December 17, Lars Tvede, author of best-selling books such as "Financial Psychology" and "The Inescapable Economic Cycle", registered financial derivatives trader of the National Futures Association (Chicago), entrepreneur and investor ) was a guest at the "Alpha Investment Summit" hosted by Wall Street Insights, where he talked about the catalytic effect of innovation on the economic cycle, and boldly looked forward to the future disruptive technological progress driven by AI.
Wonderful views
- The historical trend of GDP per capita shows that although economic growth faces cyclical fluctuations, it is mainly driven by innovation.
- cloud computing has created crazy efficiency improvements, and the time for doubling computing power has been shortened from 10 months to five weeks.
- The era of super-intelligent economy is coming, when computer-generated intelligence will be close to human intelligence, and may even surpass human intelligence in 2040.
- AI technology development will significantly change multiple fields. For example, AI will write 8% of software codes by 2025, and commercial nuclear reactors will become a reality by 2030.
- In the next few years, automated research laboratories will take off, running 365 days a year, operated by robots and analyzed by AI. The way of scientific research will be completely subverted, and it will also accelerate human scientific exploration. process.
Text
Hello everyone, I am honored to come to the Wall Street Insights alpha forum.
What I want to talk about today is technology. What will technology bring? How will it change our lives? How will it change our economy?
I'm going to introduce you to some innovations that I think are pretty amazing and will likely lead to breakthroughs in the coming years.
Of course, the basis of what I share is a long-standing interest and I have written many books in this area. Also the founder of a company called Super Trends, where we track innovation around the world.
Our 135 experts have created a timeline of Earth's innovation history, which stretches back 3 billion years and thousands of years later, and looks like this on the screen.
We use AI to find truly important innovations, both in the past and in the future. We have compiled 4,000 predictions for the future and 16,000 past innovations. Then use AI to analyze it.
We have discovered connections between approximately 200,000 different innovations.
This is typical, innovation is a rearrangement of old things. As a whole, different great innovations are connected to each other.
We look for patterns everywhere, trying to understand what it all means.
One of the models we use in our thinking is called the Kaufman model. This model divides innovation into three different categories.
The first is a shadowy future where anything is possible.
The second is the adjacent "possibility zone", where things have just been invented and can be recombined in innovative ways.
Think about electric cars represented by Tesla. Musk discovered that we have batteries, cars, and smartphones. Can we combine these things?
Many people can't see these things, but innovators like Apple and Tesla can. The third area of
is the "real area", which is the node we are currently at.
Innovation is the core force that promotes economic growth
The process of innovation has been ongoing for 3 billion years, and the innovation process began to accelerate around 1450. Since then it has grown exponentially. Innovation is an engine that creates growth.
So when you talk to economists about the path forward, they talk about how much labor do you have, how much capital do you have? Sure, they might talk about business cycles, but that's not really labor in the long run. Nor is it real capital.
What really drives growth is innovation. To illustrate this point, I created this chart, which represents GDP per capita, in this case GDP per capita in the United States from 1870 to the present. There is a straight line in the middle, representing continued growth. But everyone knows that this is impossible. This is not the case because growth cannot be a continuous straight line.
The actual situation is like this.We can see that in the 1930s, the per capita GDP of the United States fell sharply, which was the famous Great Depression period. Later, the per capita GDP increased rapidly.
Later, per capita GDP also experienced a large decline during the epidemic, but then resumed growth.
So, in the long run, economic growth always fluctuates around this line, which is driven by innovation. Innovation is always there, which is why the world is getting richer.
Global real GDP per capita grows by about 20% every decade. The consequence of this is that by the end of this century, the world may be experiencing massive disruption.
Therefore, we should plan for a richer world. Throughout my life, there have been a lot of people saying that this is all going to end because we are running out of resources because of this, that and various factors.
But the thing is, everything continues. So, it seems to me that, for the most part, the world is moving in a good direction.
There is also population growth. Current population growth rates mean that global real GDP increases by 30% every decade
But if you look at all 200,000 innovations, the world economy can be divided into four different periods. The first is Malthusian economics. In this economy, the result of innovation is a small increase in population, more people. But GDP per capita has not grown, which is called a small island economy.
Then the industrial revolution happened and the economy kept growing. At the same time, the population grew.
The super smart economy will begin to explode in 2022
Around 1980, the Internet entered the focus of innovation, and the measurement of innovation value began to become more precise. It's digital, it's computers, it's telecommunications, it's biochemistry and so on. With this kind of precision, we can create more with less.
