@MLGaming Why did this resolve N/A? Isn't it clearly a NO?

@Mira Why are you on the NO side in this question, but YES in GPT-4 market? Do you believe that multimodal model won't we able to read a grid of numbers from an image? If it will, then it's basically the same problem onward

@MrLuke255 probably because nobody will try (but I don't know if that was as obvious 2months ago)

@Gen I tested Bing and GPT-4V's image-reading capabilities months ago and realized they would need a lot of work to get working consistently.

I notice that nobody has put in that work, despite reading the grid being independent from the main problem. It's not going to solve itself, and the market creator is unlikely to want to pay API costs judging tons of examples to get a precise percentage anyways. So they will rely on ChatGPT, which will be harder to adapt Emily's solution.

@Mira It's just solving the same problem with extra steps. You will need to convert it to a text representation anyways.

@Shump Are you trying to explain this market to me as if I don't understand what needs to be done?

@Mira No I am just replying to the thread and Manifold automatically tags you. Chill.