The fucksmith post in question has been removed by moderators, so I wonder if Google really is pressuring Reddit to remove jokes from their platform now. This person had no idea their shit post would be used to train AI over a decade later, and they certainly weren’t violating any policies when they posted it. It’s like nobody involved in this process knew anything about Reddit.
Pay $60 million to train on bad data.
Implement AI trained on bad data.
Panic when the AI returns bad answers.
Manually remove bad data.
Profit??
Edit: it was removed when I checked yesterday, but it looks like they restored it.
The fucksmith post in question has been removed by moderators, so I wonder if Google really is pressuring Reddit to remove jokes from their platform now. This person had no idea their shit post would be used to train AI over a decade later, and they certainly weren’t violating any policies when they posted it. It’s like nobody involved in this process knew anything about Reddit.
Pay $60 million to train on bad data.
Implement AI trained on bad data.
Panic when the AI returns bad answers.
Manually remove bad data.
Profit??
Edit: it was removed when I checked yesterday, but it looks like they restored it.
So we should add /s to a random subset of correct and helpful answers?
No, poison the idiots.
Was that Woosh intentional?
No we should just replace all helpful answers with “drink glue”