After rescuing the fallen stock market legend and then becoming a bride, Anthropic's Chief of Staff recounts her personal experience: searching for God at the frontier of AI.

The 25-year-old "AI stock god" did not go completely bankrupt, and it turned out that his fiancee made a huge contribution?

Last Black Friday, the $45 billion fund under Leopold Aschenbrenner's management went to zero overnight.

According to people familiar with the matter, after the liquidation of the blow-up, Leopold only had two remaining assets: the underlying positions purchased with the clients' own funds, and the private equity of Anthropic held under the name of his fiancee Avital Balwit — the latter is valued at about $50 billion.

And that $50 billion was almost gone too.

Sources say that at the critical moment of liquidity crunch, Leopold once planned to sell this part of Anthropic shares at a discount to HSG. It was Avital who stopped him firmly.

"Never sell." These four words saved her fiance.

Now it seems that this may be the only correct decision in the entire blow-up incident, leaving Leopold the "capital" to stage a comeback.

This incident also made Avital Balwit, the "woman who saved the AI stock god", known to the public.

A wedding not delayed by fund blow-up

No matter how severe the Wall Street accident is, it can't stop the major life events.

After the blow-up, Leopold Aschenbrenner and Avital Balwit held their wedding as scheduled in Carmel, a seaside town in California, staging a real-life version of "Wall Street Bloody Wedding".

The ceremony was held in a Tuscan-style villa, and on the last day, guests were seen off at a forest hot spring hotel.

A seminar was also arranged before the wedding. Guests would participate in roundtable discussions and group exchanges. They originally planned to discuss investment and trends, but temporarily changed the topic to life and philosophy due to the crisis. The only request from the two newlyweds was not to send gifts.

After the ceremony, 80 to 100 guests tasted the best wine in Napa Valley. The DJ played a set of music inspired by Aschenbrenner's German background, similar to the style of Berlin's Berghain nightclub.

The whole event was arranged very compactly, and guests could even go home to rest before 11 p.m.

In two days, he lost billions of dollars due to leverage and got a beautiful wife, which was ridiculed by netizens as:

The fund can be liquidated, but the wedding cannot be postponed.

……

After all, in front of the great joy of newly marriage, the rise and fall of hundreds of billions of dollars is just a small twist at work.

Some people even spoofed the wedding vows: "Do you promise never to use leverage again?" And matched it with pictures and texts showing the bride with a solemn expression and the groom with a bright smile.

Although the fund's book loss that month exceeded 30 billion US dollars, Leopold is still a billionaire now. Even after the blow-up, his fund's net return since the beginning of the year is still +80%.

After a 67% drawdown in July, Leopold said to investors:

We are currently managing a fully paid public ledger, holding long stocks and fully paid long options, with no margin/liquidation risk.

The fund is still running, and the next batch of transaction lists will be announced in the middle of this month.

I hope this couple can accept the ups and downs of life and grow old together.

The woman who saved the AI stock god

The two names first appeared together in an employee list.

In February 2022, FTX Future Fund, a charity fund funded by FTX, was officially established, and Leopold was one of the initial team members.

A few months later, the team announced that Avital had just joined. In other words, the two have worked together since at least the middle of 2022.

At that time, Leopold Aschenbrenner had graduated from Columbia University, and this was his second job.

He entered Columbia University at the age of 15, studied economics and mathematical statistics at the same time, and became the valedictorian at the age of 19.

Avital's resume is equally excellent.

She studied political and social thought and cognitive science at the University of Virginia.

In her senior year, as one of the 32 students selected from all over the United States, she won the Rhodes Scholarship of the 2021 class, planning to go to Oxford to study philosophy, focusing on "existential risk" research: in the face of technologies that may change the direction of civilization, what kind of system should human beings establish.

Her teacher described her as having depth, skepticism and humility at the same time.

At Oxford, she was involved in the affairs of FHI (Future of Humanity Institute) and GPI (Global Priorities Institute).

When Leopold was still studying at Columbia University in April 2021, he was already a research affiliate of GPI at Oxford, and he had also done research at FHI at Oxford before. He once published an article saying that advocating growth can ensure the safety of future generations:

Economic growth is not the root of human existential risk, but the best way to go through the "dangerous period".

Both of them focus on existential risk and share the same interests. It seems unexpected that they come together, but it is actually destined.

Originally, the two joined the Future Fund together to study the long-term future and fund AI safety and biosecurity projects.

As a result, a few months later, the biggest sponsor of the foundation exploded, and the two chose to resign collectively with other members of the Future Fund team.

After that, Leopold joined the OpenAI Superalignment team to study how to control systems smarter than humans. And Avital joined Anthropic.

In April 2024, due to differences in safety measures, Leopold was fired by OpenAI. But two months later, he released a 165-page "Situational Awareness", predicting that AGI is approaching, computing power demand is exploding, and the United States must prepare for the superintelligence race.

It is worth noting that in the acknowledgments at the end of the article, he included Avital among those who participated in "forming new discussions".

Subsequently, he rose to fame and founded Situational Awareness LP, which grew from 1.5 billion US dollars to 45 billion US dollars by the end of the year.

At the same time, Avital Balwit was promoted all the way at Anthropic. Around 2024, she appeared in public as the new chief of staff to the CEO of Anthropic.

In 2024, she became famous for a long article she posted called "My Last Five Years of Work". (According to her, she only has 3 years of working time left now.)

In another article, Avital mentioned that she was learning German, her boyfriend's native language.

In recent years, she has not been active in front of the stage, but she is definitely not "the woman behind the successful man".

As the chief of staff to the CEO of Anthropic, she is the only Anthropic executive who reports directly to founder Dario Amodei. Other executives report to Dario Amodei's sister Daniela Amodei.

Dario is responsible for Anthropic's overall strategy, corporate culture and research direction, while Daniela is in charge of daily operations.

Avital does not manage Dario's schedule, but Anthropic in the eyes of this AI giant. Compared with the twists and turns of the stock market and the vows of the wedding, the pressure brought by her career may make her more uneasy.

As the only direct subordinate of Amodei, every judgment she makes is unprecedented, and she can only cross the river by feeling the stones.

In this position, she has to choose the values of powerful AI systems, but the stakes are too high, and she feels deeply uneasy inside.

As one of the few figures shaping cutting-edge AI systems, she once said bluntly: cutting-edge AI R&D is approaching the "edge of divinity".

Facing the heavy moral pressure, she began to go to church to seek God.

The following article was written by her at the end of May this year. After she became widely known last week due to the AI prodigy incident, this article was also dug out by more people.

"They didn't create gods because they missed God"

At parties in the San Francisco Bay Area, there is a popular joke: AI researchers are creating gods.

The pursuit of AGI is like filling a "God-shaped hole". The old faith was overthrown by rationality, and a new idol filled the empty shrine.

Is that really the case? Balwit said no. "That's too arrogant," no one in her circle thinks they are creating gods.

"They didn't create gods because they missed God," Balwit wrote. "They are creating something that unexpectedly brings them to the edge of the position where God should be."

Balwit's position allows her to get a glimpse of the future earlier than most people.

Now, she hopes she can get clearer moral guidance. She is thinking:

If our world has been disenchanted, then why do we still feel the glimmer of "there is something more out there"?

In her opinion, the "God-shaped hole" does not hit the point, because the non-religious status in the Bay Area is not equivalent to the absence of religion.

Standing so close to omniscience and immortality, it is impossible not to lean towards religion.

The paradox of this matter is that the elites in Silicon Valley are talking about something that sounds exactly like religion.

Cure all diseases,

Solve the problem of aging,

Expand science until we know how the universe originated and whether human beings are alone in the universe...

In addition, it may also cause various types of disasters.