Hits and Whiffs

At the halfway point, our 2025 predictions are mostly the latter

We’re at the halfway mark of 2025 and it’s time to check in on JEP’s annual predictions, fearlessly reporting on our hits and whiffs in attempting to discern where our world is headed. If anything, our performance so far has underlined, in triplicate, the truth of Yogi Berra’s dictum that it’s hard to make predictions, especially about the future. We’ve had a preponderance of whiffs at the halfway point, but that may owe as much to the lack of analytical discipline at JEP (and discipline in general) as it does to Yogi’s wisdom. So, instead of beating our chest we stand at the halfway point chagrined, humbled, chastised and corrected.

But then again, who could have predicted what we’ve seen so far in 2025, from an imperial presidency and a 12 Day War (maybe) to Trump watches, Trump cellphones and the mantle of global shuttle diplomacy being donned by a real estate developer from the Bronx?

To paraphrase the great Mink DeVille, it’s a mixed-up, shook-up world. Here’s how we fared.

“Change will continue to sweep through the media industry, good and bad,” we predicted. We were right about the continuing decline of print newspapers (specific data for the full year of 2025 isn’t available, but projections suggest the U.S. will have lost one-third of its newspapers since 2005 by the end of 2025, leaving roughly 5,600 newspapers remaining, with 80% being weeklies). But we were wrong about Jeff Bezos selling the Washington Post; he’s still hanging on even as discontent continues to decimate the morale of the newsroom. Compared to the costs of his upcoming wedding, the Post is a rounding error.

As digital content continues to swamp traditional print, supercharged by the rise of AI, we borrowed a prediction from Sara Morrison at the Nieman Lab, who warned about the rise of “AI slop” — content that is wrong, nonsensical or boring. She was right. Approximately 18 percent of financial consumer complaints are at least partially generated by AI, according to a March 2025 report, and it’s estimated that over half of longer English-language posts on LinkedIn are AI-generated. Overall, so-called AI slop sites increased 717 percent over the past year. Without curation, AI is worthless or worse.

(Related to AI slop is the intersection of AI and polarization, which colleague Bob Feldman, founder of The Dialogue Project, and I explored in this piece published by the University of Southern California.)

We also predicted that Polymarket, the crowd-sourced prediction marketplace, would supplant media punditry in its oracular power. It’s hard to argue we were wrong about that. Data trackers put the accuracy of Polymarkets at 88 to 94 percent. Polymarket has the current chance of a recession at 26 percent, so we should feel good about that.

We whiffed big time on DOGE, predicting it would generate savings of $1 trillion or more and a reduction in regulatory overhang of 20 percent. We were so wrong. DOGE appears to have achieved little besides cacophony and chaos. We didn’t see DOGE bro Vivek Ramaswamy falling out of favor so quickly, but we did foreshadow the breakup between Trump and Elon; there just wasn’t enough oxygen in the White House for both of them.

Our prediction that the U.S. economy would continue to outpace the rest of the world was notionally correct (we predicted 3 percent GDP growth, which currently stands at 3.4 percent), but we’re still a ways off from the 45,000-point Dow we predicted and the market for IPOs and M&A is still far below the level we envisioned.

Ditto for energy. We got the trend right (demand continuing to outpace supply), but oil prices, which we predicted would hit $50, have hovered in a range of $65 to $75. “Data will be the big energy hog” of 2025, we predicted, but even Stevie Wonder could see that one coming. We were correct in predicting the rise of behind-the-meter power solutions for Big Data. Chevron, for example, announced a partnership with Engine No. 1 and GE Vernova to build out a power generation business for data centers. They’re just one of many on the same track.

We correctly predicted a rapid acceleration in the use of AI agents to do mundane tasks. Surveys show that 79% of companies are already using AI agents, and 96% of enterprises plan to expand their use within the next year, with half aiming for significant, organization-wide deployment. That’s one of the reasons why Amazon CEO Andy Jassy said earlier this year that AI agents will eliminate thousands of corporate jobs. We did not, however, foresee the rise of AI chatbots and the damage they can do to users with psychological vulnerabilities, as chronicled in this Times piece.

We’ve learned over the years that JEP is supremely unequipped to make sports predictions and that hasn’t changed much. It doesn’t look like there’s a $700 million contract for an NFL quarterback on the horizon (Patrick Mahomes holds the benchmark at $450 million), the San Francisco 49ers didn’t let go of Brock Purdy but instead inked a new $265 million deal with him, and the Eagles, not the Bills, prevented the Kansas City Chiefs from notching a three-peat in the Super Bowl. We predicted Steph Curry would announce his retirement this year, but that isn’t likely until his contract expires in 2027. Shohei Otani, however, is on track to fulfill our prediction of 55 home runs this year (he’s hit 26 beaners as of this week) and an American woman, as predicted, won a Grand Slam tennis title (thank you Coco Gauff).

We were right about Joe Biden. As predicted, in the last hours of his term he issued a blanket pardon for his brother James and other family members. Wonder what they have to hide.

We predicted that “to maximize the sales price, Denmark will list Greenland with Sotheby’s auction house,” but that was just for a laugh. However, we’ve still got six months to go.

Finally, we stand by one of our boldest predictions — Joe McLaughlin, the great raconteur, self-anointed expert on everything, and valued paid subscriber, will indeed celebrate his 90th birthday.