The gigantic and costly industry Silicon Valley is building around generative AI is still struggling to explain the technology's utility.

Why it matters: AI chatbots and image generators are making headlines and fortunes, but a year and a half into their revolution, it remains tough to say exactly why we should all start using them.

• The most common rationale is a kind of circular reasoning: Everyone's going to be using these tools, the argument goes, so you might as well get ahead of the parade.

Driving the news: This spring, a chorus of thoughtful critics has begun sharing the nagging feeling that, fascinating and alarming as generative AI may be, it doesn't have much practical use in their work or lives.

• In a recent podcast conversation, New York Times columnist Ezra Klein said, "I consistently sort of wander up to the A.I., ask it a question, find myself somewhat impressed or unimpressed at the answer. But it doesn't stick for me. It is not a sticky habit ... it's not really clear how to make A.I. part of your life."
• Software engineer Molly White, author of "Web3 is Going Just Great," reports that generative AI tools are "handy in the same way that it might occasionally be useful to delegate some tasks to an inexperienced and sometimes sloppy intern," but hardly a solid foundation for tech's next big platform.

Tech historian Margaret O'Mara compares today's AI industry to the big business of beaver-skins.

• In the 19th century, a bounty of beavers roamed North America, so many that demand for men's top hats drove a vast business in trapping for decades, until beavers were nearly extinct. Then hat-makers figured out how to work with silk instead, which they could have been doing all along.
• "I get a beaver-fur hat vibe from some of the AI conversations now," O'Mara told tech journalist Caitlin Dewey. "These companies have so many resources: so much money, so much talent, all these massive data centers, the ability to create incredibly powerful models. And so they are creating those models, and the market is growing to meet them."
• "But it's not always apparent if we really need this technology, in every case, or if anyone's asking those questions."

These doubters are not technophobes or Big Tech demonizers — they're experienced observers of the tech world who have voluntarily jumped into the AI pool, only to find it mostly dry.

Flashback: Every other major platform shift in tech — from the personal computer to the internet to the smartphone — also went through a period of everyone asking, "What do I do with this thing?"

• Tech historian Laine Nooney's book "The Apple II Age: How the Computer Became Personal" reminds us that it took years for consumers to be persuaded that PCs had broad value beyond the relatively limited initial use cases of spreadsheets and games.
• The industry bet on AI is that, as with those previous waves, if you get a new technology into a wide enough set of hands, users will discover what to do with it.

The other side: There is a broad consensus that generative AI already provides clear value in several specific use-cases.

• It makes a great coding assistant, relieving some drudgery in the programmer's life and providing help with unfamiliar details and even new ideas for solving thorny problems.
• It's rapidly turning into a better interface for non-programmers to use their computers and other digital tools more easily and fully.
• It can usefully digest information, summarize meetings and conduct research — as long as you never fully trust it, since you can never be sure it isn't making things up.

Yes, but: Many of the transformative leaps envisioned by AI promoters feel as distant from daily reality today as they were when ChatGPT was first unveiled.

• The realms AI is touted as transforming — such as health care, legal practice and education — are precisely the sorts of areas where generative AI's "mostly right but frequently wrong" accuracy rate could cause havoc.

What's next: Wharton professor Ethan Mollick told Klein he urges everyone to take 10 hours using ChatGPT or one of its competitors in their work.

• "Use it in an area where you have expertise," he advises, "so you can understand what it's good or bad at, learn the shape of its capabilities."

White, however, argues that the modest capabilities of today's AI tools mean the hype and cash-flooded field is heading for a giant let-down.

• She writes: "You can't build a hundred-billion-dollar industry around a technology that's kind of useful, mostly in mundane ways, and that boasts perhaps small increases in productivity if and only if the people who use it fully understand its limitations."