> Surely the AI is going to out-perform the human in this task.
Therein lies the problem. We don't have good AI to achieve these types of tasks and are not nearly close enough to achieving it for military applications. The notion that we can just make an autonomous killing machine at this stage is absolutely laughable marketing hype.
If you invest a few billions dollars to the task, you will have your specialized AI with all the sensors needed to out-perform human at this task in less than a decade. There is a reason Google employees refused to work for military.
> If you invest a few billions dollars to the task, you will have your specialized AI with all the sensors needed to out-perform human at this task in less than a decade.
Just how much money do you think the USG has invested in this space? DARPA ain't no slouch. In fact, you're using one of their projects to shit-post.
> There is a reason Google employees refused to work for military.
> I fail to see what this has to do with anything?
If the Pentagon was going to partner with Google to develop drones, maybe it means the talented people required to develop such tech simply don't want to work in the military field, but they exist elsewhere working on others projects. Money can solve all the material problems, but if people still don't want to work toward autonomous jet fighters or killing machines, then it can explains why it doesn't exist at all.
Today's systems, such as planes or engines, are so complex that the only way to achieve today's level of efficiency is to use numerical simulations, aka Computer aided designs. Google excels at organizing information, processing big data and getting answers with predictive models, they do it better than anyone else and it can be potentially applied to just about any field.
I haven't seen any evidence that Google is better at computational fluid dynamics than the large aerospace companies. This is just silly fanboyism with no basis in reality.
Your naive faith in software developers is touching, but this is entirely speculative. Until someone actually does it we don't know how much it will cost or how long it will take. Meanwhile we need to make plans based on technologies that are actually available, or at least have a clear development path.
> Your naive faith in software developers is touching, but this is entirely speculative. Until someone actually does it we don't know how much it will cost or how long it will take.
What about using those 1.4 trillions dollars to support R&D, businesses and help people to live longer and have a good life?
What about investing 1.4 trillions in genome research? in cancer research? in renewable and nuclear energy production?
Sorry, but I prefer to be a naive software developer helping connecting people together around the world than someone who helps random people to get kill.
We are not in the cold war anymore, people don't communicate with papers. We can solve problems by other means than sending missiles to each others, we can see beyond the military propaganda of the countries. We know when something is possible but not preferable : killing drones are one of those project. Technically feasible, ethically not acceptable.
Therein lies the problem. We don't have good AI to achieve these types of tasks and are not nearly close enough to achieving it for military applications. The notion that we can just make an autonomous killing machine at this stage is absolutely laughable marketing hype.