>>12056674If you warp space in such a fashion to reduce the distance you have to travel, that replicates the results of going faster than light. Which is all any reasonable person cares about.
If you're so confident about AI being harmless and unthinking, why don't you go tell all the Oxford Professors, nobel prize winners and Elon Musk to stop worrying about it? What stops it from thinking, from using resources to achieve its reward function? These programs do all kinds of weird shit.
https://aiweirdness.com/post/172894792687/when-algorithms-surprise-usEven if it can't 'think' it might still kill us all, developing a superweapon based on some weird interpretation of a command. Paid for by market manipulation, printed out in a lab, delivered by hired van, instructed over wifi: it can do a lot of damage without a body at all. Nanotech can be very powerful, given superhuman intelligence.
AI dungeon, especially the GPT2-based parts can be kludgy. Bigger unreleased models developed in the future will be more dangerous.
