Pages

Monday, July 6, 2020

The Illogic Abounds!

By now, we’re all familiar with the nonsensical fantasy of small units, deep in enemy territory, wreaking havoc, and winning wars.  Of course, the key to this (well, one of the keys, at any rate) is the ability to resupply these units deep inside enemy territory.  No one has yet explained how this would occur.  Defense News website, however, has an article describing the general concept and the article perfectly illustrates the degree of wishful thinking and hand-waving away of problems that has to occur to even begin to contemplate this kind of resupply.

… delivery of logistical payloads to Marines on a beachhead can be done completely with unmanned platforms. For example, small to medium robotic ground vehicles loaded with supplies could be carried by one of these unmanned logistical craft. Rather than Marines exposing themselves to hostile fire while unloading supplies on the beach, robotic ground vehicles or aerial drones disembark from the vessels and deliver cargo directly to the Marines in a more secure location. (1)

Let’s break those sentences down and analyze them.

‘delivery … can be done completely with unmanned platforms’  -  This assumes the ability of a small, non-stealthy, unarmed vessel to self-navigate hundreds or a thousands of miles through enemy controlled waters and skies without being detected and attacked – and to do so on a regular, recurring basis.  Does this seem even remotely realistic?

Wait, it gets better!

Rather than Marines exposing themselves to hostile fire while unloading supplies on the beach, robotic ground vehicles or aerial drones disembark from the vessels and deliver cargo directly to the Marines’  -  So, if Marines were to try to unload supplies they’d be exposed to hostile fire and killed but robotic ground vehicles will, somehow, be immune to destruction by that same hostile fire?  Does this seem like anything but pure fantasy?

Still better … !

Further, if Marines trying to unload supplies would be exposed to hostile fire, wouldn’t the supply vessel, itself, beached on the shore for unloading, also be exposed to hostile fire and, being small and having no defensive armament or armor, be quickly destroyed along with its entire cargo?  Are you beginning to grasp the degree of fantasy and delusion that goes into this concept?

Finally, we note from the article that the authors are described thusly,

Wayne Prender and David Phillips are senior vice presidents at Textron Systems.

Hmm …  Just out of curiosity, I wonder what Textron Systems makes?  A quick check shows that, among other things, it makes …     military robotic vehicles.  That’s right, they make the very robotic vehicles that are at the core of this idiotic concept!  Who’d have guessed?!  This sounds like Textron executives contriving a poorly thought out marketing pitch to try to sell their product regardless of the degree of insanity associated with the entire concept.  If they can make sales, who cares whether it works or not, right?  The really unfortunate part of all this is that the Marines seem to have bought into this nonsensical scheme hook, line, and sinker.

I’m also disappointed in Defense News for even printing such an idiotic article.



_________________________________

(1)Defense News website, “Use existing and planned craft for unmanned logistical resupply”, Wayne Prender and David Phillips, 9-Jun-2020,
https://www.defensenews.com/opinion/commentary/2020/06/09/use-existing-and-planned-craft-for-unmanned-logistical-resupply/

67 comments:

  1. I am trying to put what I know about AI and deep learning into the context of military operations (about which I know a lot less) : isn't it one important quality of a military commander to be able to tactically outsmart his oponents ? Which means behaving in a unpredictable manner ... And AI, especially deep learning which is vastly used in autonomous vehicles cannot account for this for a very simple reason : it tries to imitate known behaviour and solve known problems, that is what the deep learning part is about. Don't get me wrong I think AI is brilliant but in the right context beacuse it can do things much quicker and more reliably than a human being but I am very dubious about autonomous systems trying to outsmart a good commander or even a soldier.

    ReplyDelete
    Replies
    1. A good comment.

      One of the problems with all militaries is summed up in the old adage that generals are always preparing to fight the last war. To an extent, this is what you're saying about AI when you talk about imitating known behavior and solving known problems.

      Delete
    2. No one becomes a good commander or a good soldier without lots of training and mentoring to develop the required skills and experiences to succeed as a commander or a soldier. And, I think the same would apply to AI for military applications. Whether all that can be boiled down to a download remains to be seen.

      But, the trick is to get AI to reason like a human being, to make decisions based on incomplete information, and take risk as appropriate to the mission. Things that we do all the time.

      At the same time, I wouldn't expect an AI ship to be totally on their own. They probably would do the day-to-day navigation, but I'm sure they would get regular updates on the weather, positions of friendly forces, and the positions or suspected positions of enemy forces. And, any other updates that would affect its mission. Plus, AI ship would be reporting back what it finds with its sensors.

      Delete
    3. Very nice comment!

      "make decisions based on incomplete information, and take risk as appropriate to the mission. Things that we do all the time."

