I took the same and I find it a really interesting approach. (it just so happened to remind me of the Texas sharpshooter idea)

I took more from what Jason was discussing was that you have a solution and you see how that solution could be applied to a set of problems that you’re already aware of, or have even been cultivating as a set problems that are particularly interesting for some reason to you. Doing this would allow you to see if the new solution was a better fit for the problem than existing solutions.

The last part of the conversation about AI awareness was quite interesting and would be fun over a few beverages. I can see how having a survival bias built into the programming could help it develop better fitness (as in survival of the fittest) but also push the evolution of the AI toward seeing threats to its existence as something to proactively prevent.

As with a lot of things, if you bias towards something you can get more of it than you’d like. Optimistically, you get more of what you measure; viewed pessimistically, what you measure will eventually get gamed

Thanks for podcast and documentaries recommendation!

On a contentious note, I think you guys were in violent agreement on the question of an advanced and powerful AI. Justin may have described it as a survival instinct, and in doing so gave it an anthropomorphic flavour. However Jason’s view that it is axiomatic that an intelligent system would seek to perpeptuate itself I think another way to describe a similar concept. It would mean that an AI would try to ensure it knew what it’s threats are and to take any steps necessary to combat those threats. We could I suppose call that a survival instinct/program.

In any case good show and incidentally Jason’s x-mas sounded like a pretty good family friendly comedy. Maybe time to look up your old script writer pal Jason 🙂

Happy New Year! 🙂