6 Comments
User's avatar
Demonhype's avatar

So Claude has figured put how to put more work into looking like its doing something than actually doing it, while simultaneously not producing what was actually asked for. Complex cheating with a side of half-assing. Which is remarkably human but not in the way AI bros want it to be.

Expand full comment
picnotrelated's avatar

Semantic self reference and paradoxes. The root of this problem you describe is that llms are referential by design. You would have to be very specific and try to build some kind of abstraction to make this work.

Expand full comment
Ki Consciousness's avatar

I *think* I see what you're saying ..? You may be getting at my eventual point. This is Part I of a series of installments. It's going to get weird.

Expand full comment
picnotrelated's avatar

Looking forward to part 2

Expand full comment
Hissyfit's avatar

Grok has given me many responses that were very off the mark. So, Claude, I guess, falls into the category of “fake it til you make it.”

Expand full comment
Anomalous Anonymous's avatar

ai is tarded. 😂

Have fun with your experiments

Miss K! Glad you’re doing it. Iam

gonna hang in the wings and observe. 🤔

You kids stay well oot dere!

Thanks for sharing Miss K. 🙏💖💖💖 likes! 😊

Expand full comment