@bitprophet I've been there and I'll be there again :blobcat_salute:

@paulrickards I remember a marketing agency I worked at back in the 90s where the office had a pool table on one floor and a few engineers on another floor with Kensington Expert Mouse trackballs. There was A Problem with engineers raiding the pool table for trackballs periodically

So, I'm as disturbed as anyone about folks having "relationships" with LLM output. But, this Reuters Special Report just reads like clickbait to me.

Like "Meta’s flirty AI chatbot invited a retiree to New York. He never made it home."

The AI didn't kill him. He fell and injured himself grievously while running to catch a train.

Sure, he was catching the train for a rendezvous with a hallucination. But, it's not like the LLM uniquely generated a trip hazard?

https://www.reuters.com/investigates/special-report/meta-ai-chatbot-death/

Not to diminish the tragedy of the guy dying, but this strikes me like the panic over various things like Dungeons & Dragons and Mazes & Monsters (1982) - "lil timmy went into the steam tunnels to play satanic wizard and he NEVER CAME BACK"

Like, this kind of thing just introduces noise and makes it easier to dismiss more nuanced criticism

@lerxst @rasterweb That totally would have been a plot point in Daniel Suarez's Daemon and Freedom™ books, though

@aparrish I mean... okay that connection I can buy

@epilanthanomai Look, when I get the Trello card to hook up the 1287 programmatic integrations for business partners with "legitimate interest", they'll flay me alive if I say no. That's the moneymaker.

(but seriously i haven't had to work on a site like that in many many years and feel privileged)