The growth in per capita consumption of goods in rich countries has become a straight line. Agricultural land is even starting to decline, which is a good development from a sustainability perspective.
But I think in, say, 30 or 40 years, when we look back at this chart, we may find that the fourth economic era begins in 2022. This is the superintelligence economy, and this is the moment when computer-generated intelligence really explodes and starts to become very similar to human intelligence.
Let’s delve deeper. One of the main drivers is the sudden acceleration in the price effectiveness of computing power. In terms of compute per second per dollar, as this graph shows, we can see on the left side of the graph, it shows exponential growth. Each step up is 1000 times, from 1900 to now.
First in the field of computing, there will be an efficiency improvement every 16 months, but with the development of electronic technology, it has accelerated and the time has been shortened to 10 months.
Then we got cloud computing, and we started to see absolutely crazy productivity growth and efficiency growth, and the doubling time was shortened to five weeks.
Cloud computing is driven by a combination of many different factors, including better chips, telecommunications and software. This is the main driving force for the development of technologies such as AI. Just in the past few years, we have seen AI defeat humans in one game after another.
For example, in this picture, the orange line represents the ability of ordinary people, and the lines in other colors represent AI.
We can see that a few years ago, AI started to become better at writing than humans, and then speech recognition, image recognition and text understanding, and contextual understanding, it can actually read the entire paper and really understand all of it. content. In these respects, it took only a few years for it to defeat humans.
What surprises me the most is that, for at least the past 20 years that I have been tracking AI, I thought AI was some kind of magical supercomputer, but I didn't expect it to be so creative. Others thought so too, but no one predicted that generative AI would be so creative.
Now we say that AI can hallucinate, which is bad, but if you can use AI correctly, its efficiency is amazing.
There is a very, very thin blue line at the top of this image. You may not be able to see it because it is too thin.This picture shows the number of protein folds in different 3D forms in the experiment. This is a very, very complex experiment. If you want to simulate it on a computer, you need the world's largest supercomputer to complete it. If you do it in the laboratory, It is also very time consuming and very expensive.
But in recent years, AI has become very good at this. What you see now is orange. The number of proteins we have simulated using computers is 772 million, and it is still increasing. This is an example of a super-smart economy.
On the website HuggingFace, there are many specially trained AI models. I've been keeping track of the number of models above.
I found that the number of uploaded models is becoming more and more as the economy grows. When I took this screenshot, there were 348,000. There are now thousands of new models uploaded every day.
This is the model store. You can think of it as a brain capable of completing different tasks. There are a lot of things you can do with the
big model. You can ask it almost any question. It can all be tried to answer. Works like the human brain. With these AI stores, there are brains corresponding to different tasks.
This is also very similar to the human brain, because the human brain has more than 130 different types of brain cells, which have different brain functions and are designed to solve special tasks. The human brain is like a large model loaded with various apps.
In our company, we use many special models for markup, instructions, etc. So this also connects to the Internet, and then the Internet connects to the Internet of Things. So a lot of different devices, big data are being used to analyze this.
Then you have the connection to robotics. So this is a huge virtual and physical machine that humans are building. And then within the next few years, quantum computing will kick in and solve some tasks, not all, I don't know, but some tasks it will be millions of times faster than what we can solve today.
So let’s look at the impact this will have on the overall economy, as predicted in 2017 by the consulting firm Accenture. They took a standard macroeconomic model and then looked at trend growth forecasts out to 2035 for different countries and got these numbers that I'm showing here, which are typically between 1.5% and 3% trend growth.
But then they did industry-specific research on the impact of AI, and they revised that and got a higher trend growth rate. As I said, this was done in 20 years, but now we know that in the next 12 months or so, AI is actually exploding, accelerating a lot of predictions.
For example, Goldman Sachs recently suggested that unit labor productivity growth could more than double in the United States due to AI. What this means is that if we go back to the growth trend model I showed earlier, in this case the AI will make the trend growth curve slope upward for a period of time.
We don't know how long it will last. Some say five to seven years, some say ten to twenty years. But we're looking at, I'm not talking about business cycle effects and so on. I'm not talking about the impact of the housing market or interest rates or labor. I'm just talking about the impact of innovation, which seems to have started to accelerate in the last 12 months.
AI will disrupt everything we know
Starting first in computing, AI will write 8% of software code in 2025, and no-code will soon become common; in the same year, 10% of portable devices will have personal AI.
