Since the rollout of ChatGPT in November 2022, many people in science, business and media have become obsessed with AI. A cursory glance at my own published work from that period identifies me as one of the culprits. My defense is that I share with those other obsessives the belief that big language models are the vanguard of a game-changing transformation. Maybe I’m swimming in the generative Kool-Aid, but I believe that the advances in AI within our reach will change not only the way we work, but the structure of businesses and ultimately the course of humanity .
Not everyone agrees, and there has been a backlash in recent months. According to some experts, AI is oversold and overhyped now opinion. Self-proclaimed AI chief critic Gary Marcus recently said of the LLM boom: “It wouldn’t be surprising if this whole thing were to fail to some extent. Others argue that AI is stuck in the “by disillusionment.”
This week we received some data that doesn’t solve the bigger questions, but does provide a snapshot of how the US, if not the world, views the advent of AI and large language models. The Pew Research Center – which conducted similar studies during the rise of the Internet, social media and mobile devices – published a study on how ChatGPT was used, viewed and trusted. The sample was taken between February 7 and 11 this year.
Some of the numbers initially seem to indicate that the LLM controversy could be a parochial disagreement that most people don’t care about. A third of Americans haven’t heard it from ChatGPT. Just under a quarter have used it. Oh, and despite all the panic about how AI is going to flood the public square with misinformation about the 2024 election? So far, only 2 percent of Americans have used ChatGPT to get information about the presidential election season that is already underway.
More generally, however data from the survey indicates that we are seeing a powerful technology whose emergence has only just begun. If you accept Pew’s example as indicative of all Americans, millions of people are indeed familiar with ChatGPT. And one thing in particular stands out: While 17 percent of respondents said they’ve used it for entertainment and an identical number say they’ve tried it to learn something new, a whopping 20 percent of adults say they’ve used ChatGPT for their work used. That’s dramatic compared to the 12 percent who answered affirmatively when the same question was asked six months earlier — an increase of two-thirds.
When I spoke with Colleen McClain, a Pew research associate involved in the study, she agreed that it appears to be tied to other major technological shifts. “If you look at our trend graphs over time in internet access, smartphones and social media, some of them are certainly showing this increase,” she says. There was a leveling off for some technologies, she adds. But in the ships she mentioned, the plateau only came when so many people came on board that there were not many stragglers left.
What’s crazy about that sudden jump in ChatGPT business usage from 12 percent to 20 percent is that we’re only at the beginning of human collaboration with these models. And the tools to fully utilize ChatGPT are still in their infancy. That is changing quickly. OpenAI, the maker of ChatGPT, is going full speed ahead, and AI giants Microsoft and Google are still redirecting their workforces to redesign every product line to integrate conversational AI. And startups like Sierra, which builds agents for enterprise customers, enable custom applications that take advantage of multiple models. As this process continues, more people will use AI tools. And as the base models get exponentially better, I hear GPT5 will too appear this year? – that will make them even more convincing. This raises the possibility that the quality of virtually all work will depend on how well one can bring out the talents of a robot worker.
What technology from the past can help us understand the trajectory of the rocket ship we are on? Although the almost limitless ceiling of AI makes it difficult to find an analogue, I suggest starting to use spreadsheets. Dan Bricklin and Bob Frankston invented them in 1978, and a year later the concept was embodied in VisiCalc, which at the time only ran on Apple computers. Spreadsheets had a phenomenal and disruptive effect on the business world. More than just accounting tools, they ushered in an era of business innovation and revolutionized the flow of information within companies. Still, it took a few years for the business world to widely adopt spreadsheets. The turning point came with a new and more powerful product called Lotus 1, 2, 3, which ran on the IBM PC. The current and future startups in the AI world, such as Sierra, all hope to become the lotuses of our time, but also to become much more consistent and sustainable. Spreadsheets are largely limited to the business domain. LLMs can apparently mess with it something.