Garry Tan, president and CEO of Y Combinator, told a gathering at the Economic Club of Washington this week that “regulation probably needs to be done” when it involves artificial intelligence.
Tan spoke with General Catalyst board member Teresa Carlson for a one-on-one conversation where he discussed all the pieces from learn how to get into Y Combinator to artificial intelligence, noting that “there’s no better time to work in tech than now.”
Tan said he was “generally supportive” of the National Institute of Standards and Technology’s (NIST) try and construct a GenAI risk mitigation framework, and said that “much of the EO by the Biden administration They are probably on the right track.”
The NIST framework proposes various things equivalent to defining that GenAI should comply with existing laws governing issues equivalent to data privacy and copyright; disclosing the use of GenAI to finish users; enacting laws prohibiting GenAI from producing child sexual abuse material, and so on. Biden’s executive order covers a big selection of orders, from requiring AI firms to share security data with the government to making sure small developers have fair access.
But Tan, like many Valley VC firms, was concerned about other regulatory actions. He called the AI-related bills pending in the California and San Francisco legislatures “very concerning.”
Like the California bill that is causing confusion is the one introduced by state Sen. Scott Wiener, which might allow the attorney general to sue artificial intelligence firms if their products are harmful. “Polityka” reports.
“The big policy discussion right now is what does a good version of this really look like?” Tan said. “We can look to people like Ian Hogarthin the UK to be prudent. They also keep in mind the idea of concentration of power. At the same time, they’re trying to figure out how we support innovation while also minimizing the worst possible harm.”
Hogarth is a former YC entrepreneur and artificial intelligence expert who was appointed to the UK’s Artificial Intelligence Model Taskforce.
“What scares me is that if we try to address a science fiction problem that is not present,” Tan said.
As for how YC manages liability, Tan said that if the organization doesn’t agree with the startup’s mission or what the product could do for society, “YC just doesn’t fund it.” He noted that he had read several times in the media about a company applying to YC.
“We go back and look at the interview notes and say we do not think it’s good for society. And fortunately, we didn’t finance it,” he said.
AI leaders keep messing up
Tan’s guidance still leaves room for Y Combinator to create multiple AI startups as cohort graduates. As my colleague Kyle Wiggers reported, there are 86 AI startups in the Winter 2024 cohort, almost double the number in the Winter 2023 cohort and almost triple the number in Winter 2021, in keeping with YC’s official startup directory.
And recent news events have people wondering whether or not they can trust AI product vendors to be the ones to define responsible AI. Last week, TechCrunch reported that OpenAI was eliminating its artificial intelligence team.
Then the debacle concerned the company’s use of a voice that appeared like that of actress Scarlet Johansson during the presentation of the latest GPT-4o model. It turned out that she was asked to make use of her voice and she refused. OpenAI has since removed Sky’s voice, although it denied it was based on Johansson. This, in addition to issues related to OpenAI’s ability to recoup worker equity, were among several aspects that contributed to this it made people ask questions openly Sam Altman’s scruples.
Meanwhile, Meta announced the AI news itself, announcing the creation of an AI advisory board that shall be comprised entirely of white men, effectively leaving out women and people of color, many of whom have played key roles in the creation and innovation of this industry.
Tan didn’t comment on any of those cases. Like most Silicon Valley VCs, he sees opportunities for huge and lucrative latest businesses.
“We like to think of startups as a maze of ideas,” Tan said. “When new technology emerges, such as large language models, the whole maze of ideas is shaken up. ChatGPT itself was probably one of the fastest-successful consumer products to hit the market in recent times. And that’s good news for founders.”
Artificial intelligence of the future
Tan also said San Francisco is at the center of the AI movement. For example, Anthropic, founded by YC graduates, and OpenAI, which was a YC spinout, were created there.
Tan also joked that he has no intention of following in Altman’s footsteps, noting that Altman “had my job a few years ago, so he has no plans to start an artificial intelligence lab.”
One of YC’s other success stories is legal startup Casetext, which was sold to Thomson Reuters in 2023 for $600 million. Tan believed that Casetext was one of the first firms in the world to achieve access to generative AI, and then was one of the first firms to exit it. in generative artificial intelligence.
Looking at the way forward for artificial intelligence, Tan said that “we obviously have to be smart about this technology” because it poses threats from bioterrorism and cyber attacks. At the same time, he said there ought to be a “much more balanced approach.”
He also posits that there likely won’t be a winner-takes-all model, but fairly “an amazing garden of consumer choice freedom and founders who can create something that touches a billion people.”
At least that is what he desires to occur. It could be in his and YC’s best interest – many successful startups returned a number of money to investors. So what scares Tan most is not bad AI, but the scarcity of AI to decide on from.
“In fact, we could find yourself in a different, truly monopolistic situation where there is a lot of concentration in just a few models. Then you talk about rent extraction and you have a world I don’t desire to live in.