The Scene
While much of the AI hype has focused on consumer uses, Box CEO Aaron Levie and Zoom CEO Eric Yuan are among the company bosses scrambling to find where the technology will be really helpful: at work.
The two friends have a similarly rosy view of the way artificial intelligence will change the world, which has served their investors well this year. Each of their companies has boosted their market cap by 30% in 2024 as they’ve rolled out new generative AI products.
Both entrepreneurs, who talked to Semafor in a joint interview, have lived through and contributed to periods of massive technological disruption.
Box, founded nearly 20 years ago as a file sharing service, has been transforming itself into an AI company. And at Zoom, video calls and conferences are essentially unstructured data that the “Zoom AI Companion” can parse, generating summaries and suggesting action items and calendar entries.
Still, the founders of the two companies may be on the verge of experiencing the biggest tech revolution yet.
“This time, it’s different from other technology evolutions,” Yuan said. “We can’t imagine what’s going to happen. It’s essentially out of control compared to the internet or the mobile phone. It’s beyond our imagination.”
Levie, on the other hand, views the landscape from a more sober vantage point. “We’ve already had our Netscape moment. We’ve already had our iPhone moment. There was ChatGPT, and now we’re just in the rollout phase of how these technologies start to enter our life more and more,” he said.
Their views reflect this unique period in the tech industry, with Silicon Valley debating how AI will evolve in the coming years.
One view is that AI models are a kind of service that can be incorporated into every software product. Another is that AI capabilities will continue to evolve until they are so powerful that they upend current paradigms, potentially disrupting incumbents in ways they couldn’t imagine.
“For any new technology, if you do not feel scared, the potential is not powerful enough,” Yuan said. “You have to embrace that. There’s no other way around [it].”
In this article:
Step Back
Yuan was one of the key engineers of Webex, which played a significant role in advancing online video streaming and conferencing. Years later, he founded Zoom, improbably creating a videoconferencing service that gave rivals a run for their money.
Levie founded Box in 2005 with funding he got after sending billionaire investor Mark Cuban a cold email. It was a first-of-its-kind file sharing service that changed the way computer documents were stored and sent.
I asked what they thought were the “bottlenecks” in AI, and what they would build if those hurdles didn’t exist.
Levie was quick to say that Box is already developing it. “We’re building the architectures to support a future where pricing has dropped to basically zero, where quality has increased to basically infinite,” he said.
Box’s corporate customers can now apply their choice of AI models directly to their data, customizing “agents” using its Box AI Studio service and querying files using chatbots instead of file systems.
“If you’re in SaaS [Software as a Service], your job mostly is to build a layer that takes whatever workflow or system you have and connects to AI models, and creates as much value between those two things as humanly possible,” Levie said.
His point is that as long as customers store their data with Box, they’ll have access to the latest AI capabilities developed by third parties. Box’s products won’t need to change much as AI offerings get better and cheaper.
Zoom’s strategy is not drastically different. Its customers can link outside services like email and documents, so its AI Companion can help tie together corresponding information from every account. As AI models improve, so will Zoom’s products.
Yuan, though, sees a vision of the future that is cloudier than Levie’s and less certain — and may not be totally possible to plan for.
“I think we’re on more like day one,” he said. “I’m excited now, but not that excited.” Artificial general intelligence, or AGI, is when things will really change. “I cannot imagine what the world will be,” Yuan said. “Aaron and I may not even need to work.”
“The problem with that is then you don’t need documents, so I’m not in any rush to get to AGI,” Levie joked.
Reed’s view
There is no right or wrong way to look at the future of AI. Even those at the very cutting edge of the technology can only speculate on what it will be capable of a year from now, let alone in five years.
During a recent appearance on Lex Fridman’s podcast, Anthropic CEO Dario Amodei said “we have nothing but inductive inference,” suggesting AI models will continue to improve. “There’s some magic to it that we haven’t really explained on a theoretical basis yet.”
In 10 or 20 years, it will be interesting to look back at how AI was viewed by the business leaders of today, when nothing was certain.
Levie looks at AI through a pragmatic lens. He understands the landscape and is embracing it, capitalizing on it for his shareholders.
Yuan is doing the same thing, but openly acknowledges that he has no idea if today’s landscape will even exist once AI reaches a certain threshold of capability — something he told me is probably limited only by the availability of energy to power data centers.
This dichotomy made me think of Blockbuster Video, a place I loved to visit as a kid. When DVDs came out, Blockbuster started to lose market share to Hollywood Video, which embraced DVDs earlier. In 1999, Blockbuster finally rolled out the new format. Neither saw how quickly the world was changing.
It was the middle of the dot-com boom, and you could already watch short online videos. Everybody was getting mobile phones that could magically send digital information through the air. Companies were investing furiously in high-speed internet infrastructure. And still, few people realized we’d soon watch movies online instead of driving to the store and renting physical media.
Will we one day look silly for not seeing the obvious implications of AI? And will business leaders who have the greatest imaginations be able to steer their companies clear of disruption?
Yuan likes to joke that one day, Zoom meetings won’t require human involvement. It will be AI-powered “digital twins” conversing for us.
The idea always gets a laugh, but it doesn’t sound much crazier than someone in 1999 claiming they’d soon be able to instantly watch any movie from anywhere in the world on a supercomputer they carry around in their pocket.
Room for Disagreement
This decade-old essay by quantum computation pioneer David Deutsch argues that AI may not be all that transformative:
“The very term ‘AGI’ is an example of one such rationalization, for the field used to be called ‘AI’ — artificial intelligence. But AI was gradually appropriated to describe all sorts of unrelated computer programs such as game players, search engines and chatbots, until the G for “general” was added to make it possible to refer to the real thing again, but now with the implication that an AGI is just a smarter species of chatbot.”