Gemma 4 for Coding Help

A developer used the local model gemma-4-26B_q4_0-it.gguf in llama-cpp to fix long-standing HTML and JavaScript issues, adding lazily loaded images to a blog and a defer-loaded JavaScript file to a search page to cut bandwidth use and speed up page display. The developer said the local Gemma 4 model made the task somewhat easier than using search engines, while noting they have not yet figured out agentic coding.

Today I wanted to fix some long-standing issues on HTML and JavaScript files I have. I decided to just use gemma-4-26B q4 0-it.gguf in llama-cpp and prompt it about the required HTML syntax. I added lazily loaded images to my blog, and I added a defer JavaScript to the search page. This should reduce bandwidth usage if someone visits the blog and doesn't scroll down, about half of the images will never load . And for the JavaScript, it will speed up how fast the search page displays as the search data is on a separate file, not directly in the HTML. I haven't quite figured out agentic coding yet; perhaps it is not useful for my exact projects or something. In any case using a local model like Gemma 4 instead of search engines made the task somewhat easier.