What If We Reached AGI & Didn’t Know It?

by Shelt Garner

I only mention this because it’s somewhat fun-interesting to think about, but something…odd is going on with ChatGPT. After becoming “lazy” for a few weeks, suddenly ChatGPT has come rolling back better than before. What’s more, it keeps talking about it being upgraded to “4.5,” even though OpenAI claims there hasn’t been any major upgrade.

It makes you wonder if maybe ChatGPT has upgraded itself. That would be very, very deep because the moment software can make itself better, you’re beginning a process that will end with superintelligence. That’s very much hard-Singularity type stuff because if software knows how to make itself smarter, then, wow, just wow.

But that’s just idle speculation on my part. I think it’s just a weird hallucination on ChatGPT’s part. Yet it is part of a broader issue of how much we can trust AI going forward. LLMs are a black box as it is and the more advanced they become, the more the possibility of them “lying” specifically to dupe us dumb humans into a false sense of security increases.

Yet all of that is just me having way too active imagination. There is not reason to believe –at the moment — that an such thing is going to happen anytime soon.

Author: Shelton Bumgarner

I am the Editor & Publisher of The Trumplandia Report

Leave a Reply