      Spot on. A child can make the right decision if every relevant piece of information is known. The challenge for business leaders - and what sets good managers apart from the run of the mill - is to make the right decisions with only partial information. The same is true of combat leaders. No one ever has all the information in combat. Some people, in business or combat, have the knack for making the right decisions based on only partial information. In my experience, it's an intuitive process that the person can't describe. As you note, how do we get an AI to replicate an intuitive process? I have no answer but you raise a great question. It will be a major challenge for AI developers.

      "I'm sure they would get regular updates"

      I would assume so. Of course, that raises the philosophical question about what constitutes 'autonomous'? We consider humans to be autonomous and yet we receive information inputs all the time. Is an AI less autonomous because it receives updates? On the other hand, if we're constantly issuing orders to the AI (ship or aircraft or whatever) then it devolves into just a glorified autopilot, which we already have. Of course, this is exactly what a ship captain is in this day of micro-managing - just a glorified human autopilot for some shore-side commander!

      Where AI will earn its money is when communications break down and there are no updates or instructions. What will the AI do then? A person in that situation would weigh mission orders, capabilities, crew morale, damage, risk, fuel supplies, and a thousand other things to decide what to do. Do we have that programming capability in the foreseeable future when we can't even program a simple logistics program (ALIS)?

      "No one becomes a good commander or a good soldier without lots of training and mentoring to develop the required skills and experiences to succeed as a commander"

      I've got a post coming on this exact point. You're going to love it!

      A really good comment! Thanks!

      Delete
  2. The obvious logic is that such a vehicle would need to be able to defend itself to have any chance of success. Personally, I'm not comfortable with AI-controlled killing machines being unleashed upon the world - if we have them, they'll have them too. AI & nano-machines need to be in the same category as Nuclear, Biological and Chemical warfare.

    ReplyDelete
    Replies
    1. Nanomachines? Have you been playing too much Metal Gear Solid 4? They don't give superpowers or anything like that.

      Delete
    2. https://www.nanowerk.com/spotlight/spotid=1015.php and this was 14 years ago!

      Delete
    3. Just a little perspective and food for thought. (That said it's also worth noting that after the Cold War, the US has pretty much taken a fairly big step backwards from autonomous weapons and is very much about having a man in the loop.)

      Delete
    4. " the US has pretty much taken a fairly big step backwards from autonomous weapons and is very much about having a man in the loop"

      Quite the opposite. First, there is no such thing as an autonomous weapon. If a weapon has an 'off' switch then it's not autonomous and has a man in the loop. Every weapon has an 'off switch' of some sort.

      That aside, the entire thrust of US weapon development has been about greater autonomy. We now try to create weapons with multiple sensors so as to allow the weapon to choose its targets and even specific aim points. Our goal is for weapons to be 'fire and forget' which emphasizes autonomy after release.

      For example, LRASM is touted for its ability to 'hunt' for targets, make its own decisions about prioritizing targets, and self-allocating weapons.

      For example, Aegis has a full auto mode.

      And so on.

      You may be confusing autonomy with robotics.

      Delete
    5. "Quite the opposite. First, there is no such thing as an autonomous weapon. If a weapon has an 'off' switch then it's not autonomous and has a man in the loop. Every weapon has an 'off switch' of some sort."

      Sure, I'm just pointing out to Anon that the common fear the mainstream media tries to stir up - weapons that select their own targets - has already been a reality for decades at this point. I was thinking of the Cold War idea of using Harpoon and Tomahawk at extreme range, fired on bearing only launches: you're sending the missile beyond the radar horizon, out of range of your control, and leaving engagement purely up to whatever target enters its seeker. That's how I interpret autonomous, anyhow: target discrimination and engagement is purely up to the missile in that instance.

      I suppose it's more accurate, perhaps, to say that an antiship missile's engagement cycle, when fired on a bearing only launch or at a specific grid, beyond sensor range and control range of the firing warship, overall, is semi-autonomous. The initial shoot-don't shoot decision is made by a human officer, but target selection is performed by the missile once it arrives at the grid point.

      Call it autonomy, call it robotics, it's one answer to the problem of weapons that outrange the firing ship's sensors. The question for discussion then shifts to whether this _should_ be the pursued option. (Or maybe it's the only workable option, if we assume that it'll be impossible to push offboard sensor assets far enough of the firing ships in sufficient numbers for sensor density and resilience against attrition from enemy fires.)

