I tested several local models for tool calling performance. Until now, it’s the qwen2.5:3b leading the “board”. To do this, I built a small tool selection(e.g. get_weather, etc.), quite a lot of typing work. But then, I found the agent-friend toolkit in this blogpost(https://dev.to/0coceo/ollama-tool-calling-in-5-lines-of-python-3h5f) and it’s just amazing! I can now get straight into the tool development without hundreds of lines in JSON.
#1 (real) Devlog: I started a graphical interface, the core structure works and linux works well with the external program for the sudo entry. 1 Hour was enough to understand first the surprisingly complex structure of tkinter, but it looks… well, it’s like Windows XP (but worse). Apropos, if you ask why this will be linux-only, there’s one answer: Windows is TERRIBLE and MacOS is restricted. I can’t wait the final result.
I changed the plan. Now, I want to build a better password entry for sudo… with chess. Every cell on the chess board stands for two ascii characters, depending on the color of the piece on it, it returns a ascii string. So your sudo password is transformed into a chess position and every time when you legitimate as sudo, you enter the position that gives the password to stdout. I think about the gui, should I use tkinter or an alternative (kivy, etc.) ?
Just started the key generation out of uci chess-moves… Not finished yet! I also want to make a hash out of it, security question.
Just starting…
I finally started my chess-encrypt project. The goal is to use the infinite number of chess combinations to encrypt data(files, passwords, etc…). You just play a chess game against the computer(stockfish) and this will generate a unique key… I know, it’s just some nerd stuff, but why not?(The Project is entirely written in Python :-) )
