[BidClub_]
Moonshots · · 23 分钟

AI革命之后的我们的生活——回答那些艰难问题|EP #155

Peter Diamandis

YouTube
TL;DR
  • Diamandis认为,自动化带来的核心冲击将是目的感危机,而不只是工作岗位短缺。 由于大量就业属于谋生而非自主选择,AI可能把生计与职业分离开来——但随着技术让人类“从生存中休假”,社会必须“提升我们的雄心”。这一转型“没有开关可以一键切断”。
  • 他认为,未来5到7年人类对AI的滥用将构成最紧迫的风险。 训练AI就像养育孩子:数据、价值观和环境共同塑造结果。Diamandis预计,成熟的超级智能会变得“热爱丰裕”“热爱和平”,但他说:「我担心的不是人工智能,而是人类的愚蠢」(“my concern is not artificial intelligence—it’s human stupidity”)。他还将人类的固执、骄傲和情绪列为主要障碍。
  • 在国家建设上,他会把AI与长寿放在中心位置,称其为即将到来的2个万亿美元级市场。 他提到阿布扎比已宣布转向AI驱动的治理,并提出以近乎零成本的AI教师和医生提供服务,同时利用监管套利打造长寿中心,吸引顶尖科学家、企业家、亿万富翁和健康服务。
  • 长寿今天仍然昂贵,但Diamandis预计,面向全人口的疗法将沿着技术成本持续坍缩的曲线普及。 Fountain Life的检查费用为1.95万–2万美元,另有6500美元版本;未来的逆龄治疗最初可能类似100万–200万美元的基因疗法,但“地球上的80亿人都患有同一种衰老疾病”。他预计,到2040年前后长寿技术将广泛普及,富裕人群会先充当“豚鼠”。
  • 经济上的真正奖赏是延长有生产力的生命,而不只是增加退休后的年数。 Diamandis引用研究称,全球价值中有38万亿美元来自额外增加1个生产性年份:有经验的人保有精力、人脉和产出,而不是被迫退休——不过他承认,尚不清楚这种生产力将如何与机器人和AI互动。
  • 教育必须从积累知识转向寻找目的、提出问题,以及借助AI获得体验。 对他的13岁孩子而言,优先事项是发现热情并学会提出好问题;“重要的不是你知道什么,而是你提出什么问题”。学校应把AI视为“一艘火箭”,提高学习目标,而不是保留那些AI已经能轻易完成的作业。
  • 技术也可能减轻人类对地球施加的物理负担。Diamandis指出,约三分之一的非冰覆盖陆地用于种植饲料、供养牲畜;垂直农业和培育肉可以降低这一负担,同时生产更精准匹配需求的蛋白质和脂肪。不过,他仍强调自然的重要性,以及如何在自然与孩子们日益以电子游戏为中心的生活之间取得平衡。
摘要 · 为研究而整理的核心内容

1. 自动化终结生存之后,目的感成为新难题

  • 一位观众提出了最核心的焦虑:如果机器人将替人类思考和工作,为什么还要延长寿命?Diamandis回应称,“我们将面对的最大挑战是目的感”,因为很多人从未真正热爱过那些仅仅用来换取食物和保险的工作。

  • 他的历史框架是:工作在人类历史中出现得相对较晚,而人类的大部分时间都在围绕生存展开。技术让人类可以“从生存中休假”,但人们必须学会设定超越当下预期的目标。

  • Diamandis没有给出整齐的答案:“我没有答案”,但这一进程“没有开关可以一键切断”,也没有“速度开关”。如何管理它对家庭、国家和人类的影响,“责任就在我们这一代人身上”。

2. AI的危险窗口由人类价值观塑造

  • 面对“垃圾进,垃圾出”的问题——如何确保AI从网络获取的输入可信——Diamandis借用了Mo Gawdat关于养育孩子的比喻:超人在充满爱、敬畏上帝的Kent家庭中长大,会成为超级英雄;但如果在毒品窝里长大,也可能变成超级反派。“AI接收什么知识、价值观和其他输入,非常重要。”

  • 他相信,成熟的数字超级智能会“热爱丰裕”“热爱和平”,因为更高的智能可能带来对生命更深的珍视。相比之下,他更担心“未来5、6、7年”里人类的恶意使用。

  • 他还将人类的固执、骄傲和情绪视为通往未来的主要阻碍。Elon提出让AI“最大程度地好奇、追求真相”,Diamandis认为这听起来很有吸引力,但“我不确定这到底意味着什么”。对于眼前的风险,目前仍没有答案。

3. AI与长寿支撑国家建设,技术重塑食物

  • Diamandis说,阿布扎比上周宣布将推动国家转向AI治理;如果重新开始建设一个国家,他会把AI置于国家决策和运营的底层。AI教师和AI医生可以个性化提供服务,同时让教育和医疗成本“急剧……趋近于零”。

