File image.
Russia’s Rubikon advanced unmanned technology center released on August 1 a video dedicated to the anniversary of its founding, including footage showing for the first time the launch of the advanced V2U loitering munition.
The presence of the V2U was first revealed by Ukrainian media earlier this year, with the Russian side confirming the presence of the loitering munition later.
The V2U employs artificial intelligence (AI) in multiple ways, from its jamming-resistance navigation to automated optical targeting system. The loitering munition also features a unique AI-assisted protection system that detects incoming threats, like interceptor drones, and direct maneuvers to avoid these threats.
In addition, the loitering munition utilizes AI to operate in “swarm” mode, coordinating with other drones with a high level of autonomy.
The V2U is also reported to feature a very unique jamming-resistance, two-way data link, with multiple channels including one based on Ukraine’s own mobile network.
In the footage released by Rubikon, the V2U is seen being launched via a catapult. The simple launch system was spotted before.
The basic version of the V2U is propelled by an electric engine. It is said to have an endurance of 60 minutes with a speed of 60 kilometers per hour while armed with a warhead weighing 3,5 kg. A larger, longer-range version powered by a piston engine was reported, but its presence is yet to be confirmed.
The V2U is still in an experimental phase. If mass produced, the loitering munition would cause a serious problem to Kiev forces as it would be able to attack targets behind the line of sight with an unprecedented level of automation. In addition, the AI-based systems being developed for it could be later employed in other unnamed systems.
_______________________________________________________________________________________________________________________
SouthFront: Analysis and Intelligence
NOW hosted at southfront.press
Previously, SouthFront: Analysis and Intelligence was at southfront.org.
The .org domain name had been blocked by the US (NATO) (https://southfront.press/southfront-org-blocked-by-u-s-controlled-global-internet-supervisor/) globally, outlawed and without any explanation
Back before that, from 2013 to 2015, SouthFront: Analysis and Intelligence was at southfront.com
stop saying “a.i.” when you mean “program.” there is no such thing as artificial intelligence. this ai propaganda nonsense is ridiculous. by this logic, a calculator is also an artificial intelligence. it’s a homing and target-orientation program, not artificial intelligence.
ai is intelligent enough. openai’s o3 already beats humans at most tasks. ai is missing the tools to interact and rival humans.
we saw how o3’s ability to zoom into images gave it a massive leap from o1 in terms of image recognition.
it still isn’t perfect, but they are teaching them to get better.
agent-0 (chatgpt agent) also just launched. it is limited, but still very useful. agent-4/5 in 2027 would wipe the floor with putler and his barbaric regime.
nothing like a moron americunt to believe better is not thinking or feeling
you also clearly don’t understand the term “ai”; in dartmouth workshop, a few researchers coined this term “ai” (i think it was mccarthy), and it basically is defined as: “any system that is able to at least imitate human intelligence, one way or another.” we see ai in so many areas, like chatgpt, who is a transf0rmer-based model (part of deep learning, which employs artificial neural networks). sure they don’t possess “true” form of intelligence.
but any system that employs machine learning techniques (especially artificial neural nets) qualifies for the term “ai.” as they imitate human intelligence. we see lots of innovations in this realm. openai’s new imo-gold winning medal shows ai isn’t far from wiping regimes and replacing them with democracies.
amerikan totalitarian fascism needs ai to think for you since you cannot feel
so, that means zelensky’s regime is not long for this world. about time. who’ll be the next selected one?
“only in amerika do people act like machines are treated like machines–only in americuntistan are machine metaphors used to describe human behavior”. geoffrey gorer
“since amerikan political technique is inferior they require more resources devoted to propaganda”. jacque ellul
they don’t want a entity that think by itself, it would find what’s wrong in our world an find who’s responsible for all that mess.
