Randomness 3

[MT2008]

Originally Posted by Mazryonh

I was taking issue with what you said about machine learning and sensor capabilities. And you needn't be so quick to judge. We all have issues we like to talk about.

OK, but as I said, I didn't fundamentally disagree with the basic premise that it is undesirable for a machine to determine whether or not to pull the trigger (and really, it's moot given policy - which I cited). I just pointed out that 2013 was an eternity ago for those fields; even though classifier error (which you refer to) is still a problem - especially for image recognition algorithms - current multi-sensor fusion techniques are mitigating some of those shortcomings. (But...before you write another 10-page rebuttal, I will add a disclaimer: Not completely, which is why most algorithms cannot recognize objects and patterns the way that human operators can.)

I just do hope you understand that IRL, if you grandstand people the way that you just did with me, you will alienate them. Might want to temper it a bit; sometimes, showing off how much you know can backfire in certain social situations.

Originally Posted by Mazryonh

Anyway, would we be even having this conversation about drones in combat if there wasn't a manpower shortage in the US Armed Forces? Do you have a link to the specific article you were referring to?

https://www.realcleardefense.com/art...ng_115283.html

(Now that I see the date in my browser, I was wrong when I said the article just came out - it came out last month, but it only just showed up in my Facebook feed on Friday.)

One other point: I don't disagree with your article's premise that the cost savings of UAVs compared to manned aircraft have often proven dubious (cost overruns are probably one of the reasons that the U.S. Navy is pausing MQ-4C Triton production). However, IMHO, that says as much about the need for acquisition reform as it does inherent shortcomings of UAVs.