3 ms·
As a programmer who specializes in security, I worry that many of the common errors that people make will get picked up the AI and recommended to new users. It
by freedomben 5y ago
As a programmer who specializes in security, I worry that many of the common errors that people make will get picked up the AI and recommended to new users. It looks like it gets reinforcement from the code that the user selects. In my experience the majority of developers make security errors, so how is the algorithm going to learn not to do it the wrong way when it's learning from bad code and getting reinforced by developers who are wrong? (This is an honest question, not a criticism. I think this product is fascinating)
- asdfman123 5y agoMost programmers are pretty careless and just get something to "work." Maybe the first generation of this sort of completion will be bad, but I have full faith it will be better than the average human at avoiding security issues as early as the next generation.
- bruce343434 5y agofaith based on what? So far in my experience AI projects tend to peter out.