Will AI ever be truly Open?

Day 159 / 366

I read a post today about a company in China announcing its own text-to-video AI model. It is supposed to be a competitor to SORA, and it can generate 16-second videos at a 1080p resolution. Twitter is again filled with the same type of threads with people saying how you “wouldn’t believe what this new model can do”. But I am sitting here wondering, when do we ever see any of this benefiting the common man?

SORA came out way back in February. The demos were amazing, but nothing after that. To this date, it has not been released to the general public. Devon, the AI programmer. Remember how it came out a few months back? Where are all those tools being built by AI? Rabbit R1, and Humane AI pin, all now just feel like gimmicks meant to evoke a reaction and raise money.

GPT-4o came almost a month ago with the demo of the model “seeing” what is happening through the phone camera and talking about it. It is still not here yet.

So my guess is that the same will happen with this new Chinese model as well. There are a few good applications of LLMs. None of it is close to magic or any real intelligence, and that is pretty much it. These new demos with no actual product open for the general public are getting old now.