The Company That Builds a Data Ontology Will Rule an Industry




We have learned that the more parameters a model has, the more expressive its representation of the world and the richer its ability to mirror it. And the latest language models with a trillion parameters will soon be outpaced by even more powerful systems, with tens of trillions of parameters and more. Some have predicted that language models with as many synpases as exsist in the human brain- some 100 trillion connections - will be constructed within the decade. - by Alexande C. Kalp, The Technological Republic
In the modern field of AI, it is often said that the more parameters a model has, the more precisely it can represent and understand the world. However, no matter how enormous a model may be—even one with tens or hundreds of trillions of parameters—its performance cannot shine if the training data is not properly structured. If data remains merely as a lump of text or numbers, the model can hardly go beyond the level of statistically connecting scattered facts.
In contrast, data ontology systematically defines and standardizes the entities, properties, and relationships across an industry, enabling AI to understand data as a structured semantic network. In this way, the model can process more than just simple words or numbers—for example, it can recognize and handle complex information such as “which airline operated a specific flight, from which ...

제목처럼 팔란티어가 그런 기업이 되는 중이겠지요? 덕분에 조금 더 이 기업의 본질을 엿볼 수 있었습니다. 감사합니다.