A good developer also has the AI generate Unit Tests for testing the routines. I even write my own routines with input values and see if it matches expected output values. That is why it's also important to tell the AI that it creates functions that can be Unit Tested. So asking the AI to change your code would result in an additional Unit Test run.
What is the best code generation AI code assist that can generate a c++23 vulkan real time graphics rendering pipeline with optimisations from nanite/lumen in under a second?
I asked that from ChatGPT: Currently, no AI code assistant is advanced enough to generate a fully optimized C++23 Vulkan real-time rendering pipeline with techniques like Nanite (virtualized geometry) or Lumen (global illumination) under a second. These systems require complex engineering, optimization, and often custom-tailored solutions.
I would love to see full app making videos where you use different llms inside of Cline
Its one thing to spin up an app fast, but another to test it.
Imagine a calculator with a hidden bug.
A good developer also has the AI generate Unit Tests for testing the routines. I even write my own routines with input values and see if it matches expected output values. That is why it's also important to tell the AI that it creates functions that can be Unit Tested.
So asking the AI to change your code would result in an additional Unit Test run.
what is that terminal?
very quick response but doesn't follow the instructions very well. and across several iterations, the original intent is lost at times.
Wr can just use cline with local llms, i don't know what the difference between groq and cline when it comes to using llama 3.3, do u know?
Groq offers an API so you can just use it inside of Cline, but I find that its high latency after you already have some code there
What's your terminal you're using? It looks sick
Oh my posh terminal, slightly customized
Oh thanks man! 😊 @@kristianlavigne8270
@kristianlavigne8270 Thanks so much 😊
What is the best code generation AI code assist that can generate a c++23 vulkan real time graphics rendering pipeline with optimisations from nanite/lumen in under a second?
I asked that from ChatGPT:
Currently, no AI code assistant is advanced enough to generate a fully optimized C++23 Vulkan real-time rendering pipeline with techniques like Nanite (virtualized geometry) or Lumen (global illumination) under a second. These systems require complex engineering, optimization, and often custom-tailored solutions.
But I think we can change all ui and maybe use gemini api for backend
Cool
maybe i am also first 💀
wow first comment
👏