He’s Not Joking: 100% Jobs Will Disappear | Monish Darda On AI, Business, Life, Future & Money | TSA

People will still reproduce and the people who are reproduced will grow up and they need to earn. There won't be jobs for people to do. There are dark factories everywhere now. They're probably 5% of all the factories and 95% still employ people. 10 years later, you can imagine 99% of the factories being dark factories. Now, bear with me for a second, a company like Nvidia, 10 years from now, I hit a button. AI takes over my design, analyzes it, finds faults, finds ways to test, finds ways to manufacture, then goes to a factory in Taiwan, negotiate with them. AI will go to potential customers, make money for them. If a multi-million dollar enterprise deal gets screwed up because AI was doing the negotiation, who is to be held accountable? It feels brutal. It also feels, you know, maybe I smoke something.

I wouldn't call it clone. I would definitely want to call it an avatar.

Steve Jobs can be running Apple forever.

So uh it's interesting, yes, I am too old? I will be 62 next month. That's always a blessing. But it's interesting, I think as long as you're not mentally old, as long as you're thinking about new things, as long as you love building and creating, I think you can be a founder.

That's a tough one. I don't even know fully myself what I am. I'm still figuring this out. And it's a great journey, by the way. But I think I can tell you what I love, right? I love building things. I love reading a lot of fiction. I can start especially with Isaac Asimov, which is one of my favorite authors. I can read them back to back for days on end. I love that. I love games like Call of Duty and Assassin's Creed and the Ghost of Tsushima. So video games are a big fancy. And from a professional perspective, I've always been a startup person. Right after I finished my schooling, did my masters in the US, came back, spent a few years going after my passion of cars and engines and two wheelers, and then I started my first company with a co-founder. After that, I've always been doing startups. I think now I'm on my ninth probably. And I love doing that. What is very interesting though is AI is also changing a lot of this. I turned a milestone last week: in the last 9 months I have written 1.5 million lines of code or generated. I have been the most productive programmer in my life, and I owe this completely to AI. I think that gives me a very interesting edge that tells me that I can still be a founder. If I can do that and if I think personally that this is some of the best code that I've ever written, I'm not too old to be a founder.

Very interesting question. So first of all, if my daughter heard it, she's a neuroscientist, she would be really upset because dumb would be a very different definition in neuroscience. The second thing, let me ask you a question. Have you used a typewriter? You didn't have the pain of having to figure everything out before you typed it because you screwed it up. You had to type the whole thing again. You could always do a backspace. Were you a better journalist? I have always given this a thought. I think over years, perfection is slowly diminishing in human beings because there are ways to correct mistakes. While you were typewriting, I would see my grandfather writing the perfect letter. But does that make you a bad journalist? No, I was one of the better journalists. See, there you go. So I think all of this technology is actually a tool to what you have to achieve. When I was in ninth standard, the first calculators appeared and everybody in school said if you use a calculator you get dumb. You have to make sure you know the math. I argued with my teacher saying that it actually makes my math better. This is a tool. Now would you actually teach a child to not use a calculator and go back? No. You would actually give them computers, you would give them ChatGPT. You would give them bigger problems to solve at a younger age. In my mind it makes them smarter because at our age the problems we were solving and at her age, 2 and a half years, my granddaughter is solving, are completely different levels of problems which 100% guaranteed proves to me that generations are getting smarter and achieving more. They're not getting dumber. They're not getting lazy. They are actually packing a lot of things in shorter and shorter periods of time. I'll give you another example. IBM took decades to get to a billion dollars of valuation. Whatever they peaked at, it took decades to build. Facebook came in and said we can do it in 10 years. WhatsApp came in and said we can do it with 11 or 13 people. 13 people sold at $19 billion. It's crazy, right? Same thing with AI. I don't think ChatGPT is ever going to make anybody dumb. It's going to make them smarter and achieve more and do more in their life. That would be the sole purpose of AI. I keep on hearing this number 80% of jobs, 70% of jobs will go away. I used to throw that number away saying that's bull crap, but you seem to be painting a different picture. I think 100% of the jobs will go away in the next 10 years. There's a reason for that. So then you and me will keep doing this conversation, that's the only thing we'll do. Maybe opening companies, deploying AI, figuring out new things, having creativity, having a ball. But in all seriousness, this is going to be a huge problem and also a huge opportunity. Think about it this way. What are the biggest challenges in the world today? World hunger, how do I feed everybody, shelter everybody, clothe everybody? The second is health longevity. The third is what do these people do? If I look at AI, the intent of AI will be to take human intent and convert it into outcome.

That's a very interesting question and again this is a hypothesis. Nobody really understands this fully. But here's the take. Think about today what Nvidia does. The most successful or most highly valued company in the history of human civilization. How much do they manufacture? Hardly. They don't manufacture much. They are a design house. They are a thinking house. They are a sales and marketing machine. They subcontract the designs. Can you imagine a company like Nvidia 10 years from now? You and I say we want to design this chip. Here are the specs. Here is the creative leap that my mind has made because I've been working on this problem for 20 years. I come up with a design. I hit a button. AI takes over my design, analyzes it, finds faults, finds ways to test, finds ways to manufacture, goes to a factory in Taiwan, negotiates with them. AI will go to potential customers, say I have this new chip, are you interested? Negotiate all my contracts, set up the supply chain, actually deliver those chips, make money for me, deposit that money in my account. Meanwhile, both of us would have gone and formed another company. Is it completely bizarre to think that companies would work like that? I don't think so. At least parts of this I can see happening.

