I did lots of AI testing yesterday so today’s posts will be discussing what I have learned.
I tried out Sora which is a way to generate video from text prompts. I must admit that when I tried it out with some various prompts I wasn’t impressed. It is unusable in my opinion. Things were spelled wrong in the video, and it looked weird and unnatural.
However later that day I tried again and this time uploaded a picture. When the first result came in I said out loudly in surprise “Oh my word.” It was remarkable. It took that half of a bear and animated it in a realistic way and then filled in the missing details and then allowed me to create complex scenes and settings very easily.
Now I have the paid ChatGPT account for $25 and that only gives me a very small amount of videos I can do for a period of time. It was also low resolution and short. Which makes sense given the amount of investment video requires. If you have a need to create videos then by buying the $200 month ChatGPT Pro you can create HD video and have way more options and opportunities for different versions.
In a period of 20 minutes I made scenes that looked like a professional company has shot them for real. It was remarkable. However it only looked good when I provided a photo and I would not trust it without one.
In a few months or years I am sure that it will be capable of making things without a photo look good. Would I use Sora for my own personal use? Probably not. I want to write something that stands the test of time, and video is gimmicky. People want substance not fluff all the time.