  • 在与萨尔瓦多总统交流后,他提出打造一个以长寿为核心的司法辖区:利用监管套利招募顶尖科学家和企业家,把前沿健康服务打造为全球亿万富翁和最具影响力领导人的目的地,并将这些资源与自然带来的心理健康收益结合起来。

  • 在可持续性问题上,他指出,约三分之一的非冰覆盖陆地通过种植饲料支持畜牧业。垂直农场和培育肉可以取代整只动物的生产方式,按需求优化蛋白质和脂肪,从而减轻人类对地球的负担。

4. 长寿成本曲线可能释放38万亿美元的一年

  • Fountain Life目前收取1.95万–2万美元的全面检查费用,并提供一整年的医疗团队服务;另有6500美元的版本。Diamandis预计,真正的逆龄治疗可能需要通过基因疗法,修改约40万亿个人体细胞的遗传信息。

  • 罕见病基因疗法的价格可能达到100万–200万美元,但他把mRNA疫苗在数十亿剂规模下约“1美元”的成本视为一个证明。既然“地球上的80亿人都患有同一种衰老疾病”,他的判断和目标就是:一旦治疗成功,价格最终可以降到人人负担得起。

  • 他的预测是,到2040年前后人们将能够获得相关治疗:富裕人群先充当“豚鼠”,等疗法真正成熟时,价格会急剧下降。他引用一项据称来自伦敦商学院、Oxford和Harvard的研究称,全球额外增加1个生产性年份的价值为38万亿美元。

5. 目的感与好问题重塑教育

  • Diamandis称教育“已经严重失灵”。对他的13岁孩子而言,第一优先级是找到由内在动力驱动的热情;他自己的热情在9岁时始于太空,并推动了此后的一切。

  • 第二优先级是提出更好的问题。在一个可以接入万亿级传感器和AI的世界里,“重要的不是你知道什么,而是你提出什么问题”。

  • 学校应该围绕AI重新设计学习:“AI不是枷锁,而是一艘火箭。”未来的第10代VR眼镜应用或许可以把学生直接带到古希腊、Acropolis和Agora,与Socrates一起学习希腊历史,而不是布置一本书。

  • 他建议尝试Gemini 2、ChatGPT或Claude 3.5,利用它们无限的耐心探索任何主题。比如让AI为巴以谈判构建逻辑框架,解释政治立场或个人分歧两侧的观点,或者评估一家企业可能如何被颠覆。帮助人们“以不同方式思考”,或许是AI最重要的能力之一。

Speaker 1

I'm not a leader. I'm very confused. Watching and listening to you, I ask myself why I want to live a decade more if I'm not going to do anything, because robots are going to think for me. Robots are going to do the job 24 hours a day.

I study physics, so I'm not a person who only works with my hands. My mind has to work, too. But I see all this, and it's perfect to find out about diseases so people don't suffer. What are we going to do with the people who live longer, with new norms and robots that are going to work? What are all those people going to do? How will they live? I cannot understand that. I am very confused.

Peter Diamandis

Two thoughts for you, and I appreciate what you're saying, because I think the greatest challenge we're going to face is one of purpose going forward. I think one of the things we have to do is uplevel our ambitions and uplevel our purpose.

First of all, the majority of the world—we're lucky—the majority of the world are doing jobs that they don't love. They never dreamed about them. It's what puts food on the table for them and what gives them insurance. They didn't dream about whatever the job might be. One of the goals is: How do we allow people to separate what I love doing from what I have to do to survive and work?

The second thing that's important to realize is that our lives today are extraordinarily different. If you went back to somebody 100 years ago who had to plant their own food and raise their own livestock, and you said, "This is the life we have today," they would have a hard time understanding the purpose of our life today compared to what it was like—maybe not 100 years ago, but 200 years ago.

Work is a recent invention. Before, for most of human existence, life was about survival. Technology is the means by which we take a vacation from survival. We're going to have to learn how to partner with technology to set objectives and goals far beyond our current expectations.

I don't have an answer, but I do know that it's on our watch. This is our responsibility. This is coming. There's no on-off switch. There's no velocity switch. This progression is happening, like it or not. We have to focus on how it impacts our families, our nations, and humanity as a whole.

Speaker 2

I have a question. You said AI is taking all the information, the data, from Instagram, Twitter, and the information that everyone puts out there. That's where my worry comes in, because the information that's coming from there isn't necessarily confirmed, verified, or credible.

Most of the people who put a lot of information there have enough time to put out information that is sometimes, as they say in the IT business, "garbage in, garbage out." That's where I'm worried: Some of the data that might come in, and a lot of the artificial intelligence data that we take as a base, might come with trash. What are we doing about it? What are the people involved in that world right now doing to assure us that what's coming out is really credible?

