>> How far away are we from this being able to train, say, an LLM? Very far. Even if it was large enough and you could train it as an LLM, there's not way to r

>> How far away are we from this being able to train, say, an LLM? Very far. Even if it was large enough and you could train it as an LLM, there's not way to read out the weights or even the connectivity graph.