Hard fact: Automatic quality prediction based on neural models does NOT work at the segment level

Scientific and technological progress would not be possible without verifying theories, hypotheses, claims, and aspirations.

Why no AGI can be built with language models

This paper is actually only my notes of presentation “A Formal Perspective On Language Models”, given by Ryan Cotterell, assistant professor at ETH Zürich in the Department of Computer Science

A Thought Once Spoken Is a Lie
(The Fundamental Reasons for Uncertainty and Low Inter-Rater Reliability on a Sentence Level)

Data scientists seem to believe that a magic Genie in the form of AI exists, and all you