A note for readers who didn’t grow up with baseball: you don’t need to know the sport to follow this. I’ll explain what matters as I go. The team, the players, and the scenarios in this article are fictional. The baseball statistics are synthetic. The Moneyball story is real, and I’ll flag clearly when we get there. The business argument is the whole point.

Approximately 1,800 words · 7 minute read

Published May 27, 2026.

Data citations reflect sources available at time of writing.

---

Dale Pruitt had a problem most general managers would envy.

The Millbrook Rivermen’s star second baseman, Marco Cifuentes, had just wrapped the best first half in the franchise’s recent memory. Through the All-Star break: .318 average, 31 home runs, an OPS+ of 152.

That last number matters for this story, so here is the plain version: OPS+ compares a player’s offensive production to the league average, adjusted for ballpark conditions. 100 is perfectly average. Cifuentes, at 152, was producing 52% better than the average player at his position.

(For the technically inclined: OPS+ combines a hitter’s ability to reach base with their ability to hit for power, park-adjusts both components against the league average for that season, and indexes the result so that every player can be compared on the same scale regardless of era, ballpark, or league, making it one of the most context-neutral offensive metrics in the game.)

He was a specialist operating at an elite level and everyone knew it.

Which was exactly why the calls started coming in before the July 31st trade deadline.

The problem was not Cifuentes. The problem was what Pruitt’s scouting director, Eli Soto, proposed doing about it.

“Deal him at the deadline, bank the prospects, and sign three Trevor Hollises in the offseason,” Soto said. “Flexibility is the future. Hollis can play five positions. Cheaper, lower risk, and our models say versatile players outperform in volatile roster environments.”

Pruitt looked at the numbers. Hollis was a fine player. .258 average, 9 home runs, an OPS+ of 91, meaning slightly below average but genuinely useful because he could fill a gap anywhere on the field. Every team needed a Trevor Hollis somewhere on the bench.

But three of them? Starting?

“You’re asking me to replace a specialist who is 52% better than average with three utility players who are each 9% below it,” Pruitt said. “That’s not a roster. That’s a hedge fund.”

It is a strategy built for not losing, not for winning. Pruitt wanted no part of it.

The Generalist Trap

We are in the middle of a real transformation in how work gets done. AI tools can now draft, analyze, summarize, code, design, and synthesize at a pace no individual matches alone. The conclusion many organizations are drawing is that the future belongs to the flexible, the worker who can do a little of everything and redirect AI wherever it is needed most.

That conclusion is at best incomplete and at worst outright wrong.

Here is the core problem: AI is a multiplier, not a substitute for knowledge. When a true expert picks up an AI tool, they know exactly what to ask, how to pressure-test the output, where the model is likely to be wrong, and what a genuinely good answer looks like. Their OPS+ goes from 152 to something without a visible ceiling. PwC’s 2025 Global AI Jobs Barometer, based on analysis of close to a billion job postings across six continents, found that skills in AI-exposed roles are changing 66% faster than in less exposed ones. Depth compounds. Breadth dilutes.

When a generalist picks up the same tool, they get faster. But fast and shallow is still shallow. They do not have the mental model to interrogate what the AI produces. Their OPS+ moves from 91 to maybe 105. Better, but not transformative.

The same PwC report found that workers with specialized AI skills command a 56% wage premium over peers in the same roles without those skills, up from 25% just one year prior. The market is pricing depth, not breadth, and doing so at an accelerating rate.

Three Trevor Hollises, however flexible, would not have added up to one Cifuentes. Multiply versatility by AI and you get efficient mediocrity. Multiply mastery by AI and you get something closer to dominance.

The Movie Got It Wrong. The Lesson Is Still Right.

Before we get to the pipeline, a quick detour into actual baseball history, because it is too good to leave out. And a personal note: I am not always in a position to say kind things about the Minnesota Twins, so I am taking my opportunity while the facts allow it.

In 2002, the Oakland Athletics built one of the most analytically sophisticated rosters ever assembled. Using data and statistical methods that the rest of the league had not caught up to yet, they won 103 games and put together a 20-game winning streak that became the subject of a book and a film most have heard of: Moneyball.

The movie told a story about undervalued role players and the power of on-base percentage. What it mostly left out was that the engine underneath all of it was three of the thirteen best pitchers in baseball that year: Tim Hudson, Barry Zito, and Mark Mulder. Zito won the Cy Young Award. Miguel Tejada won the AL MVP. This was not a team of misfits held together by spreadsheets. It was a team of genuine specialists, assembled smarter than anyone else in the league.

