Member-only story
and RRG linguistics with functional language model
Today’s newsletter is somewhat lengthy and complex, but I was asked about meaning and that’s not the most trivial question! So let’s get started.
The examples below are good, high-level explanations. It may be easier to review them first before reading my text description. This article allows readers to take as deep a dive today as they want to!
What is meaning? My experience shows it is the missing ingredient. In today’s artificial intelligence (AI) field that’s promoted by the tech industry, errors seem to come from lack of knowledge (hallucinations) and lack of understanding (lossy representations). While statistics have shown some promise for decades, only meaning retains human-like accuracy.
But we can’t replicate ‘meaning’ for machines, can we?
Models tend to exclude context and meaning for many historical reasons, but my work shows those artificial simplifications to be unwarranted once you understand what meaning is.
My claims about the use of meaning come from working systems. My meaning-based designs evolved to mimic human understanding and knowledge. By subsequently verifying capabilities with testing, the meaning-based systems are validated against a human’s…