Can you explain the scenario where Anthropic getting to “aligned” superintelligence first blocks their competitors from creating unaligned superintelligence?