What's your stance on a future open source model that is as capable as any commercial models? Also, I'm curious, do you consider LLMs to be incredibly error pr

What's your stance on a future open source model that is as capable as any commercial models? Also, I'm curious, do you consider LLMs to be incredibly error prone and untrustworthy? Or do you think they are going to replace software developers?