Introducing AI Answers in Forum Search (Beta)

[Burnside]

AI (Intelligence my a*se) is the work of the devil.

I hate it with a passion that knows no bounds.

Part of the issue for the uninformed is that they often take AI responses as fact and don’t have the basic knowledge to judge their veracity. All too often the “I think I smell…” video from the Chewing The Fat tv comedy comes to mind…

 

[HansD-NL]

The biggest part of the issue I think is that the AI models claim that they will learn from mistakes they make but that simply is not the case. It's a static language model that can include newly learned stuff after a new model has been generated/released.

 

[harkpabst]

The biggest part of the issue I think is that the AI models claim that they will learn from mistakes they make but that simply is not the case. It's a static language model that can include newly learned stuff after a new model has been generated/released.

Models based on neural networks actually can learn from input without the need for any regeneration, if they are configured to allow this. Systems that are not based on neural networks but just processing and collecting data based on fixed algorithms are not even considered AI by today's standards.

 

[HansD-NL]

Models based on neural networks actually can learn from input without the need for any regeneration, if they are configured to allow this. Systems that are not based on neural networks but just processing and collecting data based on fixed algorithms are not even considered AI by today's standards.

Well in that case both ChatGPT and Grok are not AI. Both explained it the way I described above when asked about it. Can you give an example of a real AI (neural network)? I'm not an expert but an irregular user and don't like the way it often assumes things and all that comes with that.

 

[JAB]

So only another data kraken?