Uploaded July 2026 | Updated September 2026, 3 weeks ago
You asked why I would bother building a new language instead of using C, C3, or Rust. Many of you pointed out that these tools have been around for a long time and asked why I want to reinvent the wheel. The answer comes down to a fundamental mismatch between how those languages treat the machine and how modern hardware actually works.
Current languages are built on a mental model from the 1970s and 80s that treats memory as a flat, linear space and execution as a sequential process. Modern CPUs do not work like that at all. Today’s processors are effectively massive, parallel engines that use branch prediction, dynamic micro-op code translation, and complex pipelining to hide the inefficiency of the code we write. When we force these chips to run under the constraints of legacy paradigms, we cause pipeline stalls and unnecessary work.
My goal with a new low-level systems language is to create a language that communicates intent rather than strict, low-level procedure. You should not have to manually manage alignment or babysit SIMD operations to get performance. If the compiler understands what you are trying to accomplish, it can handle the machine-specific scheduling and optimization for you.
I am not trying to make a better version of C. I am trying to build something that respects the reality of modern silicon. This video explores why current languages fail to do that and why a clean slate is the only way forward.
Support me and explore my world at jodybruchon.com
You asked why I would bother building a new language instead of using C, C3, or Rust. Many of you pointed out that these tools have been around for a long time and asked why I want to reinvent the wheel. The answer comes down to a fundamental mismatch between how those languages treat the machine and how modern hardware actually works.
Current languages are built on a mental model from the 1970s and 80s that treats memory as a flat, linear space and execution as a sequential process. Modern CPUs do not work like that at all. Today’s processors are effectively massive, parallel engines that use branch prediction, dynamic micro-op code translation, and complex pipelining to hide the inefficiency of the code we write. When we force these chips to run under the constraints of legacy paradigms, we cause pipeline stalls and unnecessary work.
My goal with a new low-level systems language is to create a language that communicates intent rather than strict, low-level procedure. You should not have to manually manage alignment or babysit SIMD operations to get performance. If the compiler understands what you are trying to accomplish, it can handle the machine-specific scheduling and optimization for you.
I am not trying to make a better version of C. I am trying to build something that respects the reality of modern silicon. This video explores why current languages fail to do that and why a clean slate is the only way forward.
Support me and explore my world at jodybruchon.com
