Part 2 of the Symbol Grounding Problem: is grounding something we really need to worry about? Some AI programs sure seem to accomplish a lot without it ...
Someone is fighting for you
- she/her
- writerofminds.blogspot.com/
Loving author of Acuitas the symbolic AI and other robots-in-progress. Aspiring fiction writer, video game appreciator. Satellite electronics are my day job.
Trying my hand at more of a theory piece: the Symbol Grounding Problem and how I intend to deal with it in my particular AI project.
Here's my obligatory Chat-GPT test: can it write my activist e-mails?
Since I've run similar tests on previous LLMs, you can trace the improvement. And there's real improvement. But though Chat-GPT produced one or two sendable notes, it still struggles.