Don't forget the plentiful of fanboys and sycophants who constantly sing OpenAI and ilks praises, sweeping the points you listed under the rug as small bumps on the road to their "greatness". They've certainly done their parts.
On a light note, just for a laugh, here is my recent interaction with Llama 3 the latest and greatest open source LLM from Meta. As expected I got correct solution for the classic farmer/wolf/goat/cabbage crossing river problem from Llama 3 as the problem must have been within the training set. I then asked a follow-up question "What if I have two wolves in this case?". Here is Llama 3's response:
"A twist! With two wolves, you can't leave them together with the goat, as they might harm it. Here's a solution:
1. Take a goat across the river first and leave it on the other side.
2. Go back to the original side and pick up one of the wolves.
3. Take the wolf across the river and leave it on the other side with the goat.
4. Go back to the original side and pick up the second wolf.
5. Take the second wolf across the river and leave it with the other wolf and the goat.
This way, you've never left the wolves together with the goat, and everyone is safe and sound on the other side!"
So this is the kind of system whose underlying technology will lead us to AGI? UBI? A bright future for humanity?
What a joke?