      Delete
    6. I think the distinctions everyone is making and concerns they have is about platforms (ships, planes) having the ability to fire weapons on their own, and thats a huge step. But thats true, in that its only another step (or three) from weapons that, once fired, make their own decisions... But I digress... The article is ridiculous, and everyone should beware of things written by people selling the systems!! Theyre in it to make a buck, and the idea that theyre cozying up to a fanciful battle plan shows that industry is ready to fill a demand, whether its worthwhile or not. Thats all bad. What happened to the days of a company developing a weapon on their own, then competing with others and trying to sell theirs as the best one?? Now we finance companies to develop garbage that never meets expectations even after massive cost overruns. Will companies ever have a conscience again? Will the military ever look at and ask for sensible systems with solid CONOPS again?
      The world wonders...

      Delete
  3. "AI & nano-machines need to be in the same category as Nuclear, Biological and Chemical warfare."

    That's laudable. Now, how do you get China, Russia, Iran, and NKorea to agree and abide by that?

    ReplyDelete
  4. The same way we got treaties on NBC. NKorea will have to be done the hard way but the others can still be negotiated with. With the right political will, there are levers on all 3 that can bring them to the table.

    ReplyDelete
    Replies
    1. You may be forgetting that those countries have a history of violating the treaties they sign. China ignored UNCLOS and annexed illegal islands. Russia routinely violated the INF treaty. Iran freely admits violating the nuclear agreement. And so on.

      Treaties don't mean much if the parties involved have no intention of abiding by them.

      Delete
    2. The NBC treaties have worked fairly well in practice and the key point is there have not been any major NBC events. Treaties are contracts - if you don't enforce the penalties when they are breached, then more abuses occur. Weak behaviour by successive US governments is the background to all the above - the US can easily kick China out of the various islands if it wants to and could put Iran back to the Stone Age equally easily. We can even do it in a deniable way using 'fishing fleets' and third parties or even with stronger sanctions - we just lack the political will.

      Delete
    3. Just read how the two carriers in the Pacific are "operating at the pleasure of the PLA", and that the SCS is "fully within their grasp"...
      At least we see two CVNs together for a change. Its not the four that CNO (or most of us) wants but its somthing I suppose!!

      Delete
    4. "the key point is there have not been any major NBC events."

      Do not confuse correlation with causation. There is no evidence that the lack of a major NBC event is due to any treaty. Far more likely is that it's simply in the best, selfish interests of the individual countries NOT to initiate a major event.

      I would also note that events of the 'C' portion of NBC HAVE occurred. Here's a site that purports to list events: NBC events

      Delete
  5. Why does it have to be one or the other? A vehicle that can go from A to B on its own could be useful (assuming the price /reliability is right). However the ability of a human to take over is also useful and adds resilience. When the internal combustion engine was developed we didn't stop using horses the next day, as the technology and reliability improved so it gradually took over.

    ReplyDelete
  6. Perhaps we could do a comparison trial. Let's put 3 platoons in the boonies and see which can deliver a palette of ammunition to them most successfully, Textron, Dominos or Amazon?

    ReplyDelete
    Replies
    1. Domino would be 7 minutes late but contract would make you tip driver anyway.
      Amazon would deliver half on time, other shipment a day late, and damaged, requiring an insurance claim.
      Textron would arrive two years late, twelve million over budget, to the wrong place, but a MilNews article would explain how the milestone was significant, and groundbreaking in its achievements....

      Delete
    2. It seems Amazon experience differs regionally. Some basic things we can have to the house while watching the same TV show we ordered them during. Fast as Postmates who still lets you tip after the job is done.

      Delete
  7. Odds on the Domino's on would get there warmer

    ReplyDelete
  8. What small / medium UAV can lift a missile big enough to conduct a decent anti-ship strike? Especially as Western missiles generally come pre-loaded in pods? A UH-1Y or KMAX can carry only 1x ATACMS or LRASM, missiles with greater range are probably going to push their envelopes.

    ReplyDelete
  9. "Ladies and gentlemen, welcome aboard flight 1515. Your aircraft is being flown totally by artificial intelligence today. There are no humans in the cockpit. This is a recorded announcement. But do not worry, nothing can go wrong... go wrong... go wrong..."

    ReplyDelete
    Replies
    1. What makes you think the two guys up front in pressed white shirts are pilots ? You can't see if they're flying these days, locked doors etc. Maybe there up there just punching numbers in the flight computer ?
      Think of the shareholders, who can afford pilots ?

      Delete
  10. This is just getting dumber and dumber. I don't think there is a viable Marine mission from the platforms that the Navy provides.

    ReplyDelete
  11. "So, if Marines were to try to unload supplies they’d be exposed to hostile fire and killed but robotic ground vehicles will, somehow, be immune to destruction by that same hostile fire? Does this seem like anything but pure fantasy?"

