Man and Machine: #Karpathy AI Exposure x Payments

Everyone's mapping AI exposure. I'm mapping augmentation.

Andrej Karpathy just scraped 342 occupations from the Bureau of Labor Statistics and scored each one for AI exposure. The results are uncomfortable if you work on a screen.

Average AI exposure across all jobs: 5.3 out of 10. But $3.7 trillion in wages sit in the "very high exposure" zone. 25 million jobs scored 8 or above.

The pattern is simple. If your output is digital, your exposure is high. If it requires physical presence, your exposure is low. Medical transcriptionists scored a 10. Roofers scored a 0.

Now zoom into payments.

Settlement. Reconciliation. Fraud review. Compliance reporting. Payment routing. Exception handling. Every one of these is screen work. Every one scores 8 or 9 on Karpathy's scale.

But here's what the exposure maps miss.

I run a fleet of AI agents across my company and personal infrastructure. The people I work with who are using AI to compress these workflows aren't getting replaced. They're running scope that wasn't possible 18 months ago. One person doing the reconciliation work of three. One compliance analyst covering the audit surface that used to require a team.

Exposure is real. But exposure is a productivity multiplier, not a death sentence.

The people who figure this out first build a lead that compounds every single day. And that lead doesn't shrink. It grows.

-dg

P.S. I wrote about what it actually looks like to run an AI fleet as a CEO in my newsletter, The AI Operator. Link in comments.

#AI #ArtificialIntelligence #Payments #Fintech #ManAndMachine #AIExposure #Karpathy #FutureOfWork #PaymentOperations #TheAIOperator