How do humans with very little exposure to grotesque violence or extreme content universally label such content so well? This is not graduate level data that needs labeling.
What is missing in an AI model for it to intuitively understand what content is extreme from very few labeled sample in training?
I'm running virtualhere on thousands of raspberry pi's sharing various USB devices to cloud machines over vpn. It's been working without issues for years now. Seems to be a solo developer in Australia that's been working on it for a really long time. https://www.virtualhere.com/
Can this be used to click around in the browser with text prompts? Maybe after some fine-tuning on screen recordings of specific workflows in browsers.
Many chatGPT prompts with mathematical inconsistencies in the output simply clear up with the some variation of "Explain your thinking step-by-step". Anyone know how or why this works?