I think there’s an issue with lumping them all together here. There is a functional difference between a machine that does your homework and is designed to endlessly praise you, and a model specifically designed to identify viruses or medications.
That’s like saying you can have space exploration without ballistic missiles. Once the technology is there power and/or the market will do whatever it wants with it.
Except even moreso with language models. Here it is not just a similar technology, but the exact same vehicle.
Translation: The way scaling laws and gradient descent work for language models means that there is NO functional difference between a machine that does your homework and is designed to endlessly praise you, and a model specifically designed to identify viruses or medications.
I think there’s an issue with lumping them all together here. There is a functional difference between a machine that does your homework and is designed to endlessly praise you, and a model specifically designed to identify viruses or medications.
deleted by creator
In this case it’s basically the same technology (LLMs); you don’t get one without the other.
Sure you do. We don’t have to have every permutation of every technology.
That’s like saying you can have space exploration without ballistic missiles. Once the technology is there power and/or the market will do whatever it wants with it.
Except even moreso with language models. Here it is not just a similar technology, but the exact same vehicle.
Translation: The way scaling laws and gradient descent work for language models means that there is NO functional difference between a machine that does your homework and is designed to endlessly praise you, and a model specifically designed to identify viruses or medications.