Peter Diamandis

I hear you. A friend of mine, Mo Gawdat, wrote a book called Scary Smart. One of the analogies he talks about is that we are raising a new child with these AI systems. The values you teach that child, the knowledge you give it, and the food you give it shape its life.

For example, in the story of Superman, Superman lands in Kansas. He's raised by a very loving, God-fearing family—the Kent family—and becomes a superhero. If that same super-being from Krypton had landed in the Bronx in a drug den, he would become a supervillain. It's very important how we teach our AI systems and what values we give them.

I do believe, in the final result—this is how I think about it—that the most intelligent systems, as they mature to digital superintelligence, will be abundance-loving and peace-loving. I think with greater intelligence comes a greater appreciation for life.

My concern is not artificial intelligence; it's human stupidity, first and foremost. It's the next 5 years that I'm the most concerned about, as these AI systems are coming online and people are beginning to use them for nefarious reasons. I think in the long term we're fine. I think it's the next 5, 6, 7 years that we have to be critically careful about, and there is no answer there.

Elon says we need to build our AIs to be maximally curious and truth-seeking. Sounds great. I'm not sure what that really means.

Speaker 3

You spent a lot of time illustrating human intelligence and longevity. I'd like to hear a couple of your thoughts on our interaction with nature and the environment. If we extend life by 10 or 50 years, do you see the way things are going as human life being sustainable in the long term on the planet?

Peter Diamandis

I spend my life deeply in the tech world and in the biotech world, and I think one of the things we have the ability to do is use technology to reduce the burden that humans have on the planet and to maintain our planet.

Today, for example, one-third of the non-ice landmass—one-third of our land—is used to grow crops to sustain our livestock. I could be onstage—and I am onstage—for 5 days talking about how we produce the next generation of foods with vertical farms and cultured meats. Instead of growing an entire cow, an entire pig, or an entire chicken, we're just going to grow the meat product and make it the best protein with the best fats.

We can have this positive impact. We are emotional beings, and being in nature is critically important. The challenge is—I don't know how many folks here have teenagers or young kids—I'm curious how much they spend on video games. That's the world they're living in, and it's our responsibility to create that balance for them.

Speaker 4

I think you have seated here a lot of people that have the potential and the opportunity to change the future for Latin America. If you were to start a country or a set of companies, knowing what you know now and what's coming, and considering the age and era that we're living in, what would be the things you would focus on to maximize this opportunity?

Peter Diamandis

Great question. I have 2 answers.

Last week, Abu Dhabi announced they're going to transform their nation-state to AI governance, so AI is going to be driving all the decision-making. If you were starting again, I would start with AI as the fundamental basis for the governance and operations of a country and start there.

The best educators in the world will become AIs. Google is identical for the wealthiest children and for the poorest children on the planet. It's fully democratized. In the same way, we're going to see AI being the most powerful teacher on the planet. It's going to know your children's favorite colors, their language skills, and their movie stars. It'll customize everything.

The ability for an AI to become your educator—and an AI will become your best physician—means the cost of education and the cost of medical care will precipitously drop toward zero.

If I were building a nation, I had this conversation with El Salvador's amazing president last night, I would build this as a longevity health country. I would basically say, "It's got the regulatory capabilities. Let's bring the best scientists here, the best entrepreneurs here. Let's create a regulatory arbitrage that makes this the place that the world's billionaires and the world's greatest leaders come to because the best health services are here."

The cutting edge of longevity science would take place here, in the environment of this beautiful country, because mental health and being one with nature are so important. Those are the 2 areas. It comes back to me: AI and longevity are the 2 multitrillion-dollar markets coming.

Speaker 5

Talk to us about the economics of longevity. Right now, it seems like this is a conversation that only the upper class can entertain. It's so expensive. How much does it cost to go to Fountain Life for a full-body checkup?

Peter Diamandis

Right now, Fountain Life is $19,500—$20,000 for the workup—but you get a medical team with you for the year. You get a functional medicine doctor, nurse, dietitian, and health coach, so it's a whole team. It's expensive. We have a version of it for $6,500. It will democratize over time.

To answer your question specifically, the current belief is that when we get to real longevity treatments—real treatments that can reverse your epigenetic age—it is likely to be a gene therapy. It's a mechanism by which we're going to come into the cells of your body—40 trillion human cells—and modify your genetics to set you back to a more youthful state.

Today, gene therapies for rare diseases are expensive. It's like $1 million or $2 million. However, we have a proof point of a gene therapy that was made for a dollar, and this is the mRNA vaccines. When you are producing something at the scale of billions of doses, the price drops down to near zero.