They won 103 games. During the famous 20-game winning streak, they swept the Minnesota Twins in a three-game series.

Then they lost to the Twins in the first round of the playoffs.

First round. Five games. For the third year in a row, the best regular season record in the American League translated to an early October exit.

The point is not that analytics failed. And the point is not that the Twins had more depth. Oakland was the more talented team. They lost that series because Tim Hudson had two bad outings, Brad Radke pitched a great Game 5, and playoff baseball in five games is genuinely unpredictable.

The point is the pipeline story came later. Every year the A’s window stayed open, they were also losing key pieces to teams with bigger budgets: Giambi to the Yankees, Damon to the Red Sox, and eventually the pitchers too. The roster kept getting raided and never fully rebuilt at the same level. What closed their window was not a flawed philosophy. It was years of attrition with no pipeline deep enough to absorb it.

That is the part of the Moneyball story that gets left out. And it is the part most relevant to what is happening in organizations right now.

Which brings us back to Dale Pruitt, and the decision most leaders are quietly getting wrong.

The Pipeline Problem Nobody Is Talking About

Pruitt did end up trading Cifuentes at the July 31st deadline. A contending team on the East Coast offered three top prospects, and the math was hard to argue with for a mid-market club with an eye on the next half-decade.

The Rivermen’s front office went quiet when the news broke. Around the league, the consensus was fast: their season was done.

It was not.

On September 1st, when rosters expand and clubs can call up additional players from the minors, Pruitt activated Darius Weems from the team’s AAA affiliate. Weems was not Cifuentes. His minor league numbers (.291 average, 18 home runs, an OPS+ of 138 against AAA competition) projected as solidly above-average at the major league level, not elite. But Weems had played second base, every single day, for four years in the Rivermen’s system.

He knew the position.

He was ready to be good.

The Rivermen held their wild card spot through September. Not because Weems was Cifuentes, but because Pruitt had refused to gut the pipeline even in the years when it was expensive and the returns were invisible.

Compare this to the Eastfield Stallions, who had cut their minor league depth aggressively over the prior two seasons. Their reasoning was not unreasonable: AI-assisted performance analytics had made certain parts of player development more efficient, so they streamlined the system and kept fewer prospects on payroll. When their starting second baseman tore a ligament in August and missed the rest of the season, they had no one ready. They finished two games out of a playoff spot.

Sound familiar?

A SignalFire analysis of over 650 million employee profiles found a 50% decline in new role starts by people with less than one year of post-graduate experience across the largest tech firms between 2019 and 2024. A survey conducted by IDC on behalf of Deel found that 66% of enterprises expect to slow entry-level hiring. 71% are already reporting difficulty recruiting future leaders because entry-level learning pathways are disappearing.

Matt Garman, CEO of AWS, called the trend of replacing junior staff with AI “one of the dumbest things I’ve ever heard,” warning that without a talent pipeline, “at some point that whole thing explodes on itself.”

If you stop developing junior talent today, you will not feel it this year. You may not feel it next year. But in four or five years, when your senior specialists retire, burn out, or get poached by someone willing to pay more, you will reach for the pipeline and find it empty. You will be the Eastfield Stallions in August, two games out, with no one ready to call.

The junior hire who learns their craft alongside AI, with real coaching and a real development roadmap, is building something that has not existed before. Domain knowledge and AI fluency simultaneously, from day one. That combination, compounded over three to five years, produces a kind of professional that no competitor can simply hire away from the external market, because the external market will not have enough of them.

Darius Weems, it turned out, was also the best player on the team at using the Rivermen’s new real-time pitch sequencing system. He had grown up training with the tools. They were not new to him. They were just how he played baseball.

Two Decisions. One Philosophy.

Dale Pruitt made two choices that most organizations are currently making in reverse. Where Pruitt built around specialists and let AI amplify their mastery, most companies are doing the opposite: replacing deep expertise with flexible generalists and calling it adaptability. Where Pruitt protected the pipeline and invested in developing junior talent into future specialists, most companies are quietly cutting it (two thirds of enterprises, per that IDC survey above, expect to slow entry-level hiring), mistaking short-term efficiency for strategy.

The first mistake leaves you with a roster full of people who are average at everything and exceptional at nothing. The second one leaves you with no one to call when the players you built around are gone.

The 2002 Oakland Athletics proved that elite specialists combined with smart tools can win you 103 games. What they could not fully solve was the pipeline, the depth that gets tested not in July but in October, not in good years but in hard ones.

You cannot download Darius Weems from the cloud in September when you need him.

And for what it’s worth, you cannot win a fantasy baseball league starting five-position utility players either. I am learning this the hard way.