    So I think i've been a little unclear on what I was saying, so I'll unpack it a little more. It's fantasy that's insiduous because there is some glimmer of truth to it:

    Conceptually, there is a legitimate benefit in unmanned resupply aircraft and vehicles that can brave enemy fire to drop off supplies, without risking human piltos. I'm reminded of the Hueys who got so badly shot up delivering ammo to LTC Hal Moore's troops at Ia Drang valley with hot LZs - 55 damaged, 4 shot down. So if you've got a fleet of robot helicopters to fly supplies in, and logistics robots to unload those supplies, and the only thing you're risking is machines that have been paid force in a budget passed years ago, well, that starts to look pretty appealing, doesn't it?

    As a practical matter, I'm in agreement that trying to offload supplies in a hot LZ that's under direct fire from the enemy is a fool's errand, whether you do that with drones or Hueys. If you think you can unload in a hot LZ while being exposed to direct fire from the enemy, yes, that's a fantasy. (This is discounting desperation moves, like how Bruce Crandall picked up his Medal of Honor at Ia Drang, flying in ammo and flying out wounded soldiers.)


    As an aside, two of my observations on the proliferation of drone programs in the Pentagon:

    - Politically, drones are great for the optics of politicians. The American public is more concerned (and rightly so) about casualties then they are about drones shot down. I think we're definitely going to see more research and effort put into pursuing drones in combat. Recall how the American public soured on Vietnam with the casualty counts - even though the tide had turned and a military victory was possible.

    - Financially, this is a great benefit for the textron company. These drones are going to get shot down in droves, which means the pentagon will order more drones in droves to replace the shot down drones, thus stimulating the economy and the bonuses of Textron executives! :V

    (This is sarcasm and a bitter joke.)

    ReplyDelete
  12. Recently, i have the opportunity to read a book by Christian Brose called "The Kill Chain: Defending America in the Future of High-Tech Warfare". I think that the book addresses the few concerns that we discussed here pretty well. Firstly, the author (and me) believes that there is a place for autonomous systems in the military consists of specific combat requirements and logistics support. Think of it this way, you just don't give a random person a gun and then tell them to fight someone else right? There will be training and rules that these soldiers have to follow. This also applies to the technology that we are replacing them in combat, a commander will explore all the scenarios that a normal soldier have to go through, set hard-coded limitations that they cannot break (PIDs possibly). I would also like to point out the CIWS system have a fully autonomous weapon that is activated as a last resort defense (mostly because humans could not comprehend the degrees of targeting they need to do) but i think that this presented our acceptance that if the risks far outweigh the benefits having such system non-autonomous, we will gladly incorporate autonomous systems regardless of our beliefs. I also believed by developing combat systems and fully mastering them will allows us to provide the counter of such systems if our enemies if they ever developed them. Last but not least, a few people argued that autonomous systems belong into the same realm as NBCso therefore needs to be banned or restricted, isn't the best way to counter other countries WMDs is having the ability to develop (or stockpiling, in the case of nuclear weapons) warrants as the best deterrence? There is also some people advocating for having a treaty to stop the development of such weapons, I will remind you that in history, the only cases that a treaty was even considered is when major countries all have equivalent capabilities on the same weapons. On other notes, logistical support systems could present a big leap in terms of labor training and development. I think of robots trailing a technician providing them with tools and parts for repairing and system that do tasks that that will mostly stays the same in real combat conditions.For example, the USAF have developed a paint system that speeds up labor intensive tasks that requires high accuracy(https://www.pfonline.com/articles/f-35-gets-new-paint-system). This reduces the amount of people that is required to do maintenance (one but help to wonder how difficult it is to maintenance these new robotic systems, and does it require technicians to acquire new skill-sets or even specifically-trained technicians?) and frees them up for other purposes. One help but wonder that if we could have robots perform repetitive tasks without negatively affecting the ship chances of survival or its combat systems even in attrition warfare (This probably will makes electrical systems a more lucrative targets). One of the problems that the author mentions that there is a lot of information collected but they are usually scrapped because we don't have enough people to analyze them and he proposed the use of deep learning systems to track the patterns, providing the person in charge that allows them to make an informed decision. I find the idea pretty interesting but I don't think more data is the problem of the military, it's more of the willingness of current commanders to take risks but that's just my opinion.

    ReplyDelete
    Replies
    1. A nice comment. Without indicating either agreement or disagreement with your points, I note that you've offered your thoughts in a coherent and reasonable manner. These are the kind of thoughtful comments that I encourage.

      I like your point that we already have autonomous lethal weapon systems. In addition to CIWS, the AEGIS system has a fully automatic mode which is, in fact, its preferred combat mode.

      " I don't think more data is the problem of the military, it's more of the willingness of current commanders to take risks but that's just my opinion."

      Again, without agreeing or disagreeing, how would you conceptually propose programming 'risk' into a computer?

