Articles — Adriana Vesa

I write to think clearly, and to make others think clearly. My essays take on questions the field too often avoids: What does it actually mean to understand AI? What happens when models consume their own output? Can algorithmic systems make ethical judgments — and should they?

Each piece is grounded in data science and machine learning, but written for anyone willing to think carefully. All articles are published on Medium and open to everyone.

Follow on Medium

Medium · AI & Society

Model Collapse · MNIST · Creativity

Everything Becomes an 8

When AI models train recursively on their own output, something irreversible happens. Edges erode. The unusual disappears first. What survives is the average — a blob that looks like everything and nothing at once.

This piece uses MNIST digit classification as a concrete lens for understanding model collapse — the phenomenon by which AI systems, trained on AI-generated data, progressively lose the variance that makes them useful. It is also a meditation on what this means for human creativity in an AI-saturated environment.

Read on Medium → April 2026 · 7 min read

Featured Essay

"The model has lost the signal and kept the certainty. That is worse."

Finance & Ethics

"Algorithmic trading is not about prediction. It is about the asymmetry of consequences."

Medium · Finance & AI

NLP · Trading · Ethics

Sentiment Analysis & Algorithmic Trading: Salvation or Damnation?

Can NLP models read the market's emotional temperature? An investigation into whether sentiment analysis genuinely improves trading algorithms — or merely gives a veneer of intelligence to systems that still cannot distinguish signal from noise.

This piece interrogates the intersection of natural language processing and quantitative finance. Where does legitimate signal extraction end and algorithmic mythology begin? And who bears the cost when the system is wrong?

Read on Medium → 2026 · Medium

More coming.

I publish on Medium as new ideas take shape — at the intersection of machine learning, philosophy of mind, ethics, and education. Follow my Medium profile to be notified when new essays go live.

Topics I'm currently thinking about: the epistemology of large language models, the ethics of AI in educational assessment, and what Mondrian's neoplasticism can teach us about dimensionality reduction.

Follow on Medium LinkedIn

AI · Ethics · Society · Education