Elon Musk's Viral Interview: Global GDP to Increase Tenfold in 10 Years Without WWIII, Humanity to Be Marginalized by AI

Elon Musk: AI has self-improved. Optimus Gen3 will be mass-produced next year. The future economy may face deflation.

On March 11 local time, in an open dialogue at a technology summit called "Abundance Summit", Elon Musk, the founder of Tesla and xAI, talked about the progress of AI, the mass - production rhythm of Optimus 3, and the economic form after the "singularity".

In this interview, he gave several clear judgments: AI has entered the self - improvement stage, and humanoid robots are about to enter large - scale production. The economy after the AI "singularity" is difficult to predict, but he bets on deflation and "money no longer matters".

When asked about the schedule for the construction of the data center related to SpaceX, Musk did not elaborate on the grounds that "SpaceX is in a quiet period".

"Optimus 3" is about to go into production: it may enter large - scale production next year

In terms of humanoid robots, Musk revealed that Tesla's "Optimus 3" is close to completing development.

He said, "We are finishing the final stage of Optimus 3. It is probably the most advanced robot in the world at present, and no other product is close to its level."

According to his statement, Tesla plans:

Start production this summer

The initial production volume is low

It is expected to enter the high - volume production stage next year

Musk emphasized that robot production also follows the common S - curve climb in the manufacturing industry: slow at the beginning and then rapid expansion.

He said, "The production volume in the manufacturing industry usually follows an S - curve. It starts slowly and then climbs rapidly."

Meanwhile, Tesla is also designing a new robot production factory. He revealed that the design of this factory is significantly different from traditional factories. The future goal is to continuously update the robot version, "It is possible to launch a new robot design every year".

AI has entered the "recursive self - improvement" stage. In the face of AI, humans will eventually be marginalized

In terms of the development rhythm of AI, Musk's judgment is also radical.

When asked whether AI has entered the "recursive self - improvement" stage, he said, "Actually, this situation has been going on for some time."

He explained that the current development of large - scale models has formed a cycle:

The new model is trained with the participation of the previous - generation model

Humans are still supervising

But the degree of participation is decreasing

He said, "The role of humans in the loop is becoming less and less. Each generation of models helps to build the next - generation model."

He predicted that this process may soon achieve a higher degree of automation: "Fully automated self - improvement may appear by the end of this year, at the latest, no later than next year."

In his view, the breakthrough of AI has entered an accelerating stage: "Now, before I go to sleep at night, I see an AI breakthrough, and when I wake up in the morning, there is a new breakthrough."

Musk also discussed in the dialogue the vision of using solar energy to expand the global economic scale by a million times. At the same time, he pointed out that even if only one - fifty - billionth of the solar energy is used, the intelligence generated will far exceed the total intelligence on Earth. In the face of that level of AI, humans will eventually be marginalized. See the video below for details

After the AI singularity, "money no longer matters"

When talking about whether the system can keep up with the impact of AI and robots, Musk used the "singularity" as a metaphor: "The reason why it is called a singularity is that it is difficult to predict what will happen."

He mentioned that the logo of Grok is "the halo around a black hole" and said, "It's hard to know what happens inside the singularity, but it will be very interesting."

He directly attributed the inflation/deflation logic to the explosion of supply: "We will have a universal income, which is basically giving people money." The reason is that "the output of goods and services will far exceed the money supply", thus forming deflation: "Deflation is the ratio of output to money supply... If the growth rate of goods and services exceeds the money supply, you will have deflation."

Looking further ahead, he believes that the importance of money will decline: "Money will become unimportant at some point in the future."

He even proposed a pricing method for a "non - human economy": "I think future AI will not use human currency. It will only care about energy and mass - wattage and tonnage."

As long as there is no third world war, the global GDP will grow tenfold in the next 10 years

In terms of macro - judgment, he gave an optimistic range with a clear premise: he believes that there are "a series of possible outcomes in the future, not all of which are good", but "it is very likely to be good", with a probability of "maybe 80% or even higher".

