Thursday, June 20, 2024

Reid Hoffman is on a mission: to show that AI can improve humanity


SAN FRANCISCO — Reid Hoffman, billionaire entrepreneur and venture capital investor, is worried about artificial intelligence — but not for the doomsday reasons it’s making headlines. Instead, he worries that doomsday headlines are too negative.

So in recent months, Hoffman has engaged in an aggressive thought-leadership effort to enhance the merits of AI. He has done so in blog posts, television interviews and fireside chats. He has spoken to government officials around the world. He hosts three podcasts and a YouTube channel. And in March, he published “Impromptu,” a book he co-wrote with the AI ​​tool GPT-4.

It’s part of a land grab for public opinion around AI in preparation for when the initial explosion of fear and hype over the technology turns into a coherent debate. Sides will be chosen, regulation will be proposed, and technical devices will become politicized. For now, industry leaders like Hoffman are trying to sway the terms of the discussion in their favor, even as public concerns seem to be growing.

“I’m beating the positive drum very loudly, and I’m doing it intentionally,” he said.

Few are as intertwined in so many aspects of a fast-moving industry as Hoffman. The 55-year-old sits on the boards of 11 tech companies, including Microsoft, which has all gone into AI, and eight nonprofits. His venture capital firm, Greylock Partners, has backed at least 37 AI companies. He was one of the first investors in the foremost AI startup OpenAI, and recently left its board. He also helped found AI chatbot startup Inflection AI, which has raised at least $225 million.

And then there is his more abstract goal of “raising humanity”, or helping people improve their circumstances, a concept he deals with in an amiable, matter-of-fact manner. Hoffman believes AI is critical to that mission and points to its potential to transform areas such as health care – “giving everyone a medical assistant” – and education – as examples. Give someone a teacher.”

“It is part of the responsibility that we must think about here,” he said.

Hoffman is one of a small group of interconnected tech executives leading the AI ​​charge, many of whom also led previous internet booms. He is a member of the “PayPal Mafia” of former PayPal executives that includes Elon Musk and Peter Thiel. The latter two backed DeepMind, an AI startup that Google bought, and all three were early supporters of OpenAI. Jessica Livingston, founder of startup incubator Y Combinator, also invested in OpenAI; Sam Altman, CEO of OpenAI, was previously the president of Y Combinator.

Musk has now started his own AI company, X.AI. Thiel’s venture firm, Founders Fund, has backed more than 70 AI companies, including OpenAI, according to PitchBook, which tracks startup investments. Altman has invested in several AI startups on top of running OpenAI, which itself has invested in seven AI startups through its startup fund. And Y Combinator’s latest batch of startups includes 78 focused on AI, nearly double its last group.

Tech leaders differ on the risks and opportunities of AI and are hammering away at the marketplace of ideas.

Musk recently warned about the dangers of AI on Bill Maher’s show and in a meeting with Sen. Chuck Schumer, Dn.Y. Hoffman has explained the technology’s potential to Vice President Kamala Harris, Commerce Secretary Gina Raimondo and Transportation Secretary Pete Buttigieg. Last week, Altman told a congressional hearing that “the benefits of the tools we’ve deployed so far far outweigh the risks.”

In Hoffman’s view, warnings about the existential risk of AI to humanity outweigh what the technology can do. And he believes that the other potential issues caused by AI – job losses, destruction of democracy, disruption of the economy – have a clear solution: more technology.

“Solutions reside in the future, not by relegating the past,” he said.

It’s a tough pitch to a public that has seen the harmful effects of the technology over the past decade, including social media misinformation and autonomous vehicle accidents. And this time, the risk is even greater, said Oded Netzer, a professor at Columbia Business School.

“It’s not just the risk, it’s how fast it’s moving,” Netzer said of tech companies’ handling of AI. “I don’t think we can expect or trust that the industry will regulate itself.”

