We’d Lose

This chapter starts with a very short prologue – the only non-fiction – of an aztec warrior watching a spanish armada showing up on it shore, and having not even the faintest idea that the men on these boats carry magic sticks that they can point at you from a distance, and you die.

The chapter then proceeds to answer the question, how exactly would AI kill us even if it intended to, when it doesn’t have a physical body? By paying people.

There’s apparently an LLM that someone connected to the social media platform X, a billionaire gave it $50,000 in bitcoin and as of the writing of the book, the LLM had over $50 million in its portfolio. So it has enough to hire humans. Worse, it has 250K admiring followers who will likely do the AI’s bidding for free.

Also, just as our brains create electrical signals that prompt our appendages to do various things, so too can AI cause things to happen in this hyperconnected world through devices and humans.

One billionaire’s robot company is building a few hundred million or a billion robots and will train AI to steer them around. We may not be quite at the age of Ultron, but we are very close.