      Delete
    2. Risk itself is a human concept and how we perceive risks differently between culture, generations and even people. For instance, our freedom and rights define who we are and we will go to great odds to keep this democracy alive. Chinese values are placed around stability and safety of the country (synonymous exchange with the Communist Party) and etc. Let's apply the concept into actual combat. If you are given 2 choices of high risk, high reward or low risk, low reward; this is practically a coin toss. There will be a lot of people could pick one over the other but you have to realize that we have a bias inherently built into us that reflects our altitude and our past experiences. Because of that, it is likely impossible that we could program robotics to be bias-free in the risk management area. However, just like humans, the commander who deploys their subordinates into combat have to has full confidence in their abilities and the predictability of their actions. What i am proposing is the robots will be programmed with the preferences of their commander (aggressive or defensive assessment of risks)and the commander is held responsible for the actions committed by these units. Data might serves as a way to assess the situation but data could always be falsify or faulty but serves as a guidance for the person in charge. I do wanna point out that for most of history, information is usually reported with a bias and decisions are made with bias. That makes you think of events that biased decision make on information from bias sources even though all other evidences point to the contrary. If you achieve success, you are hailed as a hero and a model for people to emulate. If you failed, you are considered insane and unacceptable.

      My Opinion:
      I also wanted to address the fact that i said "...the willingness of current commanders to take risk...". What i was talking about our low-intensity conflict mindset where high risk failures are remembered and glorified in Hollywood. Officers who make errors are either pushed out and discharged where they should only be punished but let them learn mistakes.

      Delete
    3. "What i am proposing is the robots will be programmed with the preferences of their commander (aggressive or defensive assessment of risks)and the commander is held responsible for the actions committed by these units."

      Again, without either agreeing or disagreeing, what a fascinating concept. This suggests a movement away from truly autonomous, independent AIs and towards a philosophy of using the AI as a subordinate tool that reflects the commander's bias. This offers both advantages, such as closer adherence to the commander's intent, and disadvantages, such as failure of the AI by prioritizing the commander's biases (possible wrong or flawed) over the best, pure AI decisions.

      Neither agreeing nor disagreeing but fascinating to consider! You're offering a different take on AI and I appreciate that kind of thinking. Great comment.

      Delete
  13. P/S: To Anon, the one i have discussed about AI a few days prior, I have started to comprehend the picture that you tried to paint but i have a few issues with it. We talked about how unmanned planes can be semi autonomous systems where it is controlled by a group leader. Such systems are too vulnerable to outside interference, and the transferring information makes detection highly likely. I think this also puts too much stretch on the human controllers and they will be focused on doing their own tasks but I think there is a place for an autonomous wing-man that follows predefined training and if conditions allow, incorporate information from other aircraft to make better decisions. It could infuse information of any connected aircraft but if it is blocked, it will only needs it own information to make informed decisions. I think right now, there is a way such a system could start to work (deep-learning is its infancy and there still many unknowns in such development) but maybe 40-50 years in the future, we might have something like this.

    ReplyDelete
  14. (Don McCollor)...Logistics is the Achilles heel of behind-the-lines (aka covert) operations. Combat effect increases but supply requirements also increase with the size of the unit. In a real hostile environment this would seem to limit it to small relatively ineffectual groups for military operations (although intelligence information would be a force multiplier). As for the robot supply boats and robot porters, there would be no need for the enemy to destroy them, just have a watcher (or another robot) report the position. Now the enemy would know exactly where they are. UAVs air dropping supplies would be better (but again limited to small units)...

    ReplyDelete
  15. There are lots of other ways AI could contribute, to name a few :

    - looking at huge data sets to help set military priorities, for example, does building a new carrier have more real-world impact than out-bidding a Chinese company for a new port in Africa etc?
    - better cost forecasting and management for military projects
    - improved wargaming by letting AI look at enemy commanders and equipment past actions and performance rather than allowing convenient assumptions and western minds acting as the enemy.
    - exposure of fraud in military accounting and store keeping through smarter analysis.
    - etc

    The military seems to be working in the right direction with weapons and even ALIS is the right idea (badly executed) but a good chunk of damage to military effectiveness comes from bureaucracy, poor administration, fraud and politics - even the Chinese can't damage the US military as much as Congress. What if the military used AI to wargame managing politicians, analyse all their past actions(and misdemeanours), track their links to contractors and lobbyists, work out what arguments influence them most when voting?

    ReplyDelete
    Replies
    1. While nothing you said is wrong, per se, everything you said is already incredibly obvious and does not require AI. Just for illustrative purposes, let's review your itemized list.

      -Analyzing spending patterns is great but we already know where we should be spending our money - we just ignore it! Isn't this entire blog about how to better spend our money? We don't need AI for that.

      - Better project management is a goal that EVERYONE talks about. Again, we just don't do it. Instead of punishing poor program management performance and holding people accountable, we reward failure with promotions. We don't need AI, we need accountability.