“machine vision” sounds better. or automatic target selection. robotic guidance at last.
ai is already destroying putler lol. they are too slow to catch up now.
nothing like a dumb inferior amerikan to amuse
“the more totalitarian a society the more conformity is required. amerikans are the most conformist peoples on earth”. jacque ellull
usa is likely to win the agi race. openai’s launch of agent-0 (chatgpt agent) is the beginning of it all.
since all americunts are moron robots we submissively need robot to produce our immorality and our belief that shit tastes good
amerikunt = post teen ‘muricunt’
the race to superintelligence isn’t gonna stop. but putler’s russia isn’t any good at coming up with new innovations that could contribute to the development of agi.
the us already has a baby agi. in 2026-27, we might see these systems not only scale up, but get more intelligent as their techniques improve.
currently, we are in the middle of agi. 2022 was the beginning, 2025 is the middle, 2027 is the year of general intelligence declaration to the public.
lots of blabber that signifies little and looks like despair.
ai has a lot of problems. it has some uses, but has been nicknamed ‘jagged intelligence’ because it can’t achieve a lot of very dumb tasks.
the ceo of google said ia helped it achieve 10% productivity increase. which is nice but very very far from all the hype that is circulating.
ia agents are completely failing, and appart for helping in some simple tasks, ia for common people is not very impressive.
right now, the investment in ia in the us is $450 billions per year, while its revenue is $40 billions (revenue, not profits). the economic model of ia is not understood at all, and ia companies are starting to raise prices because what prices are practiced now are not sustainable. ie we don’t know what is sustainable economically.
as for agi, just thinking we are close to it just shows how ignorant you are on the subject (no surprise here).
nonsense. nvidia is now capped at $4 trillion dollars. there is clearly a profit for this. many companies are pouring billions and trillions into these gpus, tpus that power ai models.
ai models are getting better by the day. of course, a lot of their intelligence would be locked behind paywalls, nontheless, they would still have their use-cases in their domain-specific tasks for now. but in the future agi would do anything for you.
companies are buying a ton of nvidia gpus fore sure. that is the $400+ billions of investments per year. and that is great business for nvidia for sure.
but you have to be able to produce profits in excess of what you invest and spend to make it economically viable. capitalism 101. that is not the case now.
as for agi, this is not for anytime soon as i said.
reasoning models have already fixed the problem of “jagged intelligence,” demonstrating their excellence at solving tasks otherwise thought too easy by humans. but humans can also fail simple tasks despite being too intelligent. the issue with llms is lack of time, tools access and memory. once those are fixed, we might declare agi.
reasoning models don’t really work as they only reproduce things they have seen in their training data but can’t extrapolate. and they make a lot of errors. that is why for example apple has not deployed “apple intelligence” yet : the quality is not here.
everybody understand that llms cannot produce agi, and we need to produce models able to handle mental maps that really model the world.
you talk nonsense and know nothing.
reasoning models aren’t just copying training data – they chain thoughts and make analogies beyond what they saw. apple’s rollout is about privacy and hardware, not proof that llms are broken. yes, today’s ai makes mistakes, but adding memory, context and tools fixes that. these systems are already building world models, even if you refuse to see it.
yes man.
that is just why apple produced a paper called :
the illusion of thinking: understanding the strengths and limitations of reasoning models via the lens of problem complexity
that just explained what i told you.
now you can believe the bs that those that invested 10s of billions of $ and have a vested interest in selling you their stuff but change their narrative every few months (not unlike nato you’ll notice).
or you can understand that you are being lied to and that perceiving reality is the only way forward (but is painful, granted).
the race to superintelligence, that explains the 50 million braceros they imported into da good ole usa over the last 50 years. dummies like you, chat. hahahaha
since amerikans are tabula rasa robots ai helps them be more submissive weaker more stupid
lgbt pride must be forbidden in netherland amsterdam…
and everywhere.
it seems the russian army gets new and updated improved ordnance daily . meanwhile ukraine can’t rotate troops or re-supply
v2u should be v2fu.
“in addition, the loitering munition utilizes ai to operate in “swarm” mode, coordinating with other drones with a high level of autonomy.” now it’s really amazing scary.