저자: KenGriffin · 외국인
요약 보기
생성형 AI는 머신러닝의 연장선이며 투자 업무에서는 생산성 도구 수준이라 게임체인저는 아니라고 평가. 과거 데이터 기반 모델은 정적 문제에 강하고 미래 예측(투자)에는 약하며 단기 매매에만 유효. 콜센터·번역·문서 업무 등 고임금 화이트칼라 일자리 감소가 사회적 비용으로 이어질 것이라 경고. 마케팅 개인화 등 새로운 활용이 향후 10~20년 게임체인저 기업을 만들 것으로 봄.
다듬은 글
bring us to AI. Look, generative AI has just like gripped the world both in mind share impact and some degree hype. Let's let's go a little bit back through the history books. Generative AI is another branch in the holistic machine learning tree. And machine learning has been around since at least the 90s.
I mean, when I was younger, you know, I we took a look at this in the 90s. Let's be right to the point. What an interesting curiosity. Not nearly enough computational pro power to actually create machine learning models of any note. And it the whole area just sort of disappear into the dust bins of history until Google.
Google harnessing roughly the computational power of the fifth largest country in the world. Got that? One firm, fifth largest country in the world. decided to revisit machine learning for purposes of optimizing the search monetization for Google users and it worked and from that came TensorFlow which which Google gifted to the world and it's one of the greatest gifts ever given to humanity.
It was the reintroduction of machine learning at large to the broader population. We I I remember my colleagues were in my office when TensorFlow the day it was announced, they were so excited about this. They're like, "This is going to change the world." And about two weeks later, the most senior partner in the room I called.
I said, "So, like, any thoughts on on this whole machine learning thing yet?" He's like, "Thoughts? It's it's it's in production. We use it to trade equities every day today." And I'm like, "It's it's been two weeks." He goes, "It only took 10 days." But in 10 days, it took 10 days from TensorFlow to be to be a part of the decision-making in almost one in four trades that happens every day in the US stock market.
All right. And I I give you this bit of history because generative AI is just another step in the journey of the use of machine learning technologies by modern society. Do we use it in our investment business? A little bit. A little bit. I can't say it's been gamechanging. It saved some time. It's a productivity enhancement tool.
It's nice. It won't I don't think it's going to revolutionize most of what we do in finance. Why is it these models are based on what has happened to thus far to date in the history of humanity and investing is about understanding what's going to unfold tomorrow or next year or in two years. That's not really the basis by which machine learning models holistically are trained.
machine learning models are really good at at when you have an underlying data set that that the relationships within that data set represent what happens holistically today and tomorrow and so on so forth in the future. So I'll give you a simple mental model. Self-driving cars historically have worked really well in the south.
Clear skies, no snow. Got that? Snow reres havoc with self-driving cars because it changes the contours, the lines of what the car has to recognize in terms of staying on the road. All right, so machine learning models work really well with problems that are more static in nature, reading a radiological report.
But investing is about understanding how the future's going to unfold. And that's where these models really struggle. All right, they work great in short-term trading. In short term, I mean as in like the next five minutes. But when you think about the next year or two years, they really start to fall apart.
Like that's just not what they do. So having said that, it's going to change the world around you. It's going to change the world around you in a lot of profound ways. Call centers as we know them today will be extinct in 5 to 10 years. Like gone. I I mean I have friends who are who have businesses that have big call centers.
Today you call the human on the other side talks to you but every prompt is now coming from a from a generative AI model and over time it's getting better and better and better and they're going to push their customer base into Slack and email as the mode of communication. They're going to actually they're going to they're going to try to take the human out of the loop and it's going to have a profound change.
For example, call centers. Could you imagine doing language translation? So friend of mine has a multinational. They sell products in 10 and some countries. They have roughly 8,000 people in documentation. They think it'll be a thousand within three years. And those are really good highpaying white collar jobs.
And I got to tell you, it's going to be painful for them to find employment from from the skills that they have now, which is highly specialized, highly valued. they're going to have to pivot midway through their careers or even late in career to other roles. So machine learning is going to come with a cost to society.
A cost that we need to understand how do we help these people land on their feet so we don't end up with a backlash against AI and machine learning. Uh other areas that it'll come into play, I mean marketing to one, you know, a friend of mine uh in his 20s ran ran a pet insurance business. All right?
And they can look at your Facebook page and when you post a photo of a new golden retriever, they can hit you with a direct ad. Congratulations on the new golden retriever that's in your life. But nothing's more heartbreaking if that dog gets sick and we have the insurance pro solution to meet your needs.
got that marketing to one like it's gamechanging and one of the fastest growing pet insurance business in the world and they sold it for a lot of money because of this innovative use of generative AI marketing to one. So I think the really interesting generative AI stories are going to be when people think about how to use these tools in radically different ways than we currently use software and and those will be many of the gamechanging businesses of the next 10 to 20 years.
So it's going to be incredibly exciting. It's going to come with a cost to society. As a society we need to think about how to manage the cost for those who are adversely affected and those people will be very adversely affected in some cases.
Comments 0