Thank you for the comment Allan! I laughed at that last bit :)

I wouldn't say LaMDA is anywhere close to human-level conversations just yet. Human language is already complex at the syntactic and semantic levels, but the pragmatic level is simply unreachable for today's AI.

Neither LaMDA, nor GPT-3, nor a hypothetical GPT-4, nor any other neural network-based system could navigate the pragmatic level of human communication. They simply don't have access to the context. That's something to think about for the future of NLP.

Cheers :)

Writing at the intersection of AI, philosophy, and the cognitive sciences | | Get my articles for free: