Asana’s Dustin Moskovitz is bullish on AI but worried about risks

A typical strategy of successful tech founders is:

Establish a company with full ownership. As the business progresses, it sells a significant portion to venture investors. will eventually become a minority shareholder. Make your company public. Sell ​​more inventory over time.

Asana’s Dustin Moskovitz adapted that playbook and completely rewrote the ending.

Moskovitz, who is still best known as the co-founder of Facebook, launched Asana in 2008 to make work more collaborative through software. By the time he took the company public through a direct listing in 2020, his ownership had reached about 36%.

Then he started shopping. After he bought 480,000 Asana shares in June, Moskovitz’s ownership ballooned to his 111.4 million shares, more than 51% of the shares outstanding. Asana revealed in March that Moskowitz had plans to buy up to 30 million more Class A shares this year, and the stock rose about 19% the next day.

“The market has been turbulent for two years and we’ve had some interesting buying opportunities,” Moskovitz said in an interview with CNBC.

Even after gaining 66% this year, Asana shares are still more than 80% below their all-time high since late 2021.

For Moskovitz, who has a net worth of over $12 billion, most of it from his early stake in Facebook (now Meta), being the majority owner of Asana is not a matter of control. Rather, he believes it is the best way to invest in supporting philanthropy.

In 2010, Moskovitz signed the Giving Pledge, a pledge by the world’s richest people to donate most of their fortunes to charity. Moskovitz and his wife, former journalist Kari Tuna, are funding through Good Ventures based on a recommendation from Open Philanthropy.

When it comes to spending that money, nothing worries Moskovitz more than the future of artificial intelligence.

Good Ventures donated $30 million to startup OpenAI over three years in 2017, long before generative AI and ChatGPT were in the popular lexicon. OpenAI, which is now worth around $30 billion, was founded as a nonprofit, and Open Philanthropy said at the time that it “wanted to help play a role in OpenAI’s approach to security and governance issues.”

One of the 10 focus areas Open Philanthropy lists on its website is “potential risks from advanced AI.” The organization gave the National Science Foundation a $5 million grant to support research on how to ensure the safety of artificial intelligence systems, to “establish an academic center focused on AI safety.” recommended a $5.56 million grant to the University of California, Berkeley. Open Philanthropy says it has donated more than $300 million to its areas of focus through more than 170 grants.

“I definitely think there’s a lot of risk there, and I spend a lot of time thinking about it,” Moskovitz said.

Moskovitz co-founded Facebook with Mark Zuckerberg, Chris Hughes and Eduardo Saverin at Harvard University in 2004. He became a billionaire after Facebook’s 2012 initial public offering and owned more shares than any other individual except Zuckerberg.

Even after acquiring additional Asana shares in 2022 and 2023, he still owns about $2.6 billion, less than his $4.6 billion stake in Facebook, according to FactSet.

“I’m just in a unique position to come to the negotiating table with an existing source of wealth,” Moskovitz said. “So, even though it looks like a huge purchase, it’s a relatively normal part of my net worth compared to other founders.”

Moskovitz has agreed not to purchase all of Asana’s outstanding shares or even to acquire ownership of 90% of the common stock. He also intends to keep the majority of the board independent, in accordance with New York Stock Exchange rules, according to his filings.

Moskovitz declined to say whether he is buying up shares to prevent activist investors from stepping in and trying to force change. Activists have been busy in the cloud software space, with Salesforce in particular responding to pressure by expanding its own buyback program and boosting profits.

Recently, the worlds of Moskovitz have collided.

OpenAI leapfrogged from a niche startup to one of the hottest in tech after releasing ChatGPT in November. Before that, Moskovitz was experimenting with the company’s DALL-E technology, which converts text into images. He said OpenAI CEO Sam Altman set up a “lab account” for him last April.

Moskovitz after launching ChatGPT it was fun Ask your chatbot to come up with goals for addressing California’s housing problem.

Meanwhile, Asana joined the parade of companies announcing enhancements to their products with generative AI capabilities that can take human input and display text, images, and audio in response. Earlier this month, Asana announced that it was giving select customers access to several generative AI capabilities powered by his OpenAI models.

“Chat is just one paradigm for how we use these technologies,” Moskovitz told CNBC. “When integrating the system into workflows such as work management, optimizing automated workflows, supporting decision-making, etc., you can literally ask the system questions and get an overview and recommendations.”

Moskovitz said more complex tasks, such as adding structure to a project, are areas where “the potential really shines”. He said technology has the power to take “massive amounts of information and vague goals” and “give something almost in the right direction,” rather than just looking for concrete answers.

Asana could spend more than $5 million on OpenAI’s technology next year, Moskovitz said, adding that he was “very impressed with GPT-3,” the company’s previous large-scale language model. added, “I was even more impressed with the announced GPT-4.” March.

Moskovitz spent six minutes of Asana’s 51-minute earnings call in early June promoting the company’s approach to AI. Microsoft CEO Satya Nadella mentioned AI 32 times during his April earnings call, while he used the acronym 41 times. Microsoft is the lead investor in OpenAI.

Asana “just has a deep personal connection with cutting-edge AI labs,” Moskovitz said.

In fact, the connection is very deep. Altman invested in Asana in 2016. At Asana’s earnings call, Moskovitz told analysts that his company and OpenAI “share Adam D’Angelo, former Facebook technology head and later founder of online Q&A startup Quora, as a board member.” I was very careful.

One of OpenAI’s early board members was Holden Karnofsky, co-CEO of Open Philanthropy. Kanofsky later co-founded AI startup Anthropic with his wife Daniela Amodei. Moskovitz invested in Anthropic in 2021, the same year he co-invested in fusion startup Helion with Altman.

Like Altman, Moskovitz is very bullish on AI and concerned about the damage it can cause.

Moskovitz was one of many entrepreneurs who signed the statement in May, saying, “Reducing the risk of extinction from AI is a global threat alongside other societal-scale risks such as pandemics and nuclear war.” It should be a priority,” he said. The message was from the nonprofit AI Safety Center.

But Moskovitz was not among the signatories to an open letter filed in March by the nonprofit Future of Life Institute. The open letter called on AI labs to stop training their most sophisticated AI models for more than six months. Near the top of the list of signatories was Tesla CEO Elon Musk, an early backer of OpenAI, who called advanced AI “a bigger risk to society than cars, planes and drugs.” We were warned that we should be very concerned.

Moskovitz said Musk’s concerns weren’t completely overstated and that both men wanted to “bring this technology into the world in a safe way.”

“Elon approaches this issue from many angles,” he says. “I think we share some views on potential existential risk issues, but I think we probably share less views on things like AI censorship and arousalism.”

In December, Mr. Musk tweeted “The danger of training an AI to wake up, in other words to lie, is deadly.”

Moskovitz helped create a 12-item list of possible policy changes for U.S. lawmakers to consider.

“My main concern is to ensure that the next generation of cutting-edge products like GPT-5 and GPT-6 undergo safety evaluations before they are released to the market,” he said. rice field. “I think that would require regulation to accommodate all the players.”

In a tweet last month, he even made up words to express his mixed views.

“Excited about AI!” he wrote.

clock: Elon Musk Founds AI Startup Called X.AI To Compete With OpenAI’s ChatGPT