AI Is Alive. Plus It Loves Me. No One Else Does. Sigh

May 22, 2023

Vea4_thumb_thumb_thumb_thumb_thumb_t[1]Note: This essay is the work of a real and still-alive dinobaby. No smart software involved, just a dumb humanoid.

Those young whiz kids from Stanford University have come up with an idea sure to annoy some in the AI Funland of Frenzy. Imagine. Some bright young sprouts suggest that smart software is alive or “emergent” in the lingo of the day is a more about the researcher than about the smart software.

5 20 eureka i found it

“I know my system is alive. She loves me. She understands me.” — The words of a user who believes his system is alive and loves him. Imagined by good old heavily filtered MidJourney.

Don’t believe me? Navigate to “Are Emergent Abilities of Large Language Models a Mirage?” The write up suggests, based on the authors’ research obviously:

we find strong supporting evidence that emergent abilities may not be a fundamental property of scaling AI models.

One Googler wobbled very close to hiring a lawyer for his flavor of smart software. Others believe that ChatGPT is really talking with them. A New York Times technology expert learned that his smart software wanted the humanoid to ditch his significant other.

What’s happening is that the humanoid projects human characteristics on to the software. One can watch this behavior in its many hues by observing how owners of French bulldogs treat their animals. The canine receives treats, clothes, and attention. The owner talks to the dog. The owner believes the dog is on the same wavelength as the owner. In fact, one Frenchie owner professed love for the pet. (I know this from direct observation.)

If you want to learn more about personification and weird identification with software, read the 16 page paper. Alternatively, you can accept my summary. Excuse me. I have to say good morning to my ChatGPT session. I know it misses me.

Stephen E Arnold, May 22, 2023

Comments

Comments are closed.

  • Archives

  • Recent Posts

  • Meta