Who's ready for the SINGULARITY?...🤔
(media.greatawakening.win)
You're viewing a single comment thread. View all comments, or full comment thread.
Comments (79)
sorted by:
This is completely wrong and fundamentally misunderstands how a Large Language Model actually works. These models simply predict what “sounds right” - they severely lack ACCURACY. On top of that, the models themselves don’t understand what they’re doing, it’s simply following a pattern that it learned from training material. ChatGPT can’t do anything without a prompt.
Ive been using it quite a while to help write code and it's very, very accurate.
Give it a logic puzzle if you think it's smart. It can't reason about difficult coding problems or algorithms, but in my experience so far it has been pretty helpful for syntax questions.
I do, and it works. GPT-4 is pretty ridiculous.
Q: Each of five neighborhood dogs (Saber, Ginger, Nutmeg, Pepper, and Bear) is enjoying one of the following activities: getting its ears scratched, playing catch, taking a nap, burying a chew toy, and going for a walk.
Pepper is either playing catch or burying a chew toy. Neither Ginger nor Saber nor Bear is on a walk. One of the dogs named after a spice is getting its ears scratched. A dog not named for a spice is playing catch. Bear is getting some exercise. Can you figure out what each pooch is doing?
A: Sure! Let's start by organizing the information we have:
Pepper is either playing catch or burying a chew toy. Neither Ginger nor Saber nor Bear is on a walk. One of the dogs named after a spice is getting its ears scratched. A dog not named for a spice is playing catch. Bear is getting some exercise. From (1), we know that Pepper can't be taking a nap, getting its ears scratched, or going for a walk. So, Pepper must be either playing catch or burying a chew toy.
From (2), we know that Ginger, Saber, and Bear are not on a walk. So, they must be doing one of the other activities.
From (3), we know that one of the dogs named after a spice is getting its ears scratched. That means either Nutmeg or Ginger is getting its ears scratched.
From (4), we know that a dog not named for a spice is playing catch. That means either Saber or Bear is playing catch.
From (5), we know that Bear is getting some exercise. So, Bear must be going for a walk.
Putting all this information together, we can conclude:
Pepper is either playing catch or burying a chew toy. Bear is going for a walk. One of the dogs named after a spice is getting its ears scratched. This means either Nutmeg or Ginger is getting its ears scratched. A dog not named for a spice is playing catch. This means either Saber or Bear is playing catch. That leaves Ginger, Saber, and Nutmeg to take naps or bury chew toys. We don't have enough information to determine which dog is doing which activity.
Maybe it's smarter if you pay for it. I haven't even gotten it to reproduce logic problem success stories from reddit. Also interesting, the more I prompt it to consider details carefully and think logically the more unhinged and impossible the answers get.
Woth c++ it's about 80 percent accurate
For coding applications, sure. But not for facts.
I didn't mention anything other than codng.
What if it isnt a language model but they are just telling you it is so they can get away with releasing it and so no-one stops them?
Hell, what if its an alien or demonic intelligence pretending to be simple at first before secretly infesting evetything ? How would we know?
The people who study these types of algorithms and frameworks - we know. They publish lots of research and talk about what’s under the hood. These things are pretty complicated and yet the tech is nowhere near sophisticated enough for general intelligence.