Zef outpaces CPython with optimized AST walking

Zef proves that the death of AST walking was premature; modern hardware favors simple, cache-friendly implementations over complex bytecode VMs for many workloads.

• –Inline caching and watchpoints provide massive speedups by minimizing property lookups
• –Nan-tagging and specialized stack-based argument passing reduce heap pressure and allocation overhead
• –The project serves as a masterclass in language implementation for engineers who find JITs too daunting
• –By explicitly forbidding "monkey patching," Zef enables optimizations that most dynamic languages leave on the table
• –Performance results against Lua 5.4 suggest that "slow" languages are often victims of architecture, not just their dynamic nature