Jun 2, 2026

Welcome back to The Hidden Layer. I’m Ian Krietzberg, fresh off a very timely rewatch of The Big Short. As my partner Bill Cohan wrote on Sunday, this isn’t the first time investors have lost their minds over a new technology—even if each new generation insists “this time will be different!” The very next day, Anthropic confidentially filed to go public. Hold on to your 401(k)s…

In today’s jam-packed issue, my conversation with Marc Zao-Sanders, the researcher behind Harvard Business Review’s new report on actual A.I. use cases. (The results may surprise you.) Plus, news and notes on the latest I.P.O. chatter, a CNN lawsuit, and A.I.’s so-called “jagged frontier.”

Also mentioned in this issue: Sam Altman, Dean Baker, James Uthmeier, Andrew Dai, Trump, David Sacks, Dean Ball, Brad Carson, Nick Reese, and more.

Three Things You Should Know…

The E.O. has arrived!: Earlier this afternoon, President Trump quietly signed a much-hyped cybersecurity executive order directing the Departments of Defense and Homeland Security to upgrade their cybersecurity systems to defend against Mythos-like “advanced A.I.” and calling for a voluntary framework for frontier labs to grant the government early access to new models up to 30 days before their release. It is, as former White House policy advisor Dean Ball pointed out, almost identical to the leaked drafts of the prior executive order that Trump was about to sign last month—before former A.I. czar David Sacks expressed his concerns that overregulating the industry would be a gift to China. Looks like he sorta got his way: The original version of the E.O. called for 90-day predeployment access; Sacks seemed pretty happy with the final result, calling the switch to a 30-day time period a “game changer.”

Former Congressman Brad Carson, who co-leads the Public First super PAC, called the E.O. “a victory—a huge one—for those who want reasonable regulation.” Nick Reese, the former director of emerging technology at D.H.S., told me that there are a number of questions posed by the order, though he added that the “willingness of the administration to intervene even a little is significant.” Kent Walker, Google’s president of global affairs, called it an “important step forward.”
The joy of SpaceX: Just in time for the I.P.O. of SpaceX, which is targeting a roughly $2 trillion valuation despite $5 billion in losses, the Nasdaq has changed its rules for index inclusion for megacap corporations—companies in the top 40 of the Nasdaq-100 based on market cap. Instead of a months-long “seasoning” period intended to ensure price stabilization, SpaceX will now likely be included in the index within 15 days of its I.P.O. At the same time, the S&P is considering new rules that would reduce its seasoning period from 12 months to six, and would enable exceptions to its GAAP profitability requirements for megacap companies.

That’s great news for OpenAI and Anthropic, which have also filed to go public in the coming months despite multibillion-dollar annual losses. It’s a risk for passive investors—i.e., people who put their money into retirement accounts or E.T.F.s that track the big indexes—who will soon gain exposure to three highly valued yet deeply unprofitable corporations. Dean Baker, a senior economist at the Center for Economic and Policy Research, suggested in a post on X that “SpaceX shareholders, especially the accidental ones in index funds (good time to get out), could sue NASDAQ for these rule changes.”

A Nasdaq spokesperson declined to comment, but pointed me to a blog post from last month in which the company said: “Public markets have evolved significantly over the past decade, as companies stay private longer, list at larger scale, and come to market with more complex ownership and share structures. The consultation was about ensuring the Nasdaq‑100 continues to reflect the market it is designed to measure, as market structure evolves.”

A MESSAGE FROM OUR SPONSOR

Powering the Energy of Every Day

The electric grid is the backbone of America’s economy. A new Concentric Energy Advisors report highlights how continued investment in reliability and resilience supports homes, schools, offices, and communities across our nation. Watch to learn how investor-owned electric companies are meeting rising demand, maintaining reliability, and keeping customer bills as low as possible as we deliver the energy of every day for 250 million Americans.

CNN sues for peace: Last week, CNN joined the ranks of the New York Post, New York Times, Dow Jones, Encyclopedia Britannica, and Merriam-Webster by filing a lawsuit against A.I. search startup Perplexity. Similar to those other lawsuits, CNN is accusing Perplexity of illegally copying and distributing the network’s content. In a filing, CNN said that it wanted a licensing deal, and was in talks with Perplexity last year to secure one—but the two never agreed on terms.

