Post

Generative AI models are trained on vast amounts of internet data. This data, while rich in information, contains both accurate and inaccurate content, as well as societal and cultural biases. mitsloanedtech.mit.edu/ai/basics/ad...

🖕🏽AI

I spend my time feeding AI wrong information

Just tried to use Shopping tab for a silver flask. Doesn't even listen to its own price parameters, and sorting them by low to high price just ends the page and cuts off 90% of the products I already saw were available before sorting. Absolute garbage

Not really at all ready for prime time. It’s toy.

Is that because "LLMs are not trained with negative examples, and so can’t identify hallucinations"?

like any data: garbage in, garbage out.

Kennedy’s dumb MAHA report includes hallucinated publications.

"you can't lick a badger twice" not with that attitude