3 ms·
Pushing objectively false answers should render an AI doodad bad for the job, no?
by inferiorhuman 1y ago
Pushing objectively false answers should render an AI doodad bad for the job, no?
- xemane 1y agoYes. But that's why I said "if it's good for the job it's good for the job" If there's something that Grok *positively* does better than other LLMs, why wouldn't you want to use it, because, _boohoo_ Musk bad.
- AlecSchueler 1y agoBecause you won't know when it's going to peppering in Holocaust denials in the emails it composes for you.
- xemane 1y agoIf you found out that there's a task Grok excels at far better than GPT or Gemini, you're telling me you wouldn't use it?
- Larrikin 1y agoYes
- xemane 1y agoOkay. Your code, your choice.
- AlecSchueler 1y agoYes, in the same way I wouldn't hire an overt fascist. I could never trust it as a tool.
- frollogaston 1y agoEven if I didn't care about who I give business to and I just wanted the best product, it's not this simple. These AIs are incredibly complicated and unpredictable. If you don't trust the business, you can't trust its product to consistently do what you want.
- bastardoperator 1y agoIt's not good for the job because better tools exist. Yeah, I'm not keen on giving money to billionaires for subpar products, that's why I don't drive a Tesla.