The “doom loop” of drugs and homelessness that has purportedly swept over San Francisco is not affecting one group: Artificial intelligence enthusiasts — largely well-educated, young entrepreneurs — who began flocking to the city 11 months ago searching for funding, information or simply camaraderie in the burgeoning AI space.
And, while larger companies like OpenAI and Anthropic are driving the boom, there are also “a bunch of very, very small startups that are, like, five to 10 people each” across the city, said Yitong Zhang, organizer of Demos & Chill. The event series, held at 23rd and Harrison streets, attracts hundreds of participants a month who can “show off cool stuff they’ve made” in the tech space.
Unlike previous tech booms, where mature startups rented offices and hired hundreds of principal software engineers, the AI boom remains at an early stage. Instead, newcomers (who mostly come alone) become regulars in a complex network of house parties, AI social events and hacker houses. What’s more, many of the newcomers are not just workers; they are looking to start their own companies.
“There’s a lot of hype, but there actually aren’t a lot of jobs,” said Zhang. “The people who moved here are mostly entrepreneurs. … If you came here as an AI founder, you really felt a calling to be here, because your people are here.”
“San Francisco has superseded what New York was in the past.”
Arjun Rai
Investment
The biggest challenge for any startup is getting funding, but “it’s significantly easier doing it” in San Francisco, said Briar Smith, 27, a founding engineer at two AI startups who moved back to the city from Waterloo, Canada, in February.
“I just think if the company is here, it’s more likely to survive. It’s more likely to thrive as well, and reach that unicorn status,” he said. So far, Smith has met with a few dozen investors for the company he works for, Tab AI, a wearable AI assistant and the second AI company in which he has been involved in eight months. It has received media coverage and is set to debut in 2024.
Every day in San Francisco, a potential startup founder can attend one of a half-dozen AI events, where they chat in person with investors, many of whom made money in earlier tech booms. That success makes them willing to make many small bets on AI, founders said.
That’s been Smith’s colleague Tong Yu’s experience. Yu, 27, left San Francisco during the pandemic because the city was “dead,” but returned in March when the launch of GPT-4, the large language model, set the world on fire.
Within two months, she found funding from Clément Delangue, the CEO of AI firm Hugging Face, for her startup, DirtyCat, after a quick exchange at the latter’s AI meetup, which attracted some 4,000 participants. This “small but meaningful” funding allowed Yu to pivot her startup idea from consumer AI to enterprise AI when she realized she was too late to the market with an earlier idea.
“I feel amazing to be at the right time, at the right place,” said Yu.
Arjun Rai, 32, who ditched New York within two weeks after landing in San Francisco in June for a visit, was another of the many to get lucky. Rai now lives in the Presidio, just two minutes away from Peter Thiel, the billionaire whose fund wrote Rai’s first check when he was 19. Life in San Francisco is like an episode of the TV show “Silicon Valley,” he said.
Moreover, a chance introduction to the friend of an UberX Share driver he met one night led to a $100,000 investment in his AI startup, HelloWoofy.com, a marketing AI platform. Though Rai once described himself as the “biggest fan of NYC,” the current lightning-fast world of investment has changed all that. “New York is still recovering,” he said. “San Francisco has superseded what New York was in the past.”
Informal information
Jeff Wang, a founder in his 40s who moved his AI learning system ClassGaga to a WeWork office in San Francisco in June, has been amazed how the informal information circulated in San Francisco tech circles has helped cultivate his company. In some cases, the quality is “even higher than major paid industrial conferences,” he said.
At a recent coffee shop gathering for CEOs, for example, Wang learned that some companies are training large language models at a cost as low as $1 for a million “tokens” of input data. “You just wouldn’t get that from news or publications; probably it would be delayed or even you will never get that informal knowledge,” he said. “But, by going through those events, you get, first-hand, the latest developments in the industry.”
Similarly, when he was thinking about building his company’s own large language model, Wang learned that a company was hosting open-source models. Instead of hiring an engineering team to do the installation, he could get access to the model through them, which saved both time and money.
In June, when his company was having trouble getting access to the interface of GPT-4, someone in a bar told him that the current API was too slow to make it usable. “It just basically told us, ‘forget about it. Don’t use it now. Use something else or don’t use it at all for now.’ And then after a few months, they had some major upgrades and the problem was solved,” he said.
People: Founders and engineers
Now, as in previous waves, a concentration of talent may be San Francisco’s biggest strength. Many founders have come from countries across the world — Canada, Kazakhstan, Singapore — moving their entire teams to San Francisco to work and learn.
Yu met Smith at the AI For Good hackathon organized by Cerebral Valley, and the latter ended up building Yu’s app DirtyCat from scratch. “I’ve met so many people that want to be founders with me or I want to be founders with them,” said Smith, who also got interested in entrepreneurship at hackathons.
The density of founders has blessed Dos Bahá, a 29-year-old serial entrepreneur who moved here in August, with “priceless” founder-to-founder conversations with top-level serial entrepreneurs.
One entrepreneur “gave me a bunch of navigations, a bunch of advice and shared his wisdom as a founder,” said Bahá, who works on GOAT.AI, an interactive storytelling AI platform in two rented hacker houses in the SoMa and around the Mission. “Founders understand each other, because they have to handle a lot of stressful situations.”
Just as Wang moved here to stay closer to talent for AI foundation models (two of the three major large language models are here), Bahá says the city is ideal for hiring local AI engineers. He is trying to move all of his team members to San Francisco, despite the visa and expense hurdles, because in-person collaboration remains the gold standard for him and other founders. In the fast-paced AI industry, he said, “remote work lowers your speed of generating and executing new ideas.”
Such a waste of resources for a technology that can only make things worse. Realtors and slimy politicians are salivating.
No need to be negative about Yujie’s article, which I thought was uplifting and optimistic relative to all the articles here that focus on crime, road accidents and various political posturing and whining. If this heralds a rebound in the city’s economic activity, not to mention its currently moribund downtown, then it is to be welcomed.
Grasping at AI to “save” the city is desperate. This technology is dangerous and cannot be regulated. This wave will create another boom and bust cycle once its uselessness is made obvious and destructive. Banks, governments, large corporations have the funds nd the incentive since AI will result in massive layoffs. Then what? Another doom loop.
So what types of business should the city be encouraging to bring opportunity and prosperity, if not AI and other emerging technologies?
Tell us your big ideas for that.