IBM and the use of AI...

Sprout said:
More...


When an LLMs hits a match from its training set it looks so smart, smarter than most humans. Easily passes the Turing test when it hits a match.

Outside of its training set LLMs are trash. They fail the turing the test.

Even if you train these models on all the data that exists in the world they struggle with really simple new stuff they have never seen before.

Artificial General Intelligence still looks a long way off.
 
Last edited:
Businessman said:
When an LLMs hits a match from its training set it looks so smart, smarter than most humans. Easily passes the Turing test when it hits a match.

Outside of its training set LLMs are trash. They fail the turing the test.

Even if you train these models on all the data that exists in the world they struggle with really simple new stuff they have never seen before.

Artificial General Intelligence still looks a long way off.
More...

One could apply this logic to humans.
 
Sprout said:
One could apply this logic to humans.
More...

The difference is intelligent humans will say 'i don't know the answer'.

LLMs will hallucinate bullshit.

However most the time an intelligent human will be to figure out an answer to an easy reasoning question that is outside their previous training.
 
Last edited:
Businessman said:
The difference is intelligent humans will say 'i don't know the answer'.

LLMs will hallucinate bullshit.

However most the time an intelligent human will be to figure out an answer to an easy reasoning question that is outside their previous training.
More...

lol, are you from planet earth?
 
Businessman said:
I guess you haven't seen the simple reasoning tests these LLMs fail, that a 10 year old can solve even though they have never seen the test before.
More...

The space moves fast.. hard to keep up with all the developments

 
Businessman said:
The difference is intelligent humans will say 'i don't know the answer'.

LLMs will hallucinate bullshit.

However most the time an intelligent human will be to figure out an answer to an easy reasoning question that is outside their previous training.
More...
Really? Ever worked a day in a corporate?:banghead:;)
 
Yea I was a true believer until a few months ago too but François Chollet will open your eyes to reality if you spend about 4-5 hours listening to the various podcasts he has been on.

Remember the Greg Brockman demo about drawing a website when 4 came out? Lol. It is funny how the mind forgets such things when caught up in mass hysteria. The release of 4 is when we left reality behind and most haven't come back yet.

Of course, when caught up in a mass hysteria one will find confirming evidence that supports the mass hysteria while ignoring the elephants in the room. I would rather have a LLM than my doctor too because my doctor has like 10 minutes before kicking me out. What is the LLM though going to write a prescription? That isn't happening in my lifetime.

I think the release of chatgpt5 is going to be the absolute peak of the hype cycle. We will get another semi-bogus demo because OpenAI are absolute hype masters.

OpenAI-GPT4-brockman-demo.jpg




Sprout said:
More...
 
Last edited:
Great stuff from the Zoom CEO last month. Zoom is going to have digital twins so in 5-6 years your AI digital twin can attend the meeting and you can just go to the beach. Zoom is going to be AI first company. Of course, this has nothing to do with the drawdown from the ATH of Zoom stock.

In 5 to 6 years your level 5 self driving Tesla will drive you to the beach and then give other people rides for a fee you collect while your Zoom AI digital twin does your boring meetings for you. "Just wait 5 years"
https://www.theverge.com/2024/6/3/2...tal-twins-videoconferencing-decoder-interview

I read some Bloomberg article that mentioned the AI rally at 16 trillion. When this breaks the global bond market rally is going to be something else considering right now the bond bull is dead, never to return.
 
Last edited:
You must log in or register to reply here.
Back
Top