In 2011, IBM
's Watson supercomputer beat two champions in a game of "Jeopardy!"
The company then spent the next decade, and billions of dollars, trying to use Watson's artificial intelligence capabilities to solve a broad set of healthcare challenges, from helping doctors diagnose diseases based on symptoms to recommending clinical trials.
That effort failed. In January, IBM announced it was selling Watson for parts to PE firm Francisco Partners
.
AI technologies have come a long way since that game show triumph. Some AI, such as those recommending ads on Google or detecting cancer on medical scans, have become part of everyday life. While improvements for these types of AI have been mostly incremental, over the last year machines suddenly became good at generating images and writing text.
Although this so-called generative AI often creates nonsensical pictures and produces text bearing the signs of a machine, many investors are convinced the new technology will usher in a productivity revolution across various industries and forge massive companies along the way.
"[Generative AI] has the potential to transform lots of different workflows and every piece of application software in [the] enterprise," said Gaurav Gupta, a partner at Lightspeed
. "We can take this technology and customize it for a particular use case—could be advertising, could be gaming, could be entertainment, could be pretty much everything."
Earlier this month, Gupta led Lightspeed's $101 million seed round alongside Coatue
and O'Shaughnessy Ventures into Stability AI
, the developer of open-source text-to-image generator Stable Diffusion, at a reported valuation of $1 billion.
Significant investments like that reflect lofty aspirations. But those aspirations may collide with AI's long history of underdelivering on expectations.
To some, the current hype over generative AI is no different.
This article appeared as part of The Weekend Pitch newsletter. Subscribe to the newsletter here .
"Outputs from Stable Diffusion don't surpass human artists, so [it] won't be adopted widely" until further improvements are made, said Brendan Burke, a senior AI and software analyst at PitchBook. And, it can't be known how long it will take the technology to improve enough for use in design or product management, not to mention in gaming or movie production.
Just because generative-AI models have shown rapid innovations over the last couple of years, that doesn't mean they will continue to improve at the same speed.
A good example of that problem is self-driving cars. The technology has been promised for years, but so far autonomous vehicles have been rolled out only in a few limited pilot programs.
Will realizing generative AI's potential turn into the same long slog? It's possible, but investors and founders are flocking to the image-creation field even though the technologies are still experimental.
"It feels like hundreds or even more companies are now getting started [in this area]," Gupta said.
And, while these technologies are years, if not decades, away from generating text, images and video as well as humans do, some applications of generative AI are already being used in a business setting.
Companies such as Jasper
, which last week raised a $125 million round at a $1.5 billion valuation
, along with Copy.ai
and Regie.ai
, can help humans automate repetitive writing tasks such as sales emails and basic blog posts.
Meanwhile, in a recent article on its website, Sequoia
wrote that "generative AI has the potential to generate trillions of dollars of economic value."
So far, there's limited evidence to back up that kind of optimism, but predictions like that are nevertheless emboldening scores of founders to ask investors for giant valuations.
"I feel the market will be hyped for a while because people are looking for the next thing to hype," said George Mathew, a managing director with Insight Partners
, who co-led the firm's investment in Jasper. "But in the long run, a decade from now, we're going to look back at this and say, 'fundamental, world-changing things happened in the introduction of these models.'"
Investors' big bets on generative AI companies may end up revolutionizing business and entertainment—or they may turn into another Watson.
Correction: This article has been updated to say that George Mathew co-led Insight Partners investment in Jasper. The previous version said he led the firm's investment in the company (Oct. 31, 2022).
Featured illustration made with Stable Diffusion by Michaus/Shutterstock
This article originally appeared on PitchBook News