If there is no extreme exogenous shock, he said he is "quite comfortable" with the economic growth judgment: "Without a third world war... I think it is a quite safe prediction that the economic scale will grow tenfold in 10 years."

Employment and "robots building robots": no layoffs, instead, recruit more. The output per person will be "extremely high"

In response to the question of "when robots can participate in manufacturing on a large scale and replace humans", Musk emphasized that the reality is still "a large number of people": the total number of Tesla employees is "about 150,000", and "about two - thirds of them work in the factory in some form"; the manpower related to its supply chain is "possibly one million to two million people".

But his judgment is that efficiency will increase significantly: "We don't plan to lay off employees or reduce the number of staff. On the contrary, we will increase the number of employees." The real change lies in the unit output: "The output per person at Tesla will become extremely high."

This echoes his description of the technological progress path: whether it is AI or manufacturing, "it is often an S - curve, or a series of overlapping S - curves: slow at first, exponential growth, reaching a plateau, and then another breakthrough to start the next curve."

Full translation of the interview:

Host Peter H. Diamandis As you can see, I'm still trying to turn "hope" into reality.

Musk You look in great shape.

Diamandis I feel good.

Musk Did you use some kind of anti - aging serum or something?

Diamandis This is our "longevity express train", and we're moving towards that goal. You're on this path too. I think in our last conversation, you've started to accept the idea of extending lifespan.

Musk To some extent, yes. I don't know if we want everyone to live forever, but I think extending the "healthy lifespan" instead of going through a long period of aging and drooling all the time sounds like a good idea. We want to avoid that.

Diamandis First of all, congratulations on the cooperation between SpaceX and xAI. It's a brilliant move that will power humanity's first "Dyson Sphere". I'm curious about the schedule for starting these data centers. How much bandwidth do you think you can get in the first year? Please let us know the speed at which you're achieving this goal.

Musk SpaceX is currently in a quiet period, and I can't disclose information that might cause trouble.

Diamandis Okay, then let's not talk about this. I can understand, but I'm looking forward to the speed.

On Monday this week, we had a conversation here with Eric Schmidt and a person in charge of another hyperscale computing (cloud service) company. I won't mention who it is, but I'm curious about what stage you think we're at in terms of AI's "recursive self - improvement". Have we reached it? Do you think Grok is currently undergoing recursive self - improvement? How is it done? What's the schedule for artificial general intelligence (AGI) and super artificial intelligence (ASI)? Please give us a general idea.

Musk I think we've been in the recursive self - improvement stage for some time. Do you mean fully automated recursive self - improvement without human participation? Is that what you mean?

Diamandis Yes, I'm referring to AI software.

Musk In the process of recursive self - improvement, human participation is indeed gradually decreasing. Each new - generation model is built by the previous - generation model. This has largely happened, but it has not yet achieved full automation. Maybe it will be achieved by the end of this year, at the latest, no later than next year.

Diamandis Do you think there will be a "hard takeoff" (mutant development) of intelligence by then?

Musk We're in the middle of a hard takeoff right now.

Diamandis Okay.

Musk Right now.

Diamandis Yes.

Musk At this stage, when I go to sleep at night, there's a major breakthrough in AI; when I wake up, there's another breakthrough.

Diamandis Indeed.

Musk Frankly, it's hard to keep up. It's really dizzying.

Diamandis I think many of those dizzying breakthroughs are also brought about by you.

Musk Currently, Grok is developing very well. It is the best in some indicators. For example, it is the best at predicting things, which can be said to be the best indicator for measuring intelligence. The new version of Grok is very good.

We're still lagging behind in programming ability. The reason I was late just now is that I just attended an all - staff meeting about programming, sorting out all the work that needs to be done to catch up with and surpass our competitors in programming. I believe we can do it. I think we can achieve this goal around the middle of this year.

In addition, I don't think people fully understand how huge the future intelligence will be, or the extent to which it will surpass human intelligence to the point of being incomprehensible.