Given how they’ve responded to previous lawsuits, it’s likely that Perplexity will dig in. After Dow Jones sued the company, in 2024, Perplexity dismissed the mere validity of media lawsuits against generative A.I. labs, writing in a blog post, “The common theme betrayed by those complaints collectively is that they wish this technology didn’t exist. They prefer to live in a world where publicly reported facts are owned by corporations, and no one can do anything with those publicly reported facts without paying a toll.”
More lawsuits…: Meanwhile, Florida Attorney General James Uthmeier sued OpenAI on Monday, accusing both the company and C.E.O. Sam Altman of allowing “a dangerous product to reach millions of Floridians.” The complaint opens with a screenshot of OpenAI’s pledge that ChatGPT is “built with safety in mind.” “Not so,” writes Uthmeier.

The lawsuit is seeking to prevent OpenAI from collecting data from users under 18 while requiring the company to more clearly lay out the risks posed by its technology. When I reached out to OpenAI, a spokesperson told me, “A.I. is a new and powerful technology, and we believe minors need significant protection, which is why we have put in place industry-leading protections and policies.”

Quote of the Week: The Jagged Frontier

“A.I. is a jagged frontier, and we’ve been seeing that not just for the past 10 years, but for the past 50 years. Essentially, math and coding is the new chess and Go, but we’re not seeing it generalize to visual reasoning.”
—Former Google Brain and DeepMind researcher Andrew Dai, in conversation with yours truly, employing the industry’s favorite term of art for the notion that A.I.’s leading edge is still… spiky. (Elorian, his new startup, is betting that visual reasoning, rather than L.L.M.s, is going to expand the frontier.)

Capital Intelligence

IBM and Red Hat unveiled Project Lightwell last week, a $5 billion venture committed to ensuring the security of open-source software in the age of A.I. attacks.
XCena, an A.I. compute and memory startup, closed a $135 million Series B round last week, valuing the company at $570 million.
Physical A.I. infrastructure startup Mecka has raised $60 million across two rounds—a Series A and a follow-on investment—at an undisclosed valuation.

And now for the main event…

Harvard’s A.I. Gut Check

An incisive conversation with Marc Zao-Sanders, author of Harvard Business Review’s latest report on how consumers are actually using A.I. Ubiquitous transformation? “It was mostly about a little bit of revenue growth, and then cost savings and time savings,” he said.

Ian Krietzberg

For the past three years, Marc Zao-Sanders, the brain behind the A.I. in the Wild research initiative, has been trying to figure out how people actually use the technology in real life. It’s a deceptively simple question, given the major developers claim they’re servicing billions of users, but it’s never been clear how those users are really quantified. Yesterday, Harvard Business Review published the third installment of Zao-Sanders’s annual report investigating that question—a large-scale study parsing a year’s worth of social media posts to identify over 12,000 A.I. use cases.

In the end, the top 10 represented an odd mix of practical and personal applications, spanning general advice, relationship guidance, tarot card readings, fan fiction, and “fun and nonsense.” For the second year in a row, therapy/companionship topped the list, while general troubleshooting took second place. Farther down, technical use of software came in at number five, and agentic operations ranked sixth. Across the top 100 use cases, there’s even more evidence that most people are leaning on A.I. to perform psychological tasks, or what experts refer to as “cognitive offloading”: things like “career advice” (number 24), “confidence boosting” (number 43), “raising kids” (number 36), “thinking better” (number 81), or “finding purpose” (number 85).

A few days ago, I caught up with Marc to discuss the study, and what it all means. As always, the following has been lightly edited for length and clarity.

A MESSAGE FROM OUR SPONSOR

Powering the Energy of Every Day

Work vs. Play

Ian Krietzberg: The industry is locked in a race to secure long-term, ideally paying users. But there’s been a divide in focus between the major labs making a play for consumer versus enterprise customers. Are you seeing either of these products becoming sticky, integral aspects of ordinary people’s daily lives?