      - Wargaming is only as good as the assumptions we input and the results we allow. When we artificially input ridiculous assumptions and ignore bad results, there's no point to even doing the game. We'll ignore AI even more than our current process because it will give us even more bad news.

      -We don't need AI to expose fraud and waste. We already know it's there and even where it is. We just don't care. The recent audits of the services pinpointed tons of waste and we did nothing.

      What we lack is not AI, it's actions and accountability. We already know everything AI would tell us.

      Delete
    2. I'm not sure we're actually disagreeing - take wargaming for example : my point is that you don't let the generals input the assumptions, you let the AI analyse reality and set the parameters and then see how the generals perform. The AI can then analyse the performance of the generals and you sack the bottom 10%. I bet the military performance improves in leaps and bounds. Only an idiot lets the military test itself.

      Delete
    3. "you don't let the generals input the assumptions, you let the AI analyse reality"

      There's a fallacy in your statement and I've got a post coming on this exact subject. You'll enjoy it.

      "The AI can then analyse the performance of the generals "

      Actually, it can't. Post coming!

      The only problem I have with your suggestion is that you'd be imposing a MASSIVELY expensive developmental program that, even if it worked perfectly, would only tell us what we already know! I'll tell you what … pay me a tenth of the cost of developing the kind of AI you're describing and I'll tell you exactly what are problems are, where they are, and how to solve them - or you can read this blog for free and find out the exact same information.

      This reminds me of the initial reaction to computers: if it came from a computer, it must be right. Now, I see a similar tendency: if it comes from AI, it must be right. Of course, both statements are categorically false. Again, post coming.

      Delete
    4. AI is not some all-knowing God.
      The problems within the Navy and US armed forces in general are well-known and documented, they are also ignored or even made worse for ideological reasons, private profit, career advancement, corruption, etc.

      Furthermore, any kind of AI will either be used merely as cover for the current stupid trend or ignored/reprogrammed if it does not give the desired results.

      Delete
  16. I await your post with interest but it's already starting :

    https://www.theguardian.com/technology/2019/apr/07/uk-businesses-using-artifical-intelligence-to-monitor-staff-activity

    https://www.headspringexecutive.com/ai-impact-employee-relations/

    ReplyDelete
    Replies
    1. " it's already starting"

      No doubt! The question, though, is whether it's necessary and actually useful. As with so many other tools we employ, we can easily waste it or worse. That AI is coming is indisputable. Whether it becomes something useful is an open and highly debatable question. At the moment, military AI is a money pit that is returning nothing. Post coming!

      Delete
  17. AI never forgets. If it learns, that learning can loaded up in the next platforms computers, potentially instantly. I think the sifting and prioritizing data is a logical next step. Really its just a continuation of WWII code breaking, replacing people and crunching the data faster with a computer, just more difficult tasks. Plus it comes in handy when there aren't promising options. If up the middle is the only way its going to happen and high casualties would be a foregone conclusion, why not flip the switch in that helo and have it go it alone. Lets say I need an aircraft too small to be manned. AI then turns into the answer. I mean, if expensive stand off weapons are the future, is there a business case to put some landing gear on the missile and allow it to be refueled, then launched again? Think LRASM/JASSM crossong with a loyal wingman.

    ReplyDelete
  18. Hi CNO, avid follower here of your blog I wrote the following article in this month's "Proceedings" and cited a few of your works. Read my work here: https://www.docdroid.net/LDBAfBB/1-pdf

    Another recent good piece on the same issue: https://foreignpolicy.com/2020/07/06/us-marines-strategy-military-tip-spear-china-congress/

    ~ Ben

    ReplyDelete
    Replies
    1. Nice article, Ben. Well done! Now, we just have to hope someone in a position of influence reads it.

      Delete
    2. I responded to retired USMC colonel T.X. Hammes on a number of points in his article here: https://warontherocks.com/2020/04/building-a-marine-corps-for-every-contingency-clime-and-place/

      What do you think of his overall piece?

      ~ Ben

      Delete
    3. Great article Ben!!! Hope we see more of you here.

      Delete
  19. Assuming you'd been ordered to make it happen, how would you do it? Yes, this is silly but how about a narco sub with a parcel firing canon? Little parachutes instead of more expensive drones. Also, the boat would not have to negotiate coral reefs without a pilot. I'm thinking roto-moulded plastic, like a giant kayak. I think roto-molding is a 3d process so watching the 40m (possibly) mold in action could be fun but, once the mold had been paid for, the plastic narco subs would be cheap and probably wouldn't set off many mines. One a day out of the molds might be possible.

    Textron's ideas may be self-serving and daft but they might not be completely impractical.

