American AI giant Anthropic has "war" with the US President Trump administration, accusing the Department of Defense of listing the company as the only domestic American enterprise in the United States that "national security supply chain risks." But the Defense Department's ban on Anthropic was not easy, and the U.S. airstrike against Iran required the use of the AI model Claude, developed by the company to assess the situation. Anthropic co-founder and CEO, Dario Amodei, insisted on setting a moral red line. He thanked his parents for instilling moral standards and joining the AI world also stemmed from the word "filial piety".

Born in 1983, Amody comes from a working-class family whose father, Ricardo, was a leathersmith, who moved to the United States from Elba, Italy in the 1970s. His Chicago-born Jewish mother, Elena, worked as a library project manager. He also has a sister Daniela Amodei, who is four years younger than him. Amodi described the closeness of family relationships, the great love of parents, the commitment to improving the world, and the teaching of children to distinguish right from wrong. "My parents taught me what is right and wrong, knew what is important in the world, and gave me a strong sense of responsibility."

Amodi believes he must take a responsibility to society. In 2003, when the Iraq war broke out, he wrote an article in the student newspaper, criticizing his classmates for being negative and not saying anything.

Father passed away from a rare illness. Deeply hit

Amodi has loved mathematics since he was a child. He has used a computer as a toy. Amodi grew up in San Francisco, the cradle of the science web bubble during his high school years. He recalls that he had no interest in writing computer programs, immersing himself in mathematics and physics, digging into the truth of basic science and understanding the world quantitatively. After graduating from high school, Amody wanted to become a theoretical physicist. He studied at Caltech, then transferred to Stanford University to study physics, and then to Princeton University.

While studying at Princeton as a graduate student, his father got a rare disease that finally succumbed to in 2006. He was struck by the fact that mathematics and equations could not save people. The most heartbreaking thing for Ammodi was the medical breakthrough within 4 years after his father's death, and his stubborn disease changed from 50% to a cure rate of 95%. Ammodi feels that if technology comes earlier, more people can be saved.

The death of his father had a profound impact on Ammodi's life and he made some decisions to make up for his loss. Amodi's graduate studies have also shifted from theoretical physics to biophysics, working to "accelerate" the development of medical technology to save the beloved.

In 2011, Amodi received his PhD from Princeton and returned to Stanford to do postdoctoral research, where he studied metastatic cancer cells through the proteins of tumors. The research work was complicated, and Ammodi began to perceive the limitations of the human brain. He must use technology to assist him, and he saw the potential of artificial intelligence (AI), believing that this technology could lead researchers to solve problems beyond the human scale. He decided to leave academia and invest in the business world with large funds to promote the development of AI.

Amody joined the Baidu Silicon Valley AI Laboratory (SVAIL) in California in 2014, at the invitation of then Baidu Chief Scientist Ng Enda. In just one year, Amodi was mainly involved in Deep Speech 2 speech recognition project, contributing to the AI's "scaling law". The conclusion suggests that increasing computational power, data volume and model size during training of AI can improve AI's capabilities when it is anticipated.

Join OpenAI and explode differences with top management

Billionaire Elon Musk also saw the potential of AI and hosted a dinner party in California in July 2015, inviting industry pioneers such as Sam Altman, and Amody was also present. The dinner party discussed Google's growing presence in the AI industry, and Musk decided to fund the founding of the AI company OpenAI, which will be led by Altman. Worried about the future, Ammodi switched to Google Brain, an AI lab under Google. "I wasn't convinced at the time, but I was even more skeptical. The dinner was like a gathering of celebrities, tech investors and entrepreneurs." It wasn't until 2016, that Amody was attracted by the talent of OpenAI that he officially joined, leading the development of language AI models GPT-2 and GPT-3.

Amodi worked closely with the Panda team established by OpenAI. However, he is getting further and further away from other senior management in terms of personnel appointment and deployment technology management philosophy. He made no secret of his former comrades. "As the leader of an enterprise, even if it is technically possible to move the company forward, it must be trusted with sincerity, with impure motives, dishonesty and unintended actions to make the world a better place, only to do bad things."

Despite Claude's abuse. Falling out with the Defense Department

During the COVID-19 pandemic in 2020, Ammodi was promoted to Vice President of OpenAI Research. However, he still decided to resign. It is reported that his request for a moral red line was the reason for the break with Altman. Amadi opened his own portal, along with seven former core employees of OpenAI, including his sister Daniela, founded Anthropic in 2021. The company name means people-oriented, with the mission of building large-scale AI models leading the industry, committed to building useful, honest and harmless systems, and in 2023 the start-up of the shocking Claude model.

Among the AI companies that work with the US Department of Defense, Anthropic has the most rules and the most strict self-discipline. In January this year, Ammodi published a long article warning that AI posed a serious "civil challenge". This technology possesses superhuman intelligence, autonomy, and difficulty in control, fearing to bring "there is danger" or be used by dictatorships, forever depriving people of their freedom. The rejection of his request by the Secretary of Defense Pete Hegseth for Anthropic's abandonment was precisely why Anthropic was labelled by the US government as a "radical left-wing awakening enterprise" . In response to being criticized by the outside world as an AI "doomsday theorist," Ahmodi insisted that he only wanted to " accelerate" technological development as long as it was safe, everyone was a winner.