Marc Zao-Sanders: I guess the fact that we have 12 and a half thousand use cases, which are generally positive, answers that question. All of this data is consumer, and with consumers… you don’t have any of the corporate restrictions, you’re not going to have any parental controls, so people can go to town, they can experiment, they’re not worried about reputational risk.

This isn’t in the article, but there’s a different dataset that we have of corporate use cases. This is both individuals at companies and the companies themselves. Obviously, when it’s a company, there’s a gloss to it—but with the individuals, there’s a little bit more cynicism. There’s quite a lot of people that feel a bit stuck at work. They’re a little bit worried about whether they’re breaking a rule, a little bit worried about whether their boss is looking, or colleagues could be looking, or where this data is going. There’s just a difference in the freedom with which people use it at home versus at work.

Still, the negative sentiment that we see on the consumer side, we don’t see it much on the business side. Partly, this is people protecting their positions at work, but I think there is also less cynicism in the business community.

How do the real enterprise-level use cases you’re seeing stack up against the sort of radical business transformation that’s being pitched, promised, and sold by A.I. companies?

At work there are these restrictions, so that sort of curtails what an individual would do at work, but I guess the distinction I want to make is what an individual will do and feel comfortable to do at work, versus the company saying, Okay, we’re going to transform this sales process. Of course, that’s very, very different from just deciding you’re going to use A.I. for something.

That said, even in the data, where we were looking at what the corporations were saying, they weren’t claiming major transformations. It was mostly about a little bit of revenue growth, and then cost savings and time savings.

“I Just Can’t Stop Using It Now”

When you first started doing this, A.I. was still new to people. In that sense, this makes your 2026 report the most interesting one so far—they’ve had time to adjust. We’ve also seen a growing sense of anger and anxiety surrounding the industry. I’m curious if any of that sentiment appeared in the data you analyzed.

I hardly saw it last year, and it barely picked up—if at all—in 2024. This year, just under 5 percent of the posts carry some negativity; the negative themes that stand out are A.I. replacing jobs, emotional manipulation, hallucinations, and thinking less critically. I think this is partly because the mood has changed, and also because there’s been enough time for us to analyze this—and obviously, there are far more people using it now.

There’s a lot more awareness now of the negative, or potentially negative, impacts that A.I. might be having on people’s thinking. A little bit further upstream from the content output that might be crappy and sloppy is actually what’s going on in our heads. I’ll read you a couple of quotes: “I’m stuck. I’m very strongly against A.I. and try not to use it as much as I can, but I feel like I’m now stuck. I just can’t stop using it now.” Another quote is, “I’ve allowed myself to become lobotomized by relying too much on A.I. agents.”

What do you think that means for the regulatory and legislative efforts that are just starting to ramp up?

A lot of people are using this technology. What there definitely hasn’t been is a five-year longitudinal study of how it’s impacting people emotionally, cognitively. Maybe some of the lawsuits will prompt, or are already prompting, action—like what they’re doing in Illinois and probably some other places soon. But to make a kind of obvious point, people can get around regulation in all kinds of ways. If people find it useful, many, many people will find a way. The law just takes so long to actually [catch up]. To expect that you’d have a widespread iron grip on this anytime soon, I don’t think that’ll be the case.

You’ve got a huge commercial engine, and then you’ve got people that want quick fixes, the convenience of the answers this technology can give. To get regulation to work fast enough and hard enough and be enforceable enough to work against those headwinds, I think it’s very, very hard.

That’s all for today. I’ll see you on Thursday.

Ian

Join Emmy Award-winning journalist Peter Hamby, along with the team of expert journalists at Puck, as they let you in on the conversations insiders are having across the four corners of power in America: Wall Street, Washington, Silicon Valley, and Hollywood. Presented in partnership with Audacy, new episodes publish daily, Monday through Friday.

A professional-grade rundown on the business of sports from John Ourand, the industry’s preeminent journalist, covering the leagues, players, agencies, media deals, and the egos fueling it all.

Need help? Review our FAQ page or contact us for assistance. For brand partnerships, email ads@puck.news.

You received this email because you signed up to receive emails from Puck, or as part of your Puck account associated with {{customer.email}}. To stop receiving this newsletter and/or manage all your email preferences, click here.

Puck is published by Heat Media LLC. 107 Greenwich St., New York, NY 10006