    ReplyDelete
    Replies
    1. Okay, accepting that your idea is an example of out of the box, creative conceptual thinking rather than specific plans, I'd like to see you think through the actual type and volume of supplies that would have to be delivered to support a platoon+ size unit. The supplies would include things like fuel drums/bladders and lots of them, many days worth of supplies (meaning many pallets), large anti-ship missile reloads, tons of water (the Pacific is a hot region and requires lots of water to survive, and so on. With those requirements firmly in mind, now revisit your concept and see what needs to be revised to accommodate reality.

      I love the non-standard thinking but now temper it with a bit of reality in terms of the type and volume of supplies needed. Let me know what you come up with.

      Delete
  20. OK, the canon is out. So, the front needs to open so that cheap, wheeled vehicles can carry heavy objects up the beach. That means finding a way through any reefs. The guidance unit just got a lot more expensive. As I've attended a volunteers conference on a coral island, I should not have needed reminding about the water. I'm definitely thinking small, quiet, cheap and numerous for the narco subs so they could cross oceans, as smugglers have recently shown, but would be better carried closer to the platoon on a bigger ship. Obviously, the vessels would need enough buoyancy to carry more than one reload at a time but I'd keep them as small as possible so each one sunk is not too much of a loss. They might have to approach via more than one route to reduce the chances of someone lying in wait and, ideally, would linger long enough to recover the empty palette shifters. I'm hampered by not having a military background and must confess that I don't know how big a platoon is or how much their equipment weighs. If the 40m I suggested was adopted and the vessel had an average beam and draft of 2m, then the vessel plus supplies might weigh about 160 tonnes. The intention is for its deck to be flush with the surface of the ocean so it would probably have a similar density to water. I have no idea how much a plastic hull would weigh and I doubt, even at sea, that plastic could support as much as steel. Even so, the plastic narco sub might still get 10s of tonnes to the beach, particularly if the missions were organised by sneaky people because I am not proposing stealth.

    I plucked 40m pretty much out of the air. I have no idea of the limits on roto-molding.

    ReplyDelete
    Replies
    1. Okay, that's getting better! Now, setting aside any issues with the movement of supplies to the beach, how will the recipients unload the 'sub'? For example, the anti-ship missile reloads weigh around 3000-4000 lbs each. A platoon on a secret, forward base isn't going to have access to a large crane. How do they unload it?

      Delete
    2. (Don McCollor)...Perhaps use the 'sub' as the launcher built around the missile and launch tube. Let it lay doggo in a shallow protected inlet or swamp until needed, maybe sunk vertical...

      Delete
    3. How do you supply it with targeting data?

      Delete
    4. If the "sub" is cheap enough, they would become throwaways, that's basically how the narco kingpins are using them, right? One way transports. I would just beach the sub and have small explosives cut thru the hull and just drive what ever you unload. Problem is then your beach is full of broken sinking subs! Plus, not sure how you re-embark the troops and gear....which just hits me: all this talk of moving USMC to small islands, ok, we know getting onto the beach is pure BS but has anyone wondered how we getting USMC OFF THE BEACH?!?! Or we not telling the troops these are one way missions? Not that the troops would be that stupid not to realize the mission is FUBAR.

      I like the general idea of JOHN D. but I'm afraid this would be the kind of program that would just be a repeat of every other US DoD program: they can't keep anything KISS and would try to do everything in one platform, gold plate it to death and instead of a cheap, throwaway platform that could be made anywhere inside the USA, USN would turn it into another expensive flop with one prime contractor! Same with other drones and unmanned systems these 2 TEXTRON officials are touting: what could probably work is something cheap, throwaway and produced in vast quantities....that's the pretty much the opposite of any US weapon system.

      I don't really blame these 2 TEXTRON official or really any other contractors like LMT or RTN,etc that publish, I'm a lot more disgusted be Defense News, Breaking D, REUTERS,etc that let this stuff get published all the time and NEVER PUT A DISCLAIMER like you find on stock market articles when the author has a financial interest or is on the payroll. Better yet would even be for these publications to print a rebuttal or counter point, to just let these articles stand is a disservice to everyone.

      Delete
    5. The mechanism used for lifting vehicles in a garage can cope with the weight of an anti ship missile. It could also be customised and dismantled for shipping. However, I can’t envisage anyone carrying a large, anti ship missile over the kind of ridges I have seen in the first island chain in a covert manner. With any distance between the landing beach and the headland where the missiles might be set up, the Marines would have to commandeer jeepnies and drive.

      Can I be clear that my comments are solely about an automated means for getting equipment up to and across a beach. I am in no position to comment on the new vision for the American Marines. Given the length of the supply lines and what I have read on your website and at SNAFU, the island hopping tactics seem more like something the Chinese would do to America than vice versa.

