😅 sounds interesting, but i guess it also shows the need of implementing reliable safety checks in the process before publishing

I think both yes and no. Yes, all the helper libraries for single purpose abstraction are going to be replaced as the AI is writing the code itself. No, libraries providing structure and reliability will increase in value. I think smart standards will be much more relevant. E.g. PocketFlow

Probably enjoying the general concepts of automated prompt optimization based on dynamic result scoring to much but looking forward to it 😄

It is a really interesting tool/video, especially considering the current paper from DeepSeek regarding Critique Tuning. Just not understanding why you are not using textgrad for the prompt optimization?

[2] Considering that we create the prompt / character of the AI in minutes or days it's never as well rounded as the real human, but this is more a execution mistake and less a systematic issue I would say

[1] It is slightly philosophical, but isn't it exactly the same as a human building the character through experience in our childhood/ general life?

Please share if you find a solution. Youtube is getting only focused on viral and less on quality, and it would be great to adjust the algos

Let's just enjoy the magic time where we decided as a society to ignore reality and vibe the moment ✨️

Just hit -20 % YtD. It's going to be an interesting year