- Jasper, an AI startup, hosted what it claims was the first conference dedicated to generative AI.
- Taking place in San Francisco, the event hosted a packed crowd of 1,200 attendees.
"It's a step forward in humanity," one VC said of generative AI. An executive compared it to the Gutenberg printing press. One CEO called the technology "not just hype, but something real."
That's how three attendees — Sameer Dholakia, a partner at Bessemer Venture Partners; Greg Larson, the vice president of engineering at Jasper; and Dario Amodei, the CEO of Anthropic, respectively — described the hot new subset of artificial intelligence at a conference called Gen AI in San Francisco on February 14.
Gen AI served as the self-proclaimed "first" conference dedicated to generative AI, which ingests vast sums of human-generated information and, from that, learns how to mimic human creation. It could be the first of many to come.
Roughly 1,200 people gathered at Pier 27 for Gen AI, which Jasper, a startup that has raised over $140 million, hosted. It brought together leading companies in the industry, including the startups Anthropic and Stability AI.
The conference comes at an interesting moment for generative AI. This past November, OpenAI unveiled ChatGPT, the chatbot that captured the internet's imagination with its ability to respond to just about any question using full paragraphs almost indistinguishable from something a human would write. While the hype around generative AI is reminiscent of the crypto and Web3 hype that dominated the pandemic, insiders believe the technology is here to stay.
Microsoft recently invested billions into OpenAI with plans to integrate a chatbot into Bing. Google followed up shortly after by announcing its own chatbot for search called Bard.
The frenetic pace of development in the field is perhaps why Jasper held the event.
Speaking in a fireside chat, Nat Friedman, the former CEO of Microsoft's GitHub, praised Microsoft for quickly introducing the chatbot to Bing despite concerns about its readiness. The incumbent technology companies have gone from a decade of "corporate blandness to wartime," Friedman said, as Microsoft and Google battle it out to win the next generation of AI-based search.
Insiders say generative AI is not just a fad
Generative AI has already run into some speed bumps. Early beta testers discovered that Bing's chatbot can sound highly emotional. The news outlet CNET came under fire for writing news stories using ChatGPT, which audience members discovered to be riddled with factual inaccuracies. Some worry that large-language models powering the likes of ChatGPT could spread bias and misinformation or become vectors for disseminating spam and malware. Some also fret over the possibility it could take over human jobs.
It's hard to ignore how Silicon Valley has a way of jumping onto new fads only to quickly scrub them from memory and move on when all the excitement meets speed bumps.
Still, at the conference, there was little talk of technologies that VCs plowed billions of capital into, like Web3, before the industry entered the current "crypto winter." Instead, AI leaders explained how generative AI is different and is something that has practical applications today.
Amodei said that consumers, businesses, and developers alike are moving at "record speeds" to adopt generative AI.
"Every C-suite executive is asking, 'What is our AI strategy?' Every company will have AI, every piece of a company will have AI," Amodei said at a panel.
Emad Mostaque, the CEO of Stability AI, cited a tweet during a panel highlighting how a generative AI GitHub project managed to surpass bitcoin in "stars" — the symbol developers use to bookmark projects — on the code repository site within just 40 days. This is despite the fact that the bitcoin project has been on GitHub for much longer.
"People see the implications of generative AI immediately," Mostaque said at the panel.
Amjad Masad, the CEO of Replit, emphasized that developers and the public should not fear AI, but rather embrace it.
"Generative AI is going to augment humans, advance their abilities, and push humanity forward," Masad said.
What's different with generative AI is that large-language models have been quietly in development for some time, executives said. Google has worked on chatbots and related technologies for years under its Google Brain and DeepMind divisions. OpenAI changed the game by creating an application that people could use to interact with and understand the potential of generative AI.
"ChatGPT gave gen AI this face, something people could interact with," Larson said in a panel. "It shifted the conversation in a fundamental way. People are starting to get it. They're starting to see it in a more practical way, how AI fits into their lives."
AI startups on the rise
While incumbents have quickly moved on the trend, there's a lot of room for smaller startups that can target different areas with niche generative-AI models. For example, Anthropic recently raised $300 million from Google. Stability AI helped build an open-source image model called Stable Diffusion. The popular app Lensa, which can take selfies of individuals and artistically render them as portraits in various styles and settings, relies on this technology.
"There are many different directions to go in," Amodei said. "There will be infinite customization and that creates space for innovation."
These tools also won't eliminate the need for artists and other information professionals as human expertise and taste will still be required, executives said. Instead, they will increase efficiency. Masad said during a panel that more people will be able to write code, while expert software developers will just become better at their jobs as AI teaches them new tricks they hadn't known before.
Industry insiders at Gen AI echoed the same idea, painting a picture of a world where generative AI does not threaten humans, but enhances them with new superpowers.
"It's super important to remember this technology will be used for a lot of good," Dholakia said.