News AI Companies Seeking AI-Produced Data for Recursive Training

[Admin]

A number of AI companies including Microsoft, OpenAI and Cohere have been deep in research in an attempt to curb AI data costs. The plan? To have AIs generate new, synthetic datasets that can then be used to continue their own training.

AI Companies Seeking AI-Produced Data for Recursive Training : Read more

 

[PlaneInTheSky]

“If you could get all the data that you needed off the web, that would be fantastic,” said Aidan Gomez, “In reality, the web is so noisy and messy that it’s not really representative of the data that you want. The web just doesn’t do everything we need.”

What a bunch of baloney, just come out already and be honest. AI is as dumb as a rock, it's simply machine learning where the AI is a dumb parrot copy and pasting parts of sentences.

And the problem is not that there isn't enough information, there is tons of information, it's that AI has no idea what it's reading and has to be trained on millions time more data than even a 5-year-old would need to produce any reply that makes any sense.

AI is a billion $ scam created by Big Tech and Nvidia/AMD, meant to convince companies that the unreliable and inconsistent parrot replies from AI are of any use to them.

ChatGPT use is in the gutter.

 

[everettfsargent]

Oh come on PITS,

AI eating its own children, what could possibly go wrong? 😀

 

[tennis2]

Garbage in, garbage out