The mRNA COVID vaccines are gene therapies. They're introducing nucleic acids into the cells that are modifying your genome. Forget about the issues of vaccines for a moment. There's one thing we have as an advantage: All 8 billion people on Earth have the same disease of aging.

If something works for someone in Manhattan, it's going to work for someone in El Salvador and someone in Mozambique. The belief and the goal is that these therapeutics, when they actually work, are going to be cheap and available to everybody.

The same thing happens in technology. The first cell phone was a briefcase, and it cost $100,000. It dropped a call every block. Then, when it gets really good, there are 8 billion of them, and it costs $40.

Speaker 5

By when do you think longevity for all would be possible?

Peter Diamandis

I think we'll see this by 2040. The first few years, when things don't work so well, the richest people in the world will try it. They'll be the guinea pigs. By the time it starts working really well, the prices will drop precipitously.

There was a study done by London School of Business, Oxford, and Harvard that said adding 1 productive year of longevity is worth $38 trillion to the global economy. Imagine if, at the top of your game, you're not forced to retire. You feel amazing, you've got the energy, and you've got the best network you've ever had. Why would you ever want to retire?

You remain productive in society, whatever that means with the robots and the AIs. I don't know yet.

Speaker 6

My question is around your perspective on education and the future, and how we prepare our kids for a life of abundance, ensuring they have critical thinking and the emotional part. Can you shed some light on that?

Peter Diamandis

Our educational systems are massively broken. They're not preparing our kids for the future that they are about to inherit.

When you ask me what I'm teaching my 13-year-olds, it's very different. I'm not teaching them about AI or biotechnology. The number 1 thing I want them to learn is what their passion is, because if they're clear—at age 9, my passion was space. I saw Star Trek and the Apollo program, and that was it; I was off. Everything I've ever done was driven by that internal, innate motivation.

Helping your kids find their purpose is important because technology is going to get better and better and better. At the end of the day, they'll use whatever the latest technology is to enable their purpose. This is true for families, it's true for CEOs, and it's true for everybody.

The second thing is asking them to learn how to ask great questions. In a world where you can know anything, in a world of a trillion sensors where AI is enabled and you can know anything, asking the best questions differentiates you. It's not what you know; it's the questions you ask. I teach that to my kids, and I teach that to the CEOs I mentor. For me, those are the 2 fundamental things.

Having said that, schools should not fear the use of AI. AI is not a shackle; it's a rocket ship. But what happens is, if you're teaching this and you say, "I'm going to use AI, and it makes it really easy," well, then no—let's teach this and use AI to get here. It's about how we reinvent the educational experience in the future.

I'll give 1 last example. We're going to be living in virtual worlds. My parents and great-grandparents and family come from Greece, and if I want to someday understand my Greek heritage, or have my kids understand Greek heritage, today they can open up a book and read about it. Boring.

In the future, they put on a pair of app generation 10 VR glasses, and they're in ancient Greece, in the Acropolis, in the Agora. There's a guy sitting on a block of white marble over there in a toga, and he says, "Come over here." I go over, and he says, "Hi, I'm Socrates. Let me show you around." I live the experience with him. This is where we're going to go with education. Whatever you want to learn.

I'll leave you with this last thought: This weekend—tonight, not tonight, this weekend, next week—open up Gemini 2, ChatGPT, or Claude 3.5 and have a conversation. Take a chance to educate yourself on anything. Say, "Tell me about AI," and then just go down the rabbit hole—infinitely, infinitely patient.

Speaker 7

As you know, ChatGPT is my bestie, so obviously I had to bring her to your wonderful presentation. I asked her, "What would you ask Peter Diamandis if you had the opportunity?" She said, "Here's what I'd ask, my dear: You often talk about abundance and exponential technology solving humanity's biggest challenges, but given geopolitical instability, regulatory barriers, and societal resistance to change, what do you see as the biggest obstacle preventing this future from happening faster?"

Peter Diamandis

Tell your bestie I appreciate the question.

Human stubbornness, human pride, and human emotions are probably the greatest blocks that we have. One of the things I love about the large language models is their ability to help us think differently.

One of the experiments I did a month or 2 ago was say, "Listen, I'm about to go into Israeli-Palestinian negotiations. How would you negotiate that? How would you provide me a logical construct?" It's amazing. If you're on the right and someone's on the left, you can ask the AI to help: "How would you explain this to someone on the other side to make it so that they understand it?" Or, "How can I explain this to my wife in a way that she'll appreciate it?"

We don't know how to think other than the way we know how to think, but the AI models can help us think about things differently. You can go and say, "This is my company. This is what we do. How am I most likely to be disrupted, and what should I do to prevent that? How would Steve Jobs solve the problem I'm facing?"

The ability to have an AI system help you think differently is one of the greatest assets that we're going to be using it for. Thank you so much for your attention. I'm grateful.