@onsongo_k@kloss_xyz It feels like being someone, it actually give an message here, it was ridicule about token guessing in whole video and in the end it says that something happens in training
@RoyInProgress@claudeai You are right, doesn’t make much diffirence for normal people, but it makes huge diffirence in software development (better codes, faster development, especially for LLMs itself), health,science. We still need better models for complex works and problems
@suni_code AIs just read raw files. They have zero clue what calls what, how classes connect, or why your architecture looks the way it does. Grafyx fixes that.
It parses your entire codebase into a live relationship graph and feeds it directly to your AI via MCP.
https://t.co/6YPAYT7MlQ
@Tech_girlll It is about skill (not llm skill.md) issue. I created a project with +125k lines of codes and with +600 files. I also using Grafyx which is I created this for this exact issue with very large codebases. https://t.co/XDyBieb5DU