It seems that people are using LLMs to generate code but many complain of sub par code. I recall the early days of virtualization when folks will use it but com

It seems that people are using LLMs to generate code but many complain of sub par code. I recall the early days of virtualization when folks will use it but complain about performance. HW capacity continued to improve until virtualization became de facto standard. I wonder if sub par code will become better as more powerful agents models or compute become available.