We still figuring it out. I'll flip the question. If an autonomous car ends up killing a person on the road, who is responsible? It is that kind of question. Finally, there will always be humans who are in control. At some point, I also believe humans and AI become one, the neural link chip. The twin shall meet. But till then, humans are going to be responsible. It's not the tool. I'll give you another example. Consumer electronic show. Last year they had a laundry folding robot. It used to take 3 minutes to fold a shirt. This year that same company came up with a new version which folds laundry perfectly without dropping it in less than 30 seconds. Efficiencies are being built. So think about factories. There are dark factories everywhere now. They're probably 5% of all factories and 95% still employ people. But that is going to be rapidly changing. 10 years later, you can imagine 99% of the factories being dark factories. Can you not? If you can drive an autonomous car in San Francisco traffic at peak hours, you can run a dark factory to manufacture shirts and watches. Of course you can.

Bear with me for a second. If I look at electricity, right, we have been struggling with cold fusion. But what is that? The same mechanism that our sun uses to generate so much power is nuclear fusion. If I can fuse two atoms, I can produce a lot of energy without the bad things that nuclear fission causes. Nuclear fusion eliminates the radioactivity problem. That's been the ticket to free energy for humanity. The second problem is longevity. Last month, the first ever drug that claims to decrease your biological age went into clinical trials in the US. The first time ever. The third thing, one month ago, Google announced they are rolling back the calendar for practical quantum computing from 2035 to 2029. What are these three things indicating to me? They're telling me that as AI becomes ubiquitous, as scientists start using AI, the ability to write code at volume, analyze data, run experiments, simulations is accelerating all sorts of science very quickly. Next 10 years, if I can get nuclear fusion right as a practical way of creating power, my power problems are gone. So a lot more AI work can be done. There is a huge debate about AI increasing energy consumption, but free power for AI production. There has to be an end to what people will do if they grow up. That is a question that will face humanity, just like it faced humanity on communications or mobility, except now at a much bigger scale.

So you're saying that the pace at which science is evolving, if we manage to live for another 10 years, then everything that is happening in the space of medicine, biology, longevity will allow us to live till 120 years. In a capitalist society, if we have that kind of money, it will allow us to live that long. But think about 10 years later, if this happens, will money exist? Money of course will exist because you will need money to live for 120 years. Without money you would not be. The value of money will even increase even more because it will create a larger societal divide. There will be a large poor population, a very skewed rich population. The middle class will become even more sandwiched because longevity cannot be given to everybody. That's the interesting question to answer. I'm writing a book which will be released by the end of May called The DeepSeek Moment. When ChatGPT was released, you needed at least 300 million to train a foundation model. Suddenly out of nowhere, a company of like five people in somewhere in China, a hedge fund, came up with the DeepSeek model. They changed the way you need compute, changed the way you do inferencing. They released this with open weights to the world. So if you have a country of 1.5 billion people, you cannot say I will only make 20% of the population live for 120 years if they give me so much money. You can't do that. It's the end of the country.

I think and again I'll pull in Isaac Asimov a little bit. I think the big change that will happen is creativity. I think a direct impact of having a lot of power, no money, power as in electricity, and abundance and longevity, is that our birth rates are going to decline like we have seen in Japan and China, and even in India as urbanization grows. Lesser birth rates, people hate meeting each other in person because they think there's no need. Then there are holograms. So it feels like you are still in the room, but you're not. Then they become paranoid of being in the same room with a person. Because if you're in the same room with...

A person now look at this, it is actually real. This is how I would think 10 years from now if I didn't know better. You're bringing in a lot of bacteria my body is not used to. I have only one life if I don't believe in rebirth and reincarnation. You bring in a bacteria, I'm taking a risk of dying, where in an age where I can live for 300 years and maybe I'm 70 years old, 62 years old. Why would I take the risk?

You could, but you could introduce a bacteria that I don't have a medicine for and don't have the time to build the medicine for either. Fear,

Paranoia, and fear. And so ASimo is beautiful that way. He describes a world where viewing is the only way to meet a person. When you say I want to see a person, you're actually meaning you are viewing the person, not seeing them in person. And everybody breaks down. The first thing to break down is children because there are no children. So then the government has to enforce people to make children for the goodness of civilization. So then it's a forced job to make children because they're not able to make them in a laboratory, which probably would have happened or will happen in years.

But that's the world that we are going towards in my mind. And so, problems actually find their own solutions. I'm a big believer of that. So when you have these problems, what do I do? No money, no longevity, no jobs, what do I do? I think the solutions will come by themselves.

Mhm.

Almost the same answers would come to everybody if I give a prompt. What differentiates in that case? That brings me to my new venture a little bit.

Human

But I'm also good at analyzing these patterns and making connections. So I'm going to make this connection if you permit me.

To find out

Okay. Now this is very important. I think the way you phrase the question, the way I phrase the question, because of our nature and nurture, that phrasing is going to be different, which means the output is going to be different. And remember with large language models, we are still at the tip of the iceberg. We have not seen the iceberg below the water. This is just the start of that wave and the next five or six years are going to be extremely different, very different than what we can imagine today. And these models are going to be even more powerful, but they are going to be influenced by the prompts and the knowledge that is locked in your brain or your email or your documents. So you can hyperpersonalize the model.