LLMs are not conscious in any meaningful sense, but this is not a good argument for why.
This argument is like declaring humans are not conscious because a single neuron connected to a single input is not conscious.
The whole point of neural networks is function through complex connections.