I'm With Stupid wrote:

Teague wrote:Why do we celebrate these long range shots lol

The military needs heroes for its PR. That's what gets the next generation of kids interested. In an era of drone strikes and questionable conflicts, they've probably been struggling for them.

Not going to be hard to get kids to fly drones (they will have been doing it for years before they even get the chance to join the army). It does however look that as the military gets more and more technical there will be even less roles for working class poorly educated people

It's been my experience in speaking to bullet catchers that they take great comfort in the ability to hold up their hand for a sniper support check and see that laser spot appear on the back of their hand. That kind of support is simply not possible with machines.

Weaver wrote:

tuco wrote:Yes you provided one variable but I do not see how, in principle, its impossible for AI to determine it better than human. But but .. write the code! Its not on the market yet! I said in principle.

You remind me of a friend of mine, chess player, who insisted that because there is no solution for chess yet it will remain human domain. I guess it made him uncomfortable that what he believed to be his "strategy", his special sauce, was merely a crutch for his limited capabilities.

Ah, yes - "in principle" again, the ultimate get out of jail card.

And again, the reason that it is a problem is that we don't live in an theoretical world, and there are too many variables to quantify.

You are living in the land of "Wouldn't it be nice if ..." and trying to insist everyone else live there too.

I was not in any jail so no card needed.

Initially I made suggestion or rather asked question (post #18) and you replied this:

Weaver wrote:

tuco wrote:Would not AI be better at this stuff? I mean what is being called skill is accounting for variables.

Trained human instincts are far better at some sorts of estimation - particularly when having to interact with instrumentation and bare-eyeball observation to obtain data.

which is obviously a pile of possibly biased horseshit and/or not thought through assessment. Human instincts are shit.

Then I continued to explore the horseshit further simply because it did not compute to me and you said:

tuco wrote:

Weaver wrote:

tuco wrote:Would not AI be better at this stuff? I mean what is being called skill is accounting for variables.

Trained human instincts are far better at some sorts of estimation - particularly when having to interact with instrumentation and bare-eyeball observation to obtain data.

Alright, how about rifle designed especially for robot to use? I just do not see how human estimation could even be competitive to AI, just like in case of driving a car for example.

Weaver wrote:

tuco wrote:

Weaver wrote:

tuco wrote:Would not AI be better at this stuff? I mean what is being called skill is accounting for variables.

Trained human instincts are far better at some sorts of estimation - particularly when having to interact with instrumentation and bare-eyeball observation to obtain data.

Alright, how about rifle designed especially for robot to use? I just do not see how human estimation could even be competitive to AI, just like in case of driving a car for example.

A car has a defined road to follow. A sniper has a much, much wider field of engagement - including the vital factor of target discrimination - determining whether it's legal to engage or not.

Judging wind includes a shitload of variables - it isn't like there are direct measurements at any point on the trajectory except the shooting position. They have to be estimated at multiple points along the trajectory, and account for the effects on the wind by nearby trees or buildings, flight over water - literally countless factors.

It simply isn't a hard enough science to automate - there are scientifically quantified elements to the firing solution, but it remains an art. If it were all down to quantifiable science, there would be automated systems winning the long range shooting matches - or at least being put up against the best shooters and beating them. But there aren't any.

Of course its down to quantifiable science because how else? That the tech is not available yet does not mean its not down to quantifiable science. What is this art you speaking about? Special sauce. So when informed that the tech is not available yet I said:

tuco wrote:

Weaver wrote:

tuco wrote:

Weaver wrote:
Trained human instincts are far better at some sorts of estimation - particularly when having to interact with instrumentation and bare-eyeball observation to obtain data.

Alright, how about rifle designed especially for robot to use? I just do not see how human estimation could even be competitive to AI, just like in case of driving a car for example.

A car has a defined road to follow. A sniper has a much, much wider field of engagement - including the vital factor of target discrimination - determining whether it's legal to engage or not.

Judging wind includes a shitload of variables - it isn't like there are direct measurements at any point on the trajectory except the shooting position. They have to be estimated at multiple points along the trajectory, and account for the effects on the wind by nearby trees or buildings, flight over water - literally countless factors.

It simply isn't a hard enough science to automate - there are scientifically quantified elements to the firing solution, but it remains an art. If it were all down to quantifiable science, there would be automated systems winning the long range shooting matches - or at least being put up against the best shooters and beating them. But there aren't any.

So what is the special sauce human has AI does not? Because I do not see it. Art .. well, whatever but its beyond me how could anyone think that human can target better than AI, in principle.

