By skyforbes I’ve noticed an increase in posts where people attribute capabilities to language models that exceed what they are designed to do. From a Computer Science perspective, these systems generate outputs by statistical patterning, not by reasoning or internal experience.
In some discussions, terms like ‘recursion’ or ‘resonance’ get used in ways that don’t match their technical definitions. This can make it difficult to talk about how these systems actually work.
I’m interested in whether clearer, standardized explanations from AI companies would reduce misunderstandings. For example, a brief overview of how language models generate text and what kinds of tasks they are suited for.
How common is this kind of misunderstanding in your experience, and what communication tools would be most effective for addressing it?