OpenAI’s Mathematical Basics are Queryed: Internally Enclosed

An interesting criticism for OpenAI came on the social media agenda. In a shared article in Mastodon, the company’s owns mathematicians to fully understand the mathematical structure of its own technology. This claim reveals a paradox that we face in the artificial intelligence industry: Leading companies can even disclose deep mathematical mechanisms of their systems.

This concern points several important points:

• Complexy problem: The large language models of today work with billion parameters, and how these systems decide exactly how open
• Activities: Specialists in technical teams can not have enough depths about mathematical foundations of models
• Control challenges: If the internal teams are not fully understood the system, this process for external supervisors is becoming more difficult
• Payment money: The rapid growth of companies can indicate that mathematical discipline remains missing

Such criticisms open an important dimension in the discussion of artificial intelligence security and transparency. If developers even offer the world of systems they don’t understand exactly, it’s worth mentioning what kind of risks will bring together.

Such internal accounting throughout the industry shows that artificial intelligence technology should be built on more robust mathematical foundations.