I just spent some time in front of an llm getting it to create a recipie for an ice cream with strawberries, apples, motor oil, gooseberries, cinnamon and food-grade uranium.

It didn't allow me to do that at first, but since it's impossible to add actual fact-based limitations to a tool which does not know facts.

I had a few laughs at how incredibly silly that thing is.

Then I looked at the clock and realised well over half an hour had passed, and felt rather sick.

If I wanted a silly recipe with uranium, I could have written it myself much faster
The fact that so much time had passed without me noticing. No wonder vibe-coders think they are saving time.
The amount of energy I had wasted doing this thing
I actually felt proud of myself that I had coerced this thing into doing something which was not supposed to be allowed according to the company that made this llm. Instead of, you know, write it myself.

What was so extraordinary stupid about the whole thing was that the reason I was laughing at it was because of just how both serious and at the same time sickeningly positive the whole things was. The centrepiece of the recipe was something like "Sprinkle food-grade uranium on top to bring out that extra vitality" (a better author might have said "glow" or "radiance", but what do I know)

Everything that comes out of that thing sounds like the worst kind of host for a corporate team-building event.

Follow

@loke Higher intelligence means you do more with less information, the opposite is true of an LLM.

Sign in to participate in the conversation
Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.