I saw a clip of a guy saying that Ukraine is going through a command reform:

  1. Military officials will be forced to adapt or leave.

  2. Military command will be highly "streamlined" and "automated" with "new technology."

The video was claiming this as a huge win and that Russia would be outdated by comparison and that military personnel, infrantry and such, would be treated more humanely.

However, what it sounds like to me, what I'm hearing instead of those claims, is that all military power and authority is being centralized and that ChatGPT or Grok or some shit is being given reigns of a fucking drone swarm. Is that what is happening? Two things I can say quite confidently is that: consolidating such power such that less total people have any say in matters has never lead to more humane treatment, historically, and that AI chatbots have never shown any capacity for morality nor hesitation in regards to the sanctity of human life and nature.

Is the Trump admin pushing this bullshit on Ukraine? Am I reading too much into this? I'm naturally very skeptical of things like this.

EDIT: I'm pretty sure the video I saw was just some AI Techbro bullshitting and in reference to changes from Brigades (smaller departments) to Corps (big offices managing their own small departments) structure. Nobody can really find info on what he was talking about, AFAIK.

you are viewing a single comment's thread
view the rest of the comments
[–] 12 points 1 month ago (1 child)

There are other AI systems than LLM slop machines.

UA has been successfully using AI vision systems for a while, which is why Russia is trying to camouflage trucks with dazzle paint.

  • source
  • hideshow 2 child comments
  • [–] [S] 3 points 1 month ago* (1 child)

    I got no results for "UA" AI vision military in search engines, but regardless there are no AI capable of competently replacing military command or any other complex decision making task.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 6 points 1 month ago* (last edited 1 month ago) (1 child)

    I can verify that I've read about Ukraine developing and using "AI vision" for UAVs for a while.

    You may recall articles mentioning operators selecting targets outside of electronic warfare (EW) and then triggering the "AI", which would head for the target. Once the UAV hits the EW, the "AI vision" (mission learning) takes it all the way, making EW generally ineffective.

    I agree with you that LLM are slop and garbage. I don't think that is the "AI" Ukraine is using. I think it is more akin to video game AI, which is just well designed logic ("AI") that controls non-player allies/enemies/environment/economy/etc.

    If tasks are well defined and repeatable they can be logiced away ("AI-ed"). A not particularly related example of what this could look like is weekly posts I make about fediverse software.

    I gather the data from sources and then do analysis on them to get insights/takeaways from the data. I could do this by hand/spreadsheet, but that would be annoying and time consuming. I didn't use LLMs, but I could/am arguing that the code I developed is artificial intelligence (AI, in the vein of video game AI) that automates a well defined and repetitive task.

    I hope/expect this is the "AI" Ukraine is developing. If you want to check out what I mean about my post examples, you can head to !fedilytics@piefed.zip / https://piefed.zip/c/fedilytics.

    Edit: fixed spelling.

  • source
  • parent
  • hideshow 2 child comments
  • [+] [S] -7 points 1 month ago (2 children)

    A UAV isn't going to sit in the command room and make phone calls.

  • source
  • parent
  • hideshow 4 child comments
  • [–] 10 points 1 month ago

    I provided two different examples of "AI". The "AI vision" UAV (machine learning) and a command-control logic design (video game "AI").

    Would you call a video game's AI an LLM? I wouldn't because it's not. No LLMs were used in my examples/descriptions.

  • source
  • parent
  • [–] 8 points 1 month ago (1 child)

    A UAV isn't going to sit in the command room and make phone calls.

    I very highly doubt that Ukraine is trying to automate command rooms to give orders based on GPT hallucinations.

    But you can definitely have machine learning algorithms learn how to effectively control drone swarms. Or learn to improve logistics and things like that.

    Machine learning algorithms are also called AI and have been around for way longer than GPT or Grok

  • source
  • parent
  • hideshow 2 child comments
  • [–] [S] -1 points 1 month ago (1 child)

    If your logistics are being managed by AI then you're fucked, bro.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 1 point 1 month ago (1 child)
  • [–] [S] 0 points 1 month ago (1 child)
  • [–] 2 points 1 month ago (1 child)

    I specifically excluded generative AI like GPT, or Claude in my post, so why are you bringing that back up?

  • source
  • parent
  • hideshow 2 child comments
  • [–] [S] -1 points 1 month ago (1 child)
  • [–] 3 points 1 month ago (1 child)

    I very highly doubt that Ukraine is trying to automate command rooms to give orders based on GPT hallucinations.

    But you can definitely have machine learning algorithms learn how to effectively control drone swarms. Or learn to improve logistics and things like that.

    Machine learning algorithms are also called AI and have been around for way longer than GPT or Grok

  • source
  • parent
  • hideshow 2 child comments
  • [–] [S] -1 points 1 month ago* (1 child)

    "Doubt [they] automate command rooms to give orders based on GPT hallucinations."
    "But you can"

  • source
  • parent
  • hideshow 2 child comments
  • [–] 2 points 1 month ago (1 child)

    "But you can"

    Are you unable to read further than the "but you can" part, where it says "machine learning algorithm", and not "generative AI"?

  • source
  • parent
  • hideshow 2 child comments
  • [–] [S] 0 points 1 month ago* (1 child)

    An LLM is still an ML, and if you didn't mean to make an exception then there should be no "but" there should be no "however", and furthermore logistics requires understanding of complex situations that simple statistical models, regardless of simple ML or complicated LLM, is inadequate for.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 2 points 1 month ago

    My third paragraph excludes LLMs, why can't you just go back and read the whole comment instead of just insisting to be right.

    Also if you do that you will find "learn to improve" not "manage completely autonomously"

    This entire thread is you just ferociously arguing a series of strawmen. Just stop.

  • source
  • parent