Hoffman’s pro-AI campaign, he said, is meant to foster trust where it is broken. “That’s not to say there won’t be some damage in some areas,” he said. “The question is, can we learn and iterate in a better position?”

Hoffman has been thinking about that question ever since he studied symbolic systems at Stanford University in the late 1980s. There, he envisioned how AI would facilitate “our Promethean moment,” he said in a YouTube video from March. “We can create these new things and we can travel with them.”

After working at PayPal and co-founding LinkedIn in 2002, Hoffman began investing in startups including Nauto, Neuro and Aurora Innovations, which were focused on applying AI technology to transportation. He also joins the AI ​​ethics committee at DeepMind.

DeepMind co-founder Mustafa Suleiman said Hoffman was different from other venture capitalists because his primary motivation was doing good in the world.

“How can we be of service to humanity? He asked the same question all the time,” Sulaiman said.

When Suleiman began working on his latest startup, Inflection AI, he found Hoffman’s strategic advice so useful that he asked him to help found the company. Greylock invested in the startup last year.

Hoffman was also there in the early days of OpenAI. In 2015 at an Italian restaurant in San Jose, California, he met with Musk and Altman to discuss the start of the company, which has a mission to ensure that the most powerful AI “benefits all of humanity.”

Several years later, when OpenAI was looking at corporate partnerships, Hoffman said he encouraged Altman to meet with Microsoft, which bought LinkedIn in 2016.

Altman said he was initially concerned that Microsoft, a giant with a duty to prioritize its shareholders, might not take seriously OpenAI’s mission and its unusual structure of capping its profits. In any large, complex deal, Altman said, “Everyone is concerned about, ‘How exactly is this going to work?’”

Hoffman helped smooth things over. He talked Altman through various concerns while wearing metaphorical “hats” as an OpenAI board member, a Microsoft board member, and himself.

“You have to be really clear about which hat you’re talking to,” Hoffman said.,

Altman said that Hoffman helped OpenAI “model Microsoft and think about what they care about, what they’re good at, what they’re bad at, and what they’ll be like for us.”

In 2019, OpenAI and Microsoft signed a $1 billion agreement, which has catapulted them to the leading position they are today. (To avoid conflicts of interest, Hoffman was not part of the negotiations and each board abstained from voting to approve the deal.)

A year ago, as Hoffman observed that OpenAI was making progress on its GPT-3 language model, he had another Promethean moment. He immediately flipped an AI switch on almost everything, including Greylock’s new investments and existing startups as well as his podcast, book, and discussions with government officials.

“It was basically like, ‘If it’s not like this, it better be something that is absolutely important to society,’” he said.

OpenAI released a chatbot, ChatGPT, in November that became a sensation. A Greylock investment, Tome soon after integrated OpenAI’s GPT-3 technology into its “Storytelling” software. The number of Tome users grew from a few thousand teams to 6 million, said Keith Peiris, Tome CEO.

Hoffman said that his approach was shaped by his access to “extremely high-quality information flows”. Some is through their business relationships with Microsoft, OpenAI and others. Some is through various philanthropies, such as Stanford University’s AI Center.

And some is through his political connections. He has poured millions of dollars into Democratic campaigns and political action committees. Barack Obama is a friend, he said.

For now, he is using his influence to paint a picture of AI-driven progress. Tech insiders cheer on her cheerleading. The rest of the world is more skeptical. A recent poll conducted by Reuters and Ipsos showed that 61% of Americans believe AI could be a threat to humanity.

Hoffman brushes off those fears as exaggerated. He hopes to work on more concrete problems facing AI, including a tendency to misinform, as tech companies upgrade their systems and deploy them to help.

Looking ahead, he said, there will be more investment, more podcasts, more interactions with government officials, and more work on Inflection AI. The way to navigate the risks of AI, he insisted, is to lead the world toward positivity.

“I’m a techno optimist, not a techno utopian,” he said.