하면 수많은 질병을 해결할 수 있고, 이런 방향이야말로 AI 시대의 가장 고무적인 활용이라고 말한다.","insights":["거대 컴퓨트의 진짜 가치는 과학 난제를 푸는 데 있다.","AI는 코드 생성보다 생명현상 해독에서 더 큰 임팩트를 낼 수 있다.","RNA 조절 메커니즘의 이해는 질병 해결의 지름길이다.","언어모델은 데이터가 복잡할수록 더 강력한 연구 도구가 된다."],"keyClips":[{"clipId":"qhnJDDX2hhU:c2:1-4","startSegmentIndex":1,"endSegmentIndex":4,"startTime":1201.51,"endTime":1242.72,"durationSeconds":41.2,"preview":"컴퓨트의 최적 사용","mustSee":true},{"clipId":"qhnJDDX2hhU:c2:2-3","startSegmentIndex":2,"endSegmentIndex":3,"startTime":1217.76,"endTime":1231.2,"durationSeconds":13.4,"preview":"RNA와 질병 해결","mustSee":false},{"clipId":"qhnJDDX2hhU:c2:4-4","startSegmentIndex":4,"endSegmentIndex":4,"startTime":1231.2,"endTime":1242.72,"durationSeconds":11.5,"preview":"언어모델의 새 역할","mustSee":true}],"curatedSegments":[{"segmentIndex":14,"text":"I think just do it. Like there's still a lot of hesitancy and the models are changing so fast and there's all this reason to wait for the next model or you're going to sort of like wait and see if this is going to shake out this way or that or if you should build, you know, with thing that thing A or thing B. I as a general principle of technology when things are changing quickly that companies that have the quickest iteration speed um and sort of make the cost of making mistakes the lowest and the learning rate the highest win.","startTime":124.88,"endTime":154,"durationSeconds":29,"level":"C1","overallScore":9,"rationale":"실행 조언과 고급 표현이 매우 풍부."},{"segmentIndex":28,"text":"And so there are lots of application like chat bots whether it's structured or unstructured data the technology is mature you can adopt it yes you can always push the boundary on what else you can do with it there are edgier agentic applications but far away from the frontier I think this technology is actually ready for mainstream use interestingly I wouldn't have quite said the same thing last year um I would have said the same thing to a startup last year but to like a big enterprise I would have say like I would say like uh you can experiment a little bit, but this is maybe not totally ready for production use in most cases.","startTime":245.12,"endTime":281.199,"durationSeconds":36,"level":"C1","overallScore":9.2,"rationale":"기술 성숙도 판단과 표현이 매우 풍부."},{"segmentIndex":62,"text":"And as that expands to longer time horizons and higher levels, you know, at some point you get an AI scientist uh an AI agent that can go discover new science and that will be kind of a significant moment in the world.","startTime":549.12,"endTime":563.44,"durationSeconds":14,"level":"C1","overallScore":8.8,"rationale":"미래 AI의 전환점을 전망하는 고통찰 문장."},{"segmentIndex":9,"text":"And whether you declare the AGI victory in 24 or 26 or 28, um, and whether you declare the super intelligence victory in 28 or 30 or 32 is way less important than this one long beautiful, shockingly smooth exponential.","startTime":655.04,"endTime":674.2,"durationSeconds":19,"level":"C1","overallScore":8.6,"rationale":"라벨보다 성장곡선이 중요하단 통찰."},{"segmentIndex":32,"text":"I think it becomes a matter of debate like Sam is saying in uh I think sometimes it's also a philosophical question that I would liken to I don't know does a submarine swim um at one level it's absurd but of course it does and so I see these models as having incredible capabilities that we will like any person looking at what things are going to be like in 2030 would just declare that's AGI but remember you and I would to Sam's point would say the same thing in 2020 about what we are seeing in 25 to me it's the rate of progress that is truly astonishing and I sincerely believe that many great things are going to come out of it and similar to again how do we feel about the fact that a pretty decent computer can beat every person in the world that can play chess doesn't matter we still have people that play chess that are very good at it so I the definition matters.","startTime":847.04,"endTime":903.04,"durationSeconds":56,"level":"C2","overallScore":8.8,"rationale":"비유와 예시로 핵심 관점을 풍부히 전개."},{"segmentIndex":38,"text":"Um I have to ask you because we have you you're training more models you know you see the next capabilities before anybody else does uh what emergent behaviors are you seeing in the next set of models that change you know how you operate what you want to build from a product perspective how you're running open AI yeah the models over the next year or two years are going to be quite breathtaking um really there's a lot of progress ahead of us a lot of improvement to come and like we have seen in the previous big jumps you know from GPT3 to GPT4 businesses can just do things that totally were impossible with the previous generation of models and so what an enterprise will be able to do we talked about this a little bit but just like give it your hardest problem if you're a chip design company say go design me a better chip than I could have possibly had before um if you're a biotech company trying to cure some disease say just go work on this for me like that's not so far away.","startTime":931.279,"endTime":994.639,"durationSeconds":63,"level":"C2","overallScore":9,"rationale":"미래 역량과 활용상을 강하게 제시."