Perhaps you'll address this in the future, but it seems clear that blockchain technology in general theory, and ICP in specific application offers to AI the same quintessential benefits it tends to offer to pretty much any application. What about the reverse? The USE of AI to power ICP dapps?
Either by replying here or in a future article, can you please cover current and future compatibilities or restrictions on the IC for AI with respect to hardware (e.g., NVIDIA GPUs or other), software (e.g., written in C, Python, etc.), and databases, since it would be helpful to know if existing models / software could be ported to the IC. I am particularly curious about how databases of any kind could be ported to the IC, even if they are just integrated somehow and not migrated to the blockchain.
Perhaps you'll address this in the future, but it seems clear that blockchain technology in general theory, and ICP in specific application offers to AI the same quintessential benefits it tends to offer to pretty much any application. What about the reverse? The USE of AI to power ICP dapps?
Either by replying here or in a future article, can you please cover current and future compatibilities or restrictions on the IC for AI with respect to hardware (e.g., NVIDIA GPUs or other), software (e.g., written in C, Python, etc.), and databases, since it would be helpful to know if existing models / software could be ported to the IC. I am particularly curious about how databases of any kind could be ported to the IC, even if they are just integrated somehow and not migrated to the blockchain.
Can't agree more. The move by Meta to open source both the Llama code and model weights is accelerating this exact trend.