Anthropic has a tough alignment interview. Like I aced the coding screener but got rejected after a chat about values. I think they want intense people on the value/safety side as well as the chops.
You misinterpreted nothing.
Are you trying to downplay the compute required to train and run inference on these large language models by stringing together some contrived comparison to the now 'uncool' Blockchain technology? That would be absurd.