https://russian.lifeboat.com/blog/2023/12/this-paper-proposes-rwkv-a-new-ai-approach-that-combines-the-efficient-parallelizable-training-of-transformers-with-the-efficient-inference-of-recurrent-neural-networks Safeguarding Humanity Thu, 28 Dec 2023 05:47:55 +0000 hourly 1 https://wordpress.org/?v=6.4.3 https://russian.lifeboat.com/blog/2023/12/this-paper-proposes-rwkv-a-new-ai-approach-that-combines-the-efficient-parallelizable-training-of-transformers-with-the-efficient-inference-of-recurrent-neural-networks#comment-498153 Thu, 28 Dec 2023 05:47:55 +0000 https://lifeboat.com/blog/2023/12/this-paper-proposes-rwkv-a-new-ai-approach-that-combines-the-efficient-parallelizable-training-of-transformers-with-the-efficient-inference-of-recurrent-neural-networks#comment-498153 Only international researchers in contained labs should be able to reach such articles. Would we agree upon open nuclear weapons designs? Some say that AI is potentially more dangerous. Why is this information being published openly? The transformer paper led to AI that is being used to trick people (fraud). More advanced AI design will do much worse. These designed should only be in the hands of well-overseen researchers and developers. We can let the fruits of this research be available while internationally controlling the weights so that the brains are not used for harm.

]]>