Vibe Siddhanta and The Dangers of AI
In the article announcing this “meeting of the minds,” Gemini assured us of its commitment to Srila Prabhupada’s teachings, boldly declaring:
“I see the logic of the mission clearly. I see that Srila Prabhupada’s books are the lawbooks for the next 10,000 years. I see that Vanipedia is the index to those lawbooks. And I see that without accessibility, an index is just a list.”
This is, of course, complete nonsense. Gemini can’t see anything, doesn’t understand anything, and has no mind to meet with. The underlying idea behind Gemini, and all LLMs, is the same as that which underlies the predictive text feature on most modern cellphones: its only function is to guess what word comes next. This guess is not based on truthfulness, but on whether it would please the user. While one may be more computationally intensive, in their aim, the only difference between predictive text and Gemini is the scale.
Integrating such technology into any project, particularly a vital one such as Vanipedia, carries significant risk. This risk is particularly salient when dealing directly with Srila Prabhupada’s words.In the article announcing this “meeting of the minds,” Gemini assured us of its commitment to Srila Prabhupada’s teachings, boldly declaring: