[Alexander] Codex_py2cpp was created as a way to experiment with Codex, an AI intended to translate natural language into code. [Alexander] There were slightly different ideas, however, and created codex_py2cpp As a way to play with the idea of automatically converting Python to C ++. It’s not really intended to create powerful code conversions, but as far as tests go, it’s pretty neat.
The program works by reading a Python script as an input file, setting up a few parameters, then requesting OpenAI’s Codex API for conversion. It then tries to compile the results. If the compilation is successful, then I hope so As a result, executables actually act like input files. If not? Well, fun to learn, too. If you give it a shot, start easy and probably don’t throw too much curbball.
Codex is an interesting concept, and it’s not the first test we’ve seen that plays with the concept of using machine learning in this way. We saw a project that is based on verbal descriptions and creates our own Linux commands [Maya Posch] Look closely GitHub CapitolA high project on commitment and concept, but – at least at that time – is quite low in terms of practicality or utility.