Comprehensive coverage

A breakthrough in nuclear fusion - thanks to artificial intelligence

Sam Altman personally invested 375 million dollars in a company called Helion that is trying to develop a nuclear fusion reactor. So far this mission has not been successful, but now artificial intelligence may improve planning, which will lead to a revolution in the field of energy

Simulation of a nuclear fusion reactor. Courtesy of General Atomics.
Simulation of a nuclear fusion reactor. Courtesy of General Atomics.

Sam Altman is best known for being the CEO and one of the founders of the OpenAI company, which developed Chat-GPT and changed the way we work with artificial intelligence. Not surprisingly, Altman enjoys considerable wealth. What is surprising is the way in which Altman chooses to redistribute his wealth, in investments that are mainly divided into two areas: life extension and nuclear fusion. It's not about divorces: Altman invested 180 million dollars in Retro BioSciences for life extension, and another 375 million dollars in Helion Energy for nuclear fusion.

No one disputes Altman's understanding of AI, but his financial choices beg the question: What does he know that we don't? 

I believe the answer is that precisely from Altman's knowledge in the field of artificial intelligence, and the fact that the man is connected by vein to the most advanced technological trends today, he has the ability to understand the direction in which we are moving forward. In short, Altman chooses to invest in areas that would have seemed high risk just a few years ago - but also have the greatest profit potential. And he does this because he understands that we are entering a new period of accelerated technological development. One that will come, of course, thanks to the artificial intelligence behind which Altman stands.

And a collection of developments from the past year supports this statement.


nuclear fusion

When we talk about nuclear energy today, we are actually referring to only one way of producing energy: using nuclear fission. In the process of nuclear fission, a heavy nucleus of an atom is split into two lighter nuclei. The fission process also releases an abundance of energy that can be harnessed for our needs. The fissioned 'fuel' material is atoms of radioactive materials such as uranium and plutonium, and the byproducts of fission are radioactive and dangerous to the environment. 

Nuclear reactors currently rely only on nuclear fission. Although they are relatively efficient and clean (certainly more so than coal-fired power plants), the energy production process can get out of control in extreme situations, and lead to disasters on a large scale. Such disasters, for example, happened in Fukushima and Chernobyl before it. This is why there are so many nuclear reactors today stands at 437 only in the whole world. The number may sound large, but it dwarfs the more than 2,400 The coal-based power plants

Physicists have fantasized for many years about another type of nuclear energy: one that could come from a reverse process, known as "nuclear fusion." If in the first process we split atoms, then in the fusion process we cause small atoms to fuse together and become larger atoms, while they release energy. The fuel in this case is not radioactive but hydrogen atoms, and the main byproduct is helium - a gas that is not harmful and is not considered a pollutant. 

This is all well and good, but nuclear fusion doesn't just happen. It occurs in the Sun and many other stars under conditions of enormous pressure and high temperatures. It is very difficult to make such a process materialize on Earth. It requires a very large investment of energy to hit with a laser the area where the fusion will take place. Only in 2022 did he succeed for the first time United States Department of Energy to reach fusion "on equal terms". That is, that a certain investment of energy resulted in the production of exactly the same amount of energy. It doesn't sound particularly impressive, but members of the United States Congress and Senate have described reaching energy parity in nuclear fusion as 

"... a historic, innovative achievement..." and "this is a very big deal".

why? Because they understand how science and technology work, and see the trend of improvement in the production of energy from nuclear fusion. They understand and expect that in the coming years we will be able to optimize the process even more - until we reach a situation where we can truly produce green energy, with a minimum of danger to the environment. 

Members of the United States House of Representatives believe this, and they are not alone. As mentioned, Sam Altman also joined them in recent years. Probably because he knows that the rate of improvement will be higher than expected, thanks to the technology of which he himself is one of the main promoters: artificial intelligence.


Artificial intelligence and nuclear fusion

In February 2024, a new study was published in Nature - probably the most prestigious scientific magazine - which illustrates the close connection between artificial intelligence and science and technology in general, and nuclear fusion in particular. 

The researchers examined a model of innovative reactors for nuclear fusion, which would be shaped like a doughnut. Powerful magnets will cause the particles to stick together inside the donut and continue running along it in an endless circle. In this way, a fusion reaction should be created over time, which can continue to provide energy without interruption. The trouble is that if there is even one small disturbance in the magnetic field, the flow of particles will be disrupted and the process will come to a premature end at best. In the worst case, the device can be damaged and even destroyed.

This model is not only theoretical. One of the most promising reactors today in the field of nuclear fusion is called ITER – International Thermonuclear Experimental Reactor. Only a few such disturbances are enough to convert Irreparable damage to its delicate ingredients.

The scientists in the study from early 2024 used artificial intelligence to detect in advance the subtle signals indicating that the reaction is about to go out of control. The AI ​​detects the signs only about 300 milliseconds before actual damage to the reactor occurs, but that's enough time for it to rebalance the magnetic fields and allow the reactor to continue to work as usual. When the artificial intelligence was tested on a tiny, primitive reactor, it was able to preserve its operation.