},{"segmentIndex":40,"text":"Uh and these models ability to understand all the context you want to possibly give them, connect to every tool, every system, whatever, and then go think really hard like really brilliant reasoning and come back with an answer and have enough robustness that you can trust them to go off and do some work autonomously.","startTime":997.279,"endTime":1025.76,"durationSeconds":28,"level":"C1","overallScore":8.8,"rationale":"에이전트형 모델의 조건을 잘 묘사."},{"segmentIndex":48,"text":"But the amazing thing is they can reason. And if you think of it as this reasoning engine that we can then throw like all of the possible context of a business or a person's life into and any tool they need for that physics simulator or whatever else.","startTime":1076.16,"endTime":1089.48,"durationSeconds":13,"level":"C2","overallScore":9,"rationale":"모델 역할을 재정의하는 핵심 통찰."},{"segmentIndex":15,"text":"And certainly what we're seeing with enterprises and AI is the people that are making the early bets and iterating very quickly are doing much better than the people that are waiting to see how it's all going to shake out. Straight, what would you say?","startTime":154,"endTime":166.16,"durationSeconds":12,"level":"C1","overallScore":7.8,"rationale":"빠른 실행의 이점이 분명한 조언."},{"segmentIndex":35,"text":"Um and I think we'll be at the point next year where you can not only use a system to sort of automate some business processes or build these new products and services but you can really say I have this hugely important problem in my business.","startTime":314.44,"endTime":329.199,"durationSeconds":15,"level":"C1","overallScore":8,"rationale":"내년 변화상을 구체적으로 전망함."},{"segmentIndex":37,"text":"And the models will be able to go figure out things that teams of people on their own can't do.","startTime":332.8,"endTime":337.68,"durationSeconds":5,"level":"C1","overallScore":7.6,"rationale":"AI의 문제 해결 잠재력을 잘 드러냄."},{"segmentIndex":38,"text":"And the companies that are have gotten experience with these models are well positioned for a world where they can say okay you know AI system whatever go you know like redo my most critical project and here's a ton of compute think really hard just figure out the answer.","startTime":337.68,"endTime":354.8,"durationSeconds":17,"level":"C1","overallScore":8.2,"rationale":"미래 업무 방식 변화가 선명하게 제시됨."},{"segmentIndex":54,"text":"And you know maybe today it is like a sort of intern that can work for a couple of hours but at some point it'll be like an experienced software engineer that can work for days.","startTime":478.72,"endTime":488.24,"durationSeconds":10,"level":"C1","overallScore":8.2,"rationale":"에이전트 성장 비유가 매우 선명함."},{"segmentIndex":60,"text":"Um, I would bet next year that in some limited cases, at least in some small ways, we start to see agents that can help us discover new knowledge or can figure out solutions to business problems that are kind of very non-trivial.","startTime":521.8,"endTime":537.44,"durationSeconds":16,"level":"C1","overallScore":8,"rationale":"에이전트의 다음 단계 전망이 선명함."},{"segmentIndex":8,"text":"Um the thing that matters is the rate of progress that we have seen year over year for the last 5 years should continue for at least the next five probably well beyond that but hard to say.","startTime":643.24,"endTime":655.04,"durationSeconds":12,"level":"C1","overallScore":7.8,"rationale":"핵심은 지속적 진보라는 주장."},{"segmentIndex":10,"text":"Um, all of that said, to me, a system that can either autonomously discover new science or be such an incredible tool to people that our rate of scientific discovery in the world like quadruples or something.","startTime":674.2,"endTime":690.24,"durationSeconds":16,"level":"C1","overallScore":7.8,"rationale":"AGI 기준을 과학 발전으로 제시."},{"segmentIndex":20,"text":"That was like a bit of an aha moment for weight. If you could do this on the entirety of the web car corpus, you of course have search which can figure out which 10 pages to look at.","startTime":755.6,"endTime":765.519,"durationSeconds":10,"level":"C1","overallScore":8,"rationale":"요약과 검색의 연결 통찰이 큼."},{"segmentIndex":26,"text":"Even if you look at many of the post- training techniques that have come, it's a little bit of okay, take this incredibly powerful model, give it context for what's worked, what's not work, and use it to improve what it is producing.","startTime":805.68,"endTime":817.6,"durationSeconds":12,"level":"C1","overallScore":8,"rationale":"후처리의 본질을 맥락으로 설명함."},{"segmentIndex":29,"text":"There's always an infinity of context. Humans solve it by what we call attention where we focus on something.","startTime":826.079,"endTime":832.639,"durationSeconds":7,"level":"C1","overallScore":7.6,"rationale":"주의를 맥락 선택으로 비유한 통찰."},{"segmentIndex":30,"text":"I think of search as a tool for setting attention for a model.","startTime":832.639,"endTime":836.56,"durationSeconds":4,"level":"B2","overallScore":7.6,"rationale":"검색의 역할을 새 틀로 설명함."}],"generatedAt":"2026-06-24T23:34:13.973Z","keyClipsTotalSec":596}]}