In principle, not to get out of any jail but simply to point out that there is no special sauce and its quantifiable science.

AI snipers will be more accurate, will not put human snipers in danger and will be more cost efficient, just like many other robots taking over human tasks simply because humans are just hairless monkeys with limitations.

Robo-Snipers, “Auto Kill Zones” to Protect Israeli Borders - https://www.wired.com/2007/06/for_years_and_y/

and btw you are trolling.

AI "snipers", shooting from what position? What do you think will happen when the enemy learns of that position? Before answering, you should remember that technology exists right now to locate the origin of a fired bullet.

Those Israeli weapons systems are not AI robots like you described being theoretically possible - they are simply automated weapons with humans controlling them remotely. Given that the weapon in question is a Mark 19 grenade machine gun, it is hardly supportive of your claims that AI robots can perform sniper duties better than humans - it is an area effect weapon designed specifically to make up for targeting inaccuracies.

It is conceivable, perhaps even probable, that someday AI driven computer systems will be able to match human snipers in the field. But they are no where near close to that now. There are way too many variables involved, way too much non-quantifiable information to program, way too much flexibility needed.

You asked me to provide examples and said you could solve them. Your lack of even an attempt demonstrates your abject ignorance in the subject area. You have no idea about the difficulties involved, but you think that by spouting "in principle" enough, and by demeaning valid points as "special sauce" you can simply dismiss all objections as unimportant.

It is the details which matter. Those details are insurmountable for the current state of the art.

@Weaver I can read, thanks. It was to demonstrate that what I suggested is viable. Glad we cleared it up.

@The_Metatron What do you mean? They could be flying drones because AI unlike human could account for own velocity etc too. These details I am not concerned with.

tuco wrote:@Weaver I can read, thanks. It was to demonstrate that what I suggested is viable. Glad we cleared it up.

You say you can read, implying you understood the reasons why the article you cited does NOT support your position - then you come right back and say it demonstrates what you said is viable. No, it isn't fucking viable. There is no AI in the loop - there is no computer making the decisions to engage, computing the necessary data, or any of the other intricacies of a long range sniper shot. It is just a weapon pointed by servos with cameras boresighted to the sights, and completely controlled by humans. So NOTHING FUCKING LIKE what you claim is viable at all.

My position is that in principle nothing prevents AI being more accurate sniper than human. Which is obvious yes so that is why its beyond me why the opposition.

tuco wrote:My position is that in principle nothing prevents AI being more accurate sniper than human. Which is obvious yes so that is why its beyond me why the opposition.

I explained why humans are better in many activities like this. That doesn't mean AI won't overtake them, but it sure hasn't happened yet.

You explained nothing. You posted cool story about unconsciousness and how special it is, sauce. I am not idiot, rather the opposite :), I understand that. Humans walk effortlessly, unlike robots. So what? Sniping is "mechanical" activity, scientifically quantifiable, putting variables into formula. Period.



If there was an argument that AI will hardly write Shakespeare, well, I would get involved much.

tuco wrote:My position is that in principle nothing prevents AI being more accurate sniper than human. Which is obvious yes so that is why its beyond me why the opposition.

And you ignore the counter arguments that it doesn't matter what is possible in principle, what matters is what is capable with current technology.

In principle, we could just go out and grab comets from the Oort cloud and redirect them to crash on enemy tank formations - that doesn't mean that we can actually do it.

Hiding behind "in principle" is just waving a magic wand to dismiss all the details of what makes current reality different than the hypothetical world you apparently want to exist in.

tuco wrote:You explained nothing. You posted cool story about unconsciousness and how special it is, sauce. I am not idiot, rather the opposite , I understand that. Humans walk effortlessly, unlike robots. So what? Sniping is "mechanical" activity, scientifically quantifiable, putting variables into formula. Period.

Prove it. Quantify the variables in the wind videos I posted - provide the formulas that will yield wind speed and direction at the specified distances from the camera.

How might the target have appeared in the scope at 2.2 miles away? Anything as clearcut as in the following image?

tuco wrote:You explained nothing. You posted cool story about unconsciousness and how special it is, sauce. I am not idiot, rather the opposite , I understand that. Humans walk effortlessly, unlike robots. So what? Sniping is "mechanical" activity, scientifically quantifiable, putting variables into formula. Period.



If there was an argument that AI will hardly write Shakespeare, well, I would get involved much.

