kinda weird how people will just attach chatgpt to customer service things and let it fuck around practically unmonitored while they watch their human employees like a hawk and fire them at the first sign of an oopsie

my wife has the perfect addendum here:

it's because they fundamentally see employees as a frustrating expense, and anything an employee does that's not making money for you in the most efficient and perfect way possible is tantamount to stealing

whereas business owners very much still have a consumer-brain just like everyone else in society, so they're not nearly as critical of their shiny new toy that they just got

@saddestrobots As long as the autopayments go through they don't have to care.
@saddestrobots Well, there's zero chance that ChatGPT is enjoying itself on paid time, unlike dirty meat employees who might sneak a minute of joy when it's slow.
@saddestrobots Almost like it was never about providing a good product/service and instead was all about power trips. Almost
@teleute @saddestrobots Nah. They wouldn't. Would they?
@saddestrobots anthropocentrism, same reason we let automobiles kill tens of thousands every year without batting an eye and then reshape an entire industry when two or three people cause a minor incident. Possibly the single biggest flaw/weakness of our species. My therapist used to call it "misattribution"
@saddestrobots To me it seems like a great way to have a bot start promising shit to customers that they absolutely should not. That's why I wouldn't trust it from even a capitalist perspective.

@saddestrobots
maybe this is planned obsolescence for companies? 😄

helps with customer dissatisfaction to "wake them up" and crave alternatives which alternative people can work on and offer where formerly people would have no need for alternatives. ...so maybe this is collaboration from people tired of running capitalism, so they see AI as the best opportunity to end this mess and having a perfect alibi thst they just tried to make things better 😁 ...j/k ...yeah, it's fucked up

@saddestrobots Clearly we should start sending emails to our reports asking them if they could do something ridiculous like to tell us something that they should absolutely know is not an allowed thing, or even a thing at all.

If they respond with affirmative responses and not "What are you talking about? That's not how that works/that's illegal/that's a horrible idea.", they fail, but if they give those responses, they're clearly paying enough attention to their work.