This comes in frustration after dealing with a flood of spammy crap this morning. But I do think someone who launches something like Google twenty years ago with great web search, spam filtering, shopping search, and not too many ads easily would get a flood of customers. I'd be the first in line.

That certainly doesn't mean the companies aren't responsible for what they're doing. But it does mean that companies can start down this path just by picking bad metrics for what they consider the best chatbot might be. Naively picking time-using-the-product as your success metric will cause this.

One curious piece of this is that the companies easily can develop bad AI without intending to. For example, if you are testing chatbot personalities and your success metric is how long people keep chatting, you'll tend to deploy engrossing chatbots that can have negative impact on mental health.

That certainly doesn't mean the companies aren't responsible for what they're doing. But it does mean that companies can start down this path just by picking bad metrics for what they consider the best chatbot might be. Naively picking time-using-the-product as your success metric will cause this.

One curious piece of this is that the companies easily can develop bad AI without intending to. For example, if you are testing chatbot personalities and your success metric is how long people keep chatting, you'll tend to deploy engrossing chatbots that can have negative impact on mental health.

the culture war/political divide is obviously real. but stuff like this shows it's also heavily amplified by essentially synthetic actors. researchers know this but the extent is notable. the platforms like this engagement. this is a desired outcome for them. fraying the social fabric for ad revenue