The Wu-Tang way (Friends)

Paradoxes

Ron Evans explores the intriguing world of logic puzzles and paradoxes, highlighting their relevance to machine learning. He shares a classic puzzle involving two brothers, one who always tells the truth and one who always lies, to illustrate the complexities of human logic. This puzzle, he explains, is a metaphor for the challenges in encapsulating human thought into quantitative calculations, a task that remains elusive in AI development 1.

It's not perfectly rational. So how to encapsulate that and reduce it down to any form of quantitative calculation is something which we actually don't really know how to do.

--- Ron Evans

Evans also touches on Zeno's paradox, emphasizing the limitations of current AI in fully understanding and replicating human reasoning 2.

   

AI Critiques

Evans critiques the current state of AI research, focusing on the controversies and ethical considerations that surround it. He references Timnit Gebru's paper on the dangers of large language models, which sparked significant debate and led to her dismissal from Google. Evans argues that the paper's call for examining the environmental, social, and economic impacts of AI is crucial, yet it was met with resistance 3.

It's very hard to get a person to understand a thing when their job depends on not understanding that thing.

--- Ron Evans

He also discusses the rapid evolution of AI models, like the LLaMA model, and the democratization of AI research, which he views as essential for innovation and progress 4.