With the newest law school admissions disclosures coming out, something caught my attention. Alabama Law's incoming class has a 3.98 median GPA, a 169 median LSAT, and a 75th-percentile GPA of 4.06. Are we genuinely producing entire classrooms of intellectual wolves in ALABAMA??? 😠And it's not just Alabama. Look at Minnesota, California, Florida State, Ohio State, and Texas A&M. Academic credentials once considered extraordinary at elite law schools are becoming the median across numerous institutions.
For my European friends, imagine law schools across Bucharest, Warsaw, and Berlin suddenly reporting incoming academic credentials comparable to what Oxford or LSE reported fifteen years ago. Obviously, the grading systems aren't directly comparable, but that's precisely my question. Are we witnessing an intellectual revolution with hundreds of Einsteins walking around Tuscaloosa, Alabama, or have we spent decades optimizing the numbers that supposedly represent intelligence?
My argument is that American education has become an incentive system that behaves remarkably like an algorithm, rewarding numerical credentials under the promise of meritocracy. Students optimize grades and standardized scores. Schools recruit applicants who improve their admissions statistics. Those numbers influence rankings, reputation, applicant demand, and in turn increase tuition on their students. Everyone has an incentive to improve the measurement, so the system becomes increasingly efficient at producing exactly what it rewards.
Even the measurements aren't consistent. LSAC counts an A+ as 4.33, while some undergraduate universities don't award anything above an A. Two students can earn the highest grades their institutions permit, yet only one may have the opportunity to exceed a 4.0. We're comparing applicants using standardized numbers from institutions that don't even share the same grading ceiling.
Consider the University of Florida. UF Law's median GPA rose from 3.91 in 2025 to 3.97 in 2026, while its median LSAT increased from 169 to 170. Yet UF had reached No. 24 nationally in 2020 before appearing at No. 38 in the 2025 U.S. News rankings. These aren't directly matched years, and ranking methodologies change, but the contrast raises a question: What happens when a school becomes better at recruiting exceptional academic credentials without demonstrating equivalent improvements in its overall educational quality? Higher admissions medians don't automatically mean better teaching, student satisfaction, or professional outcomes.
Now look at Ohio State. Its Class of 2027 profile reported just 6% first-generation college students, while its 2025 ABA disclosures show that 90% of enrolled JD students received scholarships or grants, including 175 receiving full tuition or more. These statistics describe different cohorts, so they don't reveal individual recipients. But if first-generation representation remains that low, most scholarship recipients would necessarily come from continuing-generation families. That doesn't make those students undeserving. It raises a more important question: Who has the resources, preparation, and generational knowledge to compete for admission and the largest scholarships in the first place?
And who pays the bill? LSAC's research found that first-generation college graduates anticipated almost 20% more law school debt than continuing-generation students, while Black first-year students anticipated 43% more than the overall average. Schools can distribute enormous amounts of scholarship money while disadvantaged students remain underrepresented and disproportionately indebted. In 2022, Yale Law Dean Heather Gerken publicly criticized rankings for encouraging schools to recruit high-scoring applicants with scholarships rather than prioritize financial need. The problem isn't that nobody recognizes these incentives. People running the institutions have already called them out. Yet the financial incentives remain.
Remember Alabama football under Nick Saban? When NIL opportunities expanded, Bryce Young was reportedly approaching $1 million in endorsement opportunities before becoming the starting quarterback. Financial incentives for recruiting elite talent became increasingly visible. Law schools similarly use scholarships to attract applicants who improve institutional statistics. But here's the difference: Alabama football eventually has to play another team. Recruiting five-star athletes isn't the same as winning a national championship. So why are we treating law schools recruiting 4.0 students as though they've already produced exceptional lawyers?
This brings me to Goodhart's Law: "When a measure becomes a target, it ceases to be a good measure."
In April 2026, OpenAI published Where the Goblins Came From, explaining why its models developed a strange habit of referencing goblins and gremlins. A reinforcement-learning reward intended to encourage a particular personality inadvertently favored those responses. The models learned what received rewards, and the behavior spread beyond its original context.
Producing more goblins didn't prove a more intelligent model. It proved the model had become better at producing what its reward system favored.
Here's what fascinates me: AI has given us a laboratory where we can observe reward optimization at much quicker rates. Researchers train models, adjust incentives, and observe unintended behaviors within relatively short periods. Now imagine a similar process unfolding across American educational institutions over decades instead of AI training runs.
In our educational environment, Goblin = GPA and LSAT. But the deeper issue is compliance with the reward structure. Students learn to produce exceptional scores, universities reward academic performance, and law schools use scholarships to recruit applicants who improve their statistics. This doesn't mean high-scoring students aren't intelligent. It means increasingly perfect credentials don't necessarily establish equivalent improvements in independent judgment, creativity, or real-world problem-solving.
And we already have a stress test beyond admissions: the bar examination.
In fall 2025, 42,817 students entered ABA-approved JD programs nationwide, beginning an expensive professional education with no guarantee of eventual licensure. In July 2026, Florida reported a 75.2% first-time bar passage rate, meaning 557 of 2,250 examinees failed. While nationally, the numbers can increase in one location and go down in another. It reveals something worth examining: schools receive the prestige of extraordinary admissions statistics years before students confront the professional and financial consequences of the next benchmark.
And the consequences aren't hypothetical. A friend of mine graduated with $225,000 in student-loan debt, accumulating interest, and didn't pass the bar. Meanwhile, The Ramsey Show featured two married attorneys with $900,000–$1 million in combined student debt, discussing the possibility of working 80-hour weeks just to escape interest rates. One didn't pass the bar. The others did. All were still drowning in debt.
That's the difference between benchmarks and outcomes. Schools collect tuition, celebrate admissions statistics, and award degrees, but the financial risk largely falls on graduates. These incentives didn't emerge by accident, even if their consequences weren't all intentional. Money can also discourage dissent: a scientist dependent on institutional funding may think twice before challenging those controlling it. The system rewards compliance, charges for participation, and can make questioning it financially dangerous.
But here's my ultimate concern: What happens when Alabama has to play the real game? And I'm not talking about football. What happens when America's lawyers, doctors, scientists, and AI systems confront another country's professionals whose institutions spent decades rewarding experimentation, independent judgment, and actual problem-solving instead of academic benchmarks?
In AI Arena, a model rewarded for producing Goblin might dominate its training environment but get smoked by a model with stronger capabilities. What happens when a civilization spends generations mastering its benchmarks, only to encounter another civilization that spent those generations mastering the problems themselves?
Maybe Tuscaloosa really is churning out Einsteins by the hundreds. But GPA medians can't prove that. My fear is that we've optimized for Goblin because Goblin brings rankings, prestige, and money, while those with the fewest advantages carry the greatest financial burden. AI has simply given us a laboratory to recognize how a reward system can become extraordinarily successful at producing the wrong result.
I post this because it's personal, and I intend to back it up. I'm a first-generation Mexican American who's spent years developing my abilities through debate, litigation, government, and public service, only to face an admissions system that can reduce that experience to two numbers. The system rewards advantages I wasn't born with, and too often, those who break through still carry disproportionate debt.
I'm going after that 170+ LSAT, and then the bar. Not because a score or licensing board defines my intelligence, but because I refuse to let them define the limits of my potential.
0 comments