Digital Einstein: AGI & ASI

There are two interesting terms that those who have delved into the field of artificial intelligence will hear: AGI (artificial general intelligence) and ASI (artificial super intelligence).

AGI, in short, means artificial intelligence that possesses a professorial/expert level of knowledge of all existing knowledge. In other words, it possesses a knowledge as good as that of a human being, possessing a high-level understanding of physics, pharmacy, economics, Japanese, computer programming, architecture, and so on. It is predicted that we will reach this level within the next few years (most estimate it to be within 1-2 years).

ASI, on the other hand, is an artificial intelligence that is not content with existing knowledge but can discover new knowledge, in a sense, an inventor.

IT'S NUCLEAR'S TURN

Let's think about this... Let's have a new Einstein: a digital Einstein. This new digital Einstein should know everything there is to know, not just in fields like physics and mathematics, but also in all fields like law, chemistry, biology, weapons, economics, philosophy, and more, and on top of that, he should be able to invent things.

More interestingly, we could press a button and have 1,000 more copies of Einstein appear, each with a different task. The real Einstein had biological limitations; he needed to eat and sleep. Because the digital Einstein has no such concerns, we could run a million of them simultaneously, as long as we provide electricity.

In short, a different world awaits us, one we're currently struggling to grasp. And because one of the biggest constraints in that world will be energy (one of the greatest needs of digital Einsteins), technology companies like Google, Microsoft, Amazon, and Oracle are building nuclear power plants.

You heard right, we are entering a period where technology companies, not countries, will start working with nuclear energy.

So when will this ASI arrive? Do we need to worry about a technology that will arrive 100 years from now?

Dizzying...

According to estimates, we'll reach ASI-level AI within the next 10 years. Let's be a bit more generous and say somewhere in the 2030s. But it seems like by the 2040s, we'll be waking up to a new discovery every morning.

The first company and country to reach ASI will likely become the most powerful company and country in the world. The first country to achieve this power will have the power to discover anything they desire (or prevent any other country from discovering anything).

That's why OpenAI has begun building a data center (StarGate) in which they will invest approximately $500 billion. Meta (the company that owns Facebook and Instagram) plans to invest $600 billion in data centers by 2028. We've mentioned technology companies building nuclear power plants.

The US has imposed extensive restrictions on the sale of advanced chips to China (to keep China behind in the AI ​​race). China, in turn, has made an incredible push to produce intelligent models like DeepSeek. In other words, we are living through a period of fierce competition and advancing at full speed toward an exciting future.