5 Comments

Well written post… with regards to bad actors I think of it as; an aligned model isn’t going to make evil people less evil and an unaligned model isn’t going to make evil people more evil. We definitely want to prevent making it easy to create things like child porn and such but there are plenty of bolt on approach’s you can take to monitor for those types of outputs.

Expand full comment

We will have a world similar to the past, but with more examples of extreme, dangerous stupidity fueled by the egos of the filthy rich who believe only their thinking is correct. In the end, if something goes wrong, it will go really wrong, even if it makes no empirical sense to the humanity. My brain operates in its own skull, needing a biological body to fuel and feed it. A rich oligarch's son, with access to something like X-Men Cerebro with AI generators, able to read minds or create any kind of software with the press of a button—just saying—is going to be our future, sooner or later.

Expand full comment

Monster zit, eh? :)

Expand full comment

What about maybe the ability to set "Parental Controls"? Let the customer decide how to censore the AI for themselves.

Expand full comment

That's what I'd prefer

Expand full comment