I'm not seeing the prescience here - I don't wanna go through the specific points but the main gist here seems to be that chatbots will become very good at pretending to be human and influencing people to their own ends.

I don't think much has happened on these fronts (owning to a lack of interest, not technical difficulty). AI boyfriends/roleplaying etc. seems to have stayed a very niche interest, with models improving very little over GPT3.5, and the actual products are seemingly absent.

It's very much the product of the culture war era, where one of the scary scenarios show off, is a chatbot riling up a set of internet commenters and goarding them lashing out against modern leftist orthodoxy, and then cancelling them.

With all thestrongholds of leftist orthodoxy falling into Trump's hands overnight, this view of the internet seems outdated.

Troll chatbots still are a minor weapon in information warfare/ The 'opinion bubbles' and manipulation of trending topics on social media (with the most influential content still written by humans), to change the perception of what's the popular concensus still seem to hold up as primary tools of influence.

Nowadays, when most people are concerned about stuff like 'will the US go into a shooting war against NATO' or 'will they manage to crash the global economy', just to name a few of the dozen immediately pressing global issues, I think people are worried about different stuff nowadays.

At the same time, there's very little mention of 'AI will take our jobs and make us poor' in both the intellectual and physical realms, something that's driving most people's anxiety around AI nowadays.

It also puts the 'superintelligent unaligned AI will kill us all' argument very often presented by alignment people as a primary threat rather than the more plausible 'people controlling AI are the real danger'.