4 ms·
What you are saying is true for a lot of AI startups/labs, and it's a big reason of why I "left" the field (I'm still in ML but not in computer vision)... But I
by mardifoufs 5y ago
What you are saying is true for a lot of AI startups/labs, and it's a big reason of why I "left" the field (I'm still in ML but not in computer vision)... But I'd argue that GPT-3 is a great example of the opposite. It does not need humans to run, and the algorithm/architecture has been reproduced by different projects with similar results. Adding a filter or hard coding blacklists for offensive speech does not mean you need humans for the AI to run.