3 ms·
Again that level of AI/ML means being able to state what the intent of a comment or document is. If we ever get close to that point, that programs can reliabl
by intended 9y ago
Again that level of AI/ML means being able to state what the intent of a comment or document is. If we ever get close to that point, that programs
can reliably mark something as "risky" content, human thought is in big trouble.
"Risky" material is subjective. Something even humans can't make out reliably very often.
I think, this is the result of the tech firms putting content creators of old out of business, but not investing in the necessary editorial power, subscribing instead to the "wisdom of the crowds" school of thought.
They have not paid for (and will not be able to afford) the distributed human editorial power which was lost.
- Scoundreller 9y agoIt doesn't need to be perfect, just reliable enough to determine what needs more thorough review. A 2 minute video titled "How to fix your computer's _______", that is linked from fixit forums, that is universally liked with comments stating "Wow, that worked, thanks!" shouldn't even need human review to determine risk. The comparison alone to other videos with similar characteristics should do it.
- intended 9y agoHow, all human activities have adversarial components to it- especially when there’s a financial pay off involved. People will just got the comments. And that’s over the naturally noisy data set of human behavior. Not to mention cultural norms across all nations in the world - I know many Indian user forums where even mildly helpful or unhelpful videos/posts will receive “thank you, could you help with...” posts.