|
|
|
|
|
by transcriptase
1205 days ago
|
|
Then how to do you explain companies like OpenAI, Google, Meta, Microsoft etc sinking so much developer effort into ensuring that out of the box, technologies like GPT won’t say something like BasedGPT did? If no reasonable individual would take offence, then what exactly is the concern? |
|
It's not that any of the developers or big tech want to kill a person rather than uttering a racial slur. I'm amazed this needs saying
Edit: further down the thread, https://news.ycombinator.com/item?id=34975072 also shows (and I'm a Linux fanboy) that it'll put out unnuanced stuff in such modes. I can see why it's all toned down and made safe for work. It's just not as useful to us otherwise (there's a moral lesson in that, probably)