After having a wave of crashing of my Cursor IDE instance, I decided to give Windsurf (https://codeium.com/windsurf) a go today, and boy, it is a significant improvement. The premium Claude calls doesn't run out, look appropriate at all your libraries, do not need to be saved after you accept updates, and can look at how the app runs, correcting errors automatically. It seems to avoid the silly loops of confusion and better understand your code and the problem it is trying to fix. Let's see the progress.

Hey Ollama users. I have a 4200 word, Word document. It has some tables. I want to try some embedding with it. What Opensource models do you use for embeddings?

Hey all! Beyond stoked to connect with Brandon and the rest of the tribe🌐

What is the cheapest LLM you have found to do research (not creative) tasks? This agent would need to use tools like web searching and web scraping and be able to follow directions well. I have tried gpt-4o-mini and it does not seem to follow directions very well and does not seem to want to scrape websites although it will do an ok job of searching the web with serper. Gpt-4o does a better job, but is more expensive. What is your experience?

I have a question regarding handling data that exceeds the context window limit. I once worked on an application to generate unit tests. As you know, writing a good unit test for a function requires access to the definitions of all dependencies used by the function. While I, as a human, can access the entire codebase, this is not possible with an LLM. If I don’t provide the definitions of all the dependencies, the LLM will make assumptions, which can result in incorrect mocks for other classes or external tools used by the function I’m trying to test. Do you have any suggestions for addressing this issue? Summarization or advanced chunking solutions won't work in this case because the LLM must have full access to all dependencies to generate accurate unit tests.