      However, I am under the impression that the Belt and Road Initiative is not going well. One article described targets missed and growing resentment to the concessions given to the Chinese in Kazakhstan. The treatment of the Uigurs is also unlikely to encourage cooperation by Turkic neighbours. So, I cannot see the Chinese choosing to do anything which will hamper shipping into their main ports. America will have to start any war.

      Getting back to AI and unmanned transport, imagine a parallel universe in which the Argentinians were able to re-supply Port Stanley with semi-smart blockade runners. Would British troops still be outside trying to get in? Not if semi-smart re-supply was also available to them so each man didn’t have to yomp 1cwt across the Falklands.

      There will always be jobs AI can’t do but plenty more are ripe for automation. Getting supplies up to and across a beach could be one of them.

      Delete
    6. Sorry, ComNavOps, I misread your question. I was proposing the lift for reloading launchers.

      I don’t see the front blowing off the transport. If the plastic vessel can get to the beach once, it can probably do it twice, so re-use it. Sorry, Nico.

      Apologies out of the way, I visualised loads on palettes with each palette on a wheeled platform. The plastic fantastic beaches gently, the front opens, ro-ro style, and the wheeled platforms drive out in procession. The vulnerability at this point is why I previously mentioned sneakiness. For longer loads, join two or more of the platforms together. Once in the trees, many loads could be manhandled. Those which can’t would need the lift. If no one has realised re-supply is in progress, drive the wheeled platforms back up the ramp into the vessel and return to the main supply point.

      A platoon has as much chance as a recently hatched turtle if enemy forces have spotted them, with or without re-supply.

      Delete
    7. " the front opens, ro-ro style, and the wheeled platforms drive out in procession."

      Of course, there's always those nasty little bits of reality to deal with. For example, a single missile weighs around 4000 lbs (more with packaging!). Will a small, wheeled platform be able to move through wet beach sand and up an incline with a 4000 lb load?

      How do these small wheeled platforms negotiate their way through and over uncleared terrain, with heavy loads, and not tip over or get stuck? Do we have to build paved roads for them?

      These are the kind of details that so many conceptual ideas get lost on. I urge you to think through the entire resupply operation. You'll find a LOT of bits of reality that intrude on the concept. The challenge is to identify those bits and resolve them.

      Delete
    8. (Don McCollor)...Apologies, ComNavOps. The 'sinkable" sub ASM platform was only as an alternative to unloading and manhandling it ashore. Targeting information could be by radio or cable to shore (it is assumed that a team would be close by)...

      Delete
  21. I love reading the latest idiocies from the defense establishment. Each new toy is going to be the greatest thing since sliced bread. And all of our troubles will be gone if we only spend billions on this latest and greatest. From personal experience (that means I didn’t hear about in a think tank seminar sponsored by a defense industry contractor) the only that that matters is the ability to put ordinance on target. Everything else is BS

    ReplyDelete
  22. The sad part about what these 2 officials and others are proposing to replace humans brings to mind 2 small short SCI FI stories where after using up all the drones and unmanned gear to explore a new dangerous planet, you guessed it, they had to use humans that weren't trained or had the proper gear. The other story was the robots got so expensive, military went to using humans first then the robots....I could see US DoD one day using drones and other unmanned crap till it all gets shot up then sadly sending humans that don't have what it takes in terms of training and gear to do the job.

    ReplyDelete
    Replies
    1. "sending humans that don't have what it takes in terms of training and gear to do the job. "

      Now that's a great point and a very real possibility.

      Delete
  23. >we’re all familiar with the nonsensical fantasy of small units, deep in enemy territory

    Again, do you have a source for this? That the new plan is to operate within enemy territory rather than on friendly islands within the enemy A2/AD bubble?

    ReplyDelete
    Replies
    1. This is taken straight from the Commandant's own documents and statements. Go read them and stop wasting my time.

      Delete
    2. I too can't find that "deep in the enemy territory" in Commandant's documents. Here's the excerpt from his article which agrees with "small units" part, but disagrees with "deep in enemy territory" part:

      Marines and Sailors will have to uncover and develop solutions for the challenges of operating in the new modes our concepts suggest: in smaller units, on smaller ships, distributed over vast distances but linked by command and control systems and doctrines that allow
      such radically dispersed forces to achieve relevant, lethal effects in deterrence and in war.

      Delete
    3. Among many other documents, go read the Force Design 2030 which discusses placing units inside the enemy's "weapons engagement zone".

      Here's a quote from another article:
      "operate inside a contested maritime environment "

      Where do you think these weapons engagement zones and contested maritime environments are? Hint: they're not out around the periphery, they're INSIDE enemy territory.

      I can't believe I'm wasting my time on this! There will be no further comments on this.