In the next two years, AI will understand you and become your coach, your psychologist, your virtual trainer, your educational tutorial tutor, your personal travel organizer, etc. It can be built into any portable device, such as your laptop, watch, and smartphone.
At that time, automated research laboratories will take off, running 365 days a year, operated by robots and analyzed by AI. The way of scientific research will be completely subverted, and it will also accelerate the process of human scientific exploration.
Then in 2028, AI will be combined with synthetic biology. AI can really understand what is happening in the cell at the atomic level. You can create cells with various functions like no-code programming, and we can do whatever we want. Create life.
In 2030, AI will be able to classify genes. Today, we still don't understand the role of many genes.
In 2032, heterogeneous data processing will occur in more than 50% of computer operations. I mentioned before that the brain has more than 130 different functions. If more than half of the computers can operate different types of chips, our computing efficiency will achieve a leap. .
In 2030, our AI training costs will drop by 99.99% compared to 2020. But we have different predictions, saying that computing power that cost 4.5 million US dollars in 2020 may only cost 30 or 300 US dollars in 2030.
By 2040, the comprehensive computing power of AI will surpass that of humans. Many people speculate that by this period, superhuman AI will have self-awareness.
Let's look at the changes in media and lifestyles. One of the things I think will happen is digital encoding, where the surfaces of various objects can become screens; also, for example, many publishers want me to record audiobook audio. , but it is no longer needed because I can make a voice sample and let AI read it. By 2024, this technology will become very mainstream.
In 2026, AI will be used to track consumer emotions and use language models to analyze emotions. For example, on social media such as X, AI can tell us how people's emotions change due to a certain event. This is very valuable for marketers, politicians.
In 2026, AI will be able to turn a movie into a book, and most news reports will be generated by personal AI assistants, so news will become completely customized content.
In 2027, more than 25% of non-fiction books may be written by AI. The price of books will be greatly reduced and the number will surge. The art industry will also be completely changed, and cooperation with AI will become commonplace. There are also some more amazing technologies that may become feasible because of AI, such as nuclear fusion.
Commercial nuclear reactors may be available by 2030. With nuclear fusion, we can power the world for more than a billion years, and by 2039 perhaps most of a country's electricity will come from nuclear fusion.
Genes and life technologies will also thrive in a super-intelligent economy.
In 2025, U.S. regulatory authorities will approve more than 10 gene therapies every year, and human doctors can treat diseases through gene editing.
In 2026, we will have continuous bioreactors for culturing meat, which do not require any live animals, just growing animal cells in petri dishes. This technology is quite expensive, but with the development of AI, it is expected that its Prices will drop significantly, and less land will be needed for human production.
In 2027, previously extinct species are expected to be resurrected, such as the mammoth. We can use mammoth cells for reverse genetic engineering.
In 2028, humans may invent biotechnology that can eliminate cancer. In the 2030s, stem cells can be grown inside the human body using tiny medical devices. These tiny devices can be delivered to areas where cell regeneration is needed, such as cartilage.
Then by 2035, precision fermented protein will be 10 times cheaper than agricultural production, and we will be able to use fermentation technology to produce very cheap food.
The last area is healthcare, where the way people stay healthy will change a lot, with personal AI tracking activities such as our sleep. The previous invasive sampling will be replaced by smart watches and rings. DNA testing will tell us which factors our body is susceptible to. We can create a customized guide about healthy life, so that we can have a longer healthy life and a happier and more fulfilled life. life.
Therefore, handheld skin scanners will appear in 2024. You can check your skin with your mobile phone and tell you whether you should treat it. The mRNA malaria vaccine will also be born to completely eliminate malaria. In 2027, nerve damage could be repaired with biodegradable electronic devices. Even if you fall off a horse and break your cervical spine, you can still move again. In 2028, brain scans will be able to diagnose mental illness. In 2029, a reversal vaccine will be commercially available for the treatment of type 1 diabetes. In 2030, patients will be operated on by microrobots.
In 2031, whole genome sequencing will become synonymous with newborns. When you have a child, the supplier will ask you, do you want DNA secrets? They will tell you what risks your newborn faces and can be corrected and avoided in advance.
In 2052, because more and more people will be doing liquid biological sampling, we will have these huge databases.
What I can see is that if anyone tells you that the best stuff has already been invented, they are wrong. Our innovation is accelerating tremendously and will create effective, environmentally friendly technologies, and I believe we have a great future.
Thank you!