What the fuck are you talking about? I said humans have incredible pattern recognition abilities, often people don't understand what they see, but they see it. This is well known. You don't understand this? Do you have any idea how hard it is to program in the ability to see patterns in video data? They're still barely beginning to do this, and they're using neural networks, which will do things that the programmers have no idea how the machines came up with their solutions/results. This again is well known.

Weaver wrote:

tuco wrote:You explained nothing. You posted cool story about unconsciousness and how special it is, sauce. I am not idiot, rather the opposite , I understand that. Humans walk effortlessly, unlike robots. So what? Sniping is "mechanical" activity, scientifically quantifiable, putting variables into formula. Period.

Prove it. Quantify the variables in the wind videos I posted - provide the formulas that will yield wind speed and direction at the specified distances from the camera.

Yeah! That's what i said, but afterwards, I lead from behind.

The shot used the 50 cal BMG, which is a massive round. That means it will be bothered less by the wind. The velocity helps also.

Question, it's only tangential to this, but this seems like a decent place to ask. I saw cockpit video of an Apache taking out a person, don't know the distance, with it's chain gun/gatling gun, forgot which it is or the distinction, but gatlings are multibarreled I think. And it uses depleted uranium rounds if I'm not mistaken. First, I would think they would need far less windage corrections to aim accurately, not that that is what they do. The real question is, the shootee just disappeared. It was a 'wow" moment, but I'm not really sure how much of that was an artifact of the video. What would happen if you got struck by 3-4 of these rounds, which I think are 25 mm, and probably weigh a pound. At likely supersonic??? velocity.

I will repeat, I understand that crank. Auto exposure metering, for example, uses algorithm and can also be fed data for patter recognition. I have already said that AI can be fed much more data and much faster than any human. Period.

Weaver wrote:

tuco wrote:You explained nothing. You posted cool story about unconsciousness and how special it is, sauce. I am not idiot, rather the opposite :), I understand that. Humans walk effortlessly, unlike robots. So what? Sniping is "mechanical" activity, scientifically quantifiable, putting variables into formula. Period.

Prove it. Quantify the variables in the wind videos I posted - provide the formulas that will yield wind speed and direction at the specified distances from the camera.

Honestly I am losing my patience but that is my cool story. Its like asking to provide formulas for water drop distribution when an excrement hits water in toilet. In principle, such distribution can be accounted for precisely. Beyond me why and how would someone dispute this.

chango369 wrote:How might the target have appeared in the scope at 2.2 miles away? Anything as clearcut as in the following image?

A human facing head-on toward the observer is assumed to have a shoulder-to-shoulder width of 18 inches.

Plug that into the mil formula, width in inches * 27.77/width in mils = Range in Yards.

Assuming that the target is being ranged at the proper magnification for the optic, that means the target is 249.93 yards away (228.54 meters). Since most people zero their weapons on a meter range rather than a yards range, they'd likely use their data for 225 meters - the difference would be negligible.

So - that's what a target appears like at 225 meters - through a 10x optic.

It would appear about the same if it were 450m away using a 20x optic.

The max magnification the Canadian sniper could have used, from reports I have seen of the two optics available to him, would be either 16x or 25x. Depending on a couple of factors, he may or may not have selected max magnification when making the shot - his spotter was most likely using much more magnification for target ID, ranging and spotting purposes.

So, long answer short - no, the target wouldn't have appeared anywhere near close to that big in his reticle. It would have been far, far smaller.

crank wrote:The shot used the 50 cal BMG, which is a massive round. That means it will be bothered less by the wind. The velocity helps also.

Question, it's only tangential to this, but this seems like a decent place to ask. I saw cockpit video of an Apache taking out a person, don't know the distance, with it's chain gun/gatling gun, forgot which it is or the distinction, but gatlings are multibarreled I think. And it uses depleted uranium rounds if I'm not mistaken. First, I would think they would need far less windage corrections to aim accurately, not that that is what they do. The real question is, the shootee just disappeared. It was a 'wow" moment, but I'm not really sure how much of that was an artifact of the video. What would happen if you got struck by 3-4 of these rounds, which I think are 25 mm, and probably weigh a pound. At likely supersonic??? velocity.

Apaches do not use gatlings - the Bushmaster cannon on an Apache is a chain gun.

And it doesn't use depleted uranium.

The primary ammunition types are variants of high explosive armor penetrating ammunition - and yeah, when you shoot a person with them they pretty much go away.

And yes, they are most certainly supersonic at the muzzle - well past that line, like most non-pistol ammunition.