Once again, Russia has announced its latest Wunderwaffe is coming to Ukraine, which ZeroHedge treats seriously, because ZeroHedge.
Western media outlets flooded the airways with hope for Ukraine this week as the US prepares to send 31 main battle tanks to the wartorn country in Eastern Europe to counter Russian aggression ahead of spring. What wasn’t highly publicized is that these M1 Abrams are a modified version and will be stripped of “secret” uranium armor.
Following the news of NATO-made tanks set to flood Ukraine, the former head of Russia’s space agency Dmitry Rogozin told the Russian newspaper Pravda that “Marker,” a new robo-tank, will be able to ‘destroy Western tanks, including American Abrams and German Leopards.’
Rogozin explained the robot tank automatically recognizes and attacks Ukrainian equipment, including NATO tanks, all because of its artificial intelligence system and machine learning technology.
“The combat version of the Marker robot has an electronic catalog in the control system that contains images of targets both in the visible and in the infrared range,” he said.
The director of the Air Defense Museum, retired colonel Yuri Knutov, told Lenta.Ru, a Russian newspaper, “the robot can thus identify NATO-made tanks” and will be “armed with a machine gun and an anti-tank missile with a range of up to about six kilometers.”
Honestly, all of this is pretty hilarious stuff.
Machine learning and artificial intelligence are real disciplines, and Russia doesn’t entirely lack technological and programming talent. It’s entirely possible that you could develop and effective autonomous battle-tank driven by AI that can adequately detect between friend and foe given lots of money, lots of time, honest, hands-off project management, and sophisticated, iterative, trial-and-error proving over a decade or more of time.
All things Russia isn’t going to have or do. If they could adequately identify friend from foe on the battlefield (especially given how much kit Ukraine shares with Russia), then they’d already be using such technology to prevent the numerous, documented friendly fire instances Russia has suffered from. And training AI to do that is something like six orders of magnitude harder than training troops to do it.
And we all know Russia sucks at training its own troops as well.
Russia’s military is so demonstrably backwards that they can’t even have their army and air force communicate with each other in real time for combined arms operations. And yet we’re supposed to believe that they’ve developed cutting edge autonomous battlefield AI?
Pull the other one.
Russia has a long history of vaporware, and Russia’s previous attempt at field trials for a semi-autonomous AFV in Syria was a hilarious disaster. And it was plagued by bog-standard mechanical failures. Autonomous driving is a whole lot harder.
There’s a small possibility that they’ll get this thing into the field and immediately start blowing away its own troops, but a far more likely outcome is that it never sees the field at all, just the latest case of Russian Wunderwaffe vaporware.