> This company can hire engineers with an understanding of code (even better, engineering/safety principles behind the code) from the industry
I hate to break it to you, but previous waves of AI did exactly the same thing (e.g. expert systems). Every wave of AI insists that the problem is on the algorithmic side and that they've figured it out this time.
I agree that we should see how this plays out, but don't be surprised if it doesn't work. Much like software development, I suspect there are many strong practical reasons why the real world has a tendency to not align with the specified standards. You can't keep insisting it's mere human incompetence without this whole thing sounding like a grift.
> easy to diagnose 99% of the time
Specifications always have edge cases. Abstractions always leak. That's why everyone is still employed despite all the AI hype. It's wise to expect that 1% is a giant cliff that can totally ruin the success of any project.
I hate to break it to you, but previous waves of AI did exactly the same thing (e.g. expert systems). Every wave of AI insists that the problem is on the algorithmic side and that they've figured it out this time.
I agree that we should see how this plays out, but don't be surprised if it doesn't work. Much like software development, I suspect there are many strong practical reasons why the real world has a tendency to not align with the specified standards. You can't keep insisting it's mere human incompetence without this whole thing sounding like a grift.
> easy to diagnose 99% of the time
Specifications always have edge cases. Abstractions always leak. That's why everyone is still employed despite all the AI hype. It's wise to expect that 1% is a giant cliff that can totally ruin the success of any project.