This is just one example of the way in which artificial intelligence is currently advancing the field of nuclear fusion. She does not come alone. It joins another study from 2022 in it Artificial intelligence is used to shape and control the magnetic field in the crucible. And these two join a larger movement, within it Invested by the United States Department of Energy Almost 30 million dollars by mid-2023, in projects aimed at using artificial intelligence to jump-start the field of nuclear fusion. The International Atomic Energy Agency also launched the "Artificial intelligence to accelerate research and development in fusion", as her name is.


Artificial intelligence - for the benefit of science

The fact that artificial intelligence can assist in research and development in the field of nuclear fusion, and in monitoring and controlling the operation of the reactor, does not surprise anyone today. In a recent Nature survey of scientists, about 85 percent believed that artificial intelligence could help with repetitive, or labor-intensive, research tasks. 38 percent believed that artificial intelligence could improve the productivity of researchers.

Artificial intelligences don't just help with routine scientific tasks. We can find them today planning experiments on their own. the physicist Mario Kern, for example, used in recent years the algorithm that he proposed for experiments in quantum mechanics, and was shocked by the results - 

"I let the algorithm run, and within a few hours it found exactly the solution that we as human scientists couldn't find for weeks."

Other artificial intelligences manage entire laboratories - known as Self-Driving Labs, or "autonomous laboratories" - on their own. There they can propose hypotheses, develop experiments, control the robots that perform the experiments, decipher their results, and soon also write drafts for articles that will share the results and conclusions with the wider scientific community. 

In view of all this, it is no wonder that one of the largest research bodies in the world - the "National Academies of Sciences, Engineering and Medicine" in the United States - recently held a two-day workshop dedicated entirely to "Artificial intelligence for scientific discoveries". Among the descriptions of the meetings in the workshop one can find a clear recognition of the contribution of artificial intelligence to science - 

"Artificial intelligence has already made a significant impact on the field of research in areas that include materials research, chemistry, climate science, biology and cosmology. These developments combined the development of autonomous laboratories for the efficient performance of physical experiments, acceleration in the power of artificial intelligence for analyzing large data sets and the connection of many different information channels."


When two rivers connect

The situation today can be compared to the one where two rivers join together. One river is human intelligence in the scientific community. To date, human intelligence has driven innovation and generated the most research. Now it is joined by a second river, which is now reduced in size, and is that of artificial intelligence. Although the second river is still narrow, and its waters are still shallow, all those involved in the field of artificial intelligence understand that it is expanding and growing rapidly. The faster it grows, the faster the ocean into which the two rivers flow their waters will fill.

The ocean, of course, is the product of scientific research and development. The ocean contains our understanding of the world, and the technologies that can arise from that understanding. 

Artificial intelligence has leapt forward in the last five years by leaps and bounds. Progress does not stop for a moment. If anything, it just keeps accelerating. Investors in the field of artificial intelligence understand this. They are attentive to what is happening. They understand that the river of artificial intelligence will continue to feed the ocean of science and technology at an increasing rate, and that developments in the field of artificial intelligence will also lead to an acceleration in the discovery of new scientific insights and the development of new technologies.

Sam Altman is one of these investors and managers, and he chooses to bet big on the most important scientific and technological fields. Specifically, it concentrates on nuclear fusion and life extension. I refrain from saying whether he invests In companies The correctness in these areas, but I believe that his investment strategy is correct: artificial intelligence can cause miracles and wonders to be realized in our lifetime, and if so - it is worth investing in the areas that have the greatest impact on our lives, and with the greatest profit potential.


There are also storms in the ocean

Perhaps you were able to understand between the lines that I am optimistic about the future of science and technology. This is true. Or rather, I am optimistic that artificial intelligence will be able to elevate our scientific understanding and help us develop new technologies faster and better than ever before. I sincerely believe that artificial intelligence will allow us to deal with problems whose solutions will change the world: free energy, prolonging life, curing all diseases, copying human consciousness to a computer and much more. 

All this does not mean, of course, that there will not be dangers as well.

Soon, every child will have the superpower of artificial intelligence that can conduct research for them and explain to them how to assemble deadly viruses and bacteria. Every PhD student will learn how to build atomic bombs. Any mad genius could place an entire autonomous laboratory in the basement. And of course, advanced artificial intelligence engines will also be able to get out of control, develop technologies against humans - and implement them on their own.

These are not theoretical dangers. They are very real, and a direct path can be drawn that explains how we will reach them. I don't want to downplay their importance. We will have to deal with them as well, but we cannot forget that alongside these great concerns, we will all enjoy the fruits of artificial intelligence in all fields. There are always fears, concerns and dangers - but the successes will serve us all, sooner rather than later.

free energy. Extending healthy life span. These are just two key areas that Sam Altman believes are going to leap forward in the coming years. They will join many others, and the world will never be the same again.

More of the topic in Hayadan:

Science website logo
Search