“Powerful II” will appear by 2026
Dario Amodei, CEO of Anthropic, wants you to know: He’s not a fan of apocalyptic predictions about AI. At least this impression is created by him recent essay about 15,000 words long, published on a personal blog last Friday.
In this text, Amodeus paints a picture of a world where all the risks associated with AI are eliminated and the technology brings unprecedented economic growth, social welfare and prosperity. At the very beginning, he criticizes those who exaggerate the capabilities of AI and promote its “superpowers”. However, as one reads, one gets the feeling that his essay goes too far in the direction of techno-utopia, making claims that are not yet supported by the facts.
Contents
Powerful AI already in 2026?
Asmodeus claims that “powerful AI” will appear by 2026. By this, he understands AI, which will be smarter than Nobel Prize winners in the fields of biology and engineering, will be able to solve mathematical theorems and even write “really good novels.” In his opinion, such an AI will be able to control any software and hardware, including industrial equipment, and perform almost any job better than a human.
“This AI will be able to perform any actions – from remote operations to controlling tools through a computer, even designing robots or equipment for itself,” Amodei writes.
However, many obstacles will have to be overcome on the way to such a reality. Current AI models, even the most advanced ones, are not capable of “thinking” the way humans do. They are faster at copying and reproducing patterns from the data they have been trained on. And given the slow progress in robotics, it seems unlikely that AI will soon be able to conduct laboratory experiments and manufacture its own tools.
Ambitious forecasts in medicine and science
However, Amodei remains optimistic. He predicts that within 7-12 years AI will help treat almost all infectious diseases, eliminate most cancers, cure genetic disorders and stop Alzheimer’s disease in its early stages. Moreover, in the next 5 to 10 years, he expects AI-created drugs to help manage mental disorders such as PTSD, depression and schizophrenia.
His predictions are so ambitious that he foresees a doubling of the average life expectancy to 150 years. Amodeus calls it “the 21st century in a nutshell,” suggesting that with AI, humanity will achieve what took 100 years in just a decade.
But these statements raise doubts. Current advances in AI in medicine are far from revolutionary, and many obstacles still stand in the way. AI can help in the testing phase of new drugs, but like human development, they may not work at later stages.
AI that will solve the world’s problems?
Asmodeus argues that AI can solve global problems: eradicate hunger, reverse climate change and transform the economies of developing countries. He believes that AI can raise the GDP per capita in African countries to the level of China in 5-10 years.
These statements are reminiscent of the ideas of the “Singularity” movement, where they believe in similar results of technological development. Amodei acknowledges that such changes will require enormous efforts in global health, philanthropy and political engagement.
However, this requires a fundamental change in human behavior, because in reality, short-term benefits often outweigh long-term goals. Also, many people working on AI training are making well below minimum wage, while companies are making millions of dollars in profits.
Dangers of AI and challenges for society
Asmodeus casually mentions the risks of AI to society, suggesting a coalition of democracies to protect AI supply chains and block unfriendly countries from accessing key technologies. He also argues that AI can undermine repressive regimes and even reduce bias in legal systems — although historically AI has only exacerbated discrimination in these areas.
What will happen to people’s work when AI will perform almost all tasks? Amodeus recognizes that this will cause major economic changes, and humanity will have to rethink how the economy is organized. However, he does not offer specific solutions.
Ultimately, he argues that AI is only an accelerator of technological progress, and that humanity is naturally moving towards democracy and enlightened values. However, it ignores the real costs of AI, such as its impact on the environment and inequality. Economists such as Joseph Stiglitz warn that AI could further concentrate wealth in the hands of companies, leaving workers disenfranchised.
Conclusion
Anthropic is, after all, a business, and its product is the AI that Amodeus says will save the world. But one might question the timing of this essay, given that the company is rumored to be raising billions of dollars in investment.
Source