What happened?

Today was a huge day at Thomas Kuhn Foundation. We've been on a mission to understand the value of our KGX3 engine and if it really can do what we believe it can do, and demonstrate that to potential users.

This morning I read over our recently completed benchmarking study—Cross-Field Validation of Kuhnian Classification of Scholarly Papers vs Citation Patterns—that shows, yes, it seems that we can indeed predict paradigm shifts long before the citations catch up.

Our founder, Dr. Khalid Saqr ran a test across 500 papers spanning physics, biology, computer science, medicine, and astronomy. KGX3 read each paper—just the text without considering the citation data—and classified them into one of five stages of scientific development based on Thomas Kuhn's theories: normal science, model drift, model crisis, model revolution, or paradigm change.

The results validated our hypothesis 🎉

Papers KGX3 labelled as revolutionary or paradigm-changing had attracted significantly more citations and generated more volatility over time, exactly as Kuhn predicted. This isn't guesswork—it's validated, deterministic pattern recognition that identifies paradigm shifts in scientific fields potentially years before those breakthroughs become obvious through Impact Factor.

So what?

Imagine getting alerts when paradigm-change papers appear in your domains—not six months after your competitors have already moved, but immediately.

Revolutionary papers signal emerging trends worth investing in; crisis-stage papers reveal established models failing, which often means market opportunities opening up.

Here's what makes KGX3 different from everything else out there: citation-based tools tell you what was important, often years after publication. KGX3 tells you immediately what will be important based on the thinking in the research itself. This is predictive, not retrospective. That timesaving advantage matters enormously.

—

For publishers drowning in the submissions fire hose, increasingly created by AI, this means better editorial triage at scale. Which papers represent paradigm shifts that belong in top-tier journals? Which are solid but incremental?

KGX3 can classify at submission, before expensive peer review begins, helping match papers to the right journal tier whilst reducing screening time.

For pharmaceutical companies and manufacturers monitoring thens of thousands of new papers every month, this is an early warning system for licensable research and emerging commercial opportunities.

Universities and funding bodies face similar challenges. Which research programmes are generating genuinely transformative work versus incremental progress? Where should strategic investment focus? KGX3 provides epistemic analytics that show the shape of science changing in real-time across fields.

So, now what?

This benchmarking study used 500 papers.

Next, we're plan to scale to 100,000 papers, and pilot with publishers on real submissions rather than historical data. We want to A/B test with and without KGX3, measuring both time saved and prediction accuracy, tracking which KGX3-flagged "revolutionary" papers actually become high-impact in the coming years.

But for me today, reading through the validation results with coffee going cold on my desk, I just felt genuinely excited. This thing works! Not perfectly, not completely, but demonstrably enough that I can see the path forward clearly. That's a good day at work.

Read the full validation study KGX3 Public Benchmark