In the mid-20th century, the world entered the information age, with the shift from industry to information technology. The era began with the miniaturization of computers and culminated with the invention of the World Wide Web, which brought access to information within the reach of almost everyone. Today, with the rise of AI, those days are over, according to some technology leaders, and a new technological era has begun.
“We have moved from an information age to an intelligence age,” Prakhar Mehrotra, PayPal’s senior vice president and global head of AI, said at the press conference. Fortune Brainstorm AI Conference earlier this month.
This “intelligence age” is marked by industries moving away from the data storage and retrieval model, Mehrotra said. Fortune journalist Sharon Goldman. Instead, through AI capabilities, data can be generated more spontaneously, with the ultimate goal of achieving autonomy in certain parts of the workplace.
Companies are rushing to apply AI – with its promises of increased productivity and output – to their respective workplaces, but their successes have been mixed. A month of August MIT study found that 95% of enterprise workplace AI initiatives have failed to achieve rapid revenue acceleration.
“It’s going to be a journey… You have to get through this crawling, walking and running,” Mehrotra said. “I think that adage was true 10 years ago and it’s also true in this time.”
The future of AI factories
Marc Hamilton, vice president of solutions architecture and engineering at Nvidia, interviewed alongside Mehrotra at the conference, said the future of AI development in the workplace will come through investment in AI factories, on-premises or in the cloud. Indeed, the data necessary for businesses to operate will no longer be primarily collected by humans or computers, but rather generated by AI.
“When you say, ‘Generate a PowerPoint slide that says this,’ or ‘I’m working on this coding feature, can you go in and generate some code?’ “It’s not about retrieving it from the database, it’s about taking a model and generating that data,” Hamilton said.
Mehrotra noted that for companies to effectively scale the computing power needed to create this data, there needs to be a new atomic unit prized by businesses: tokens, or the fundamental component of text that AI needs to understand and process language. Tokens are both the snippet of information used to train the data, as well as what is generated by the AI after a model receives a prompt.
“Every company needs to think about their data in terms of tokens, because then they can derive this information from it,” Mehrotra said.
Measuring inputs and outputs, token generation has become a key metric for tech companies in particular. In May, Nvidia boasted that Microsoftwhich uses Nvidia’s chips, generated more than 100 trillion tokens in the first quarter of this year, a five-fold increase year-on-year. These production indications can help these AI companies sell themselves to investors and increase their valuations. data shows that the correlation of tokens with demand and profits is lower than tech companies suggest.
Mehrotra and Hamilton agree that many companies today see the value of tokens in empowering AI capabilities, but are considering how best to tailor them to their needs, such as which tokens should be acquired or purchased, which should be generated internally, and for what purpose? Each company then has its own AI factory, both receiving tokens and producing tokens that have value.
“I see it as just building that muscle,” Mehrotra said. “Like all employees start thinking in terms of tokens, in terms of generation processes, then yes, it’s a different company.”