You can imagine this situation: Suppose we use a million times more energy than the current total electricity consumption on Earth, that's still only about one - millionth of the solar energy output. Essentially, if you expand the US economy by a million times, the energy it consumes is still only a small fraction of the solar energy. If we really calculate based on the scale of solar energy, even if you expand the current economy and electricity consumption by about a million times, you'll only use about one - millionth of the solar energy.

But what would an economy or an intelligent agent look like, think about, and do if it uses a million times more electricity than the current entire civilization? It would be a very grand picture. The challenge we face is that it's very difficult to even vaguely understand that huge level of intelligence. But it's certain that it can solve all the problems you can think of.

Diamandis Yes, it's like a long - distance journey. You may think it sounds funny, but I really appreciate this unremitting optimism.

Musk I see you're turning "hope" into reality. It's interesting that you've taken the phrase "turn hope into reality" to heart.

Diamandis That's Grok's credit. That was the marketing advice Grok gave me when you were strongly criticizing me.

Musk Right? So you're turning hope into reality. But you were also turning pain into reality before.

Diamandis That's for sure.

Musk When artificial intelligence and robots increase economic output by several orders of magnitude, it's beyond our imagination.

Diamandis We're likely to become a small minority of the intelligence on this planet in a very short time, then a very small minority, and finally an almost negligible minority.

Musk Yes. Not only on Earth, but in the entire solar system. Because if you only develop intelligence on Earth, the best result, that is, the energy that intelligence can use, is about one - billionth of the solar energy. If you're only limited to Earth, this is the best result you can get.

Diamandis That is, the part of the energy we can intercept, right?

Musk Yes. Earth only receives a small part of the solar energy, which is most of the energy that exists in the universe and we can obtain. So in fact, the intelligence in the entire solar system will be many orders of magnitude higher than that on Earth.

Diamandis Elon, can I ask a question? How far into the future can you see? How many years into the future can you make a reasonable prediction now?

Musk It's difficult to accurately predict the future path. Many things often present as an S - curve, or a series of S - curves. It starts with slow development, then exponential growth, then enters a linear region, and finally becomes logarithmic growth (slowing down).

This is roughly what I see in AI breakthroughs. For example, after a breakthrough occurs, it will go through an S - curve and seem to grow infinitely, but then the returns will show a logarithmic decline until the next breakthrough occurs. Therefore, the progress of AI is basically a series of overlapping and connected S - curves.

Diamandis There was a time when you might have been able to predict the situation for the next ten or twenty years. What's your current view?

Musk What I'm going to say next may sound crazy.

Diamandis It doesn't matter. Because you've always been great listeners for our bold predictions.

Musk Yes, I want to say that within 10 years, the economic scale will reach 10 times or more of the current level.

Diamandis Yes. You did say that more than five years from now, the GDP will have three - digit growth, and the economic scale will reach 10 times the current level. But in terms of your prediction ability...

Musk I think a ten - fold growth in about 10 years is actually a rather conservative prediction. Unless something like a third world war disrupts these planned expectations. But without a third world war, if the current trend continues, I'd say the economy will grow tenfold in 10 years.

Diamandis I like that. Can you give us an example?

Musk Humans will establish a base on the moon.

Diamandis Yes. And we'll have humans...

Musk Land on Mars.

Diamandis We'll also build a mass driver on the moon.

Musk I think so. I think we'll have a mass driver on the moon within 10 years.

Diamandis Great. Gerard K. O'Neill's space vision is coming true.

At this year's "Abundance Summit", we had four robots on stage. I'm looking forward to "Optimus". I'm curious about the schedule for Optimus 3, especially when I can buy one or two. When do you expect Optimus to be put on the commercial market? Or will you adopt a leasing model?

Musk We're currently in the final stage of completing Optimus 3. It will be the most advanced robot in the world at present, and no other robot can match it. In fact, I haven't seen any robot demonstration as good as Optimus 3. Frankly, maybe there are such robots, or they're kept secret, but I haven't seen them. Of course, I have to make sure that what I say can be moderately disclosed.

Diamandis We're live - streaming this conversation on the X platform.

Musk Okay, then this is very public.