I have a friend who is curious about AI and I want to introduce them to the problem of AI safety. What is the best short introduction to the problem? Something roughly the length of a long blog post (shorter than a book). Maybe something that gives a good hook but also gives an overview for why it is technically very difficult?
Any medium is fine (blog posts, articles, youtube videos, podcasts, etc).
1000M to the best suggestion.
Related questions
I really appreciate Holden Karnofsky's (CEO of Open Philanthropy) "Most Important Century" series.
It has a high level, couple page summary of why he thinks AGI is plausible in the near future, why it would be world changing, and why we are not ready for it.
Then under each claim in the summary, he links to a blog post of his expanding on the claim in detail.
It seems you're looking for something in text format, but you haven't explicitly ruled out video. Since I suspect that one of your concerns is that you want the introduction to appeal to someone who might not already be interested in such topics, I think a video might be better than a short article. To that end, I would recommend the following video from Rob Miles:
https://www.youtube.com/watch?v=pYXy-A4siMw
Thank you so much for all the great resources everyone! https://www.lesswrong.com/posts/bJgEMfiD48fEJJxjm/ai-risk-intro-1-advanced-ai-might-be-very-bad From Isaac is my current favorite so I awarded the M1000 there. But I really appreciate the other contributions so I gave you a little extra. Thanks again everyone!