A new LLM-powered scam: research an author, research a publication they write for, research the title of something they've written, register a similar gmail address, go phishing.
This is still easy to see through. It won't be for long.
Researching someone and putting together a personalized attack used to be possible but labor-intensive. It can now easily be scaled up to millions of targets.
The emails will look genuinely indistinguishable from those of your health insurance this time, no more silly typos. They will ask for something right when you expect to be asked.
It's all going to get a lot more sophisticated.
