Which is why it's only a conjecture. You need static analysis, SAT solving, and run-time checks to validate that conjecture.
Using something like GPT-4 on this problem is promising. It's probably going to be right most of the time, and its errors can be caught by the next phase of the analysis. That's about what you'd get if you put junior programmers on language conversion.
Comments
Which is why it's only a conjecture. You need static analysis, SAT solving, and run-time checks to validate that conjecture.
Using something like GPT-4 on this problem is promising. It's probably going to be right most of the time, and its errors can be caught by the next phase of the analysis. That's about what you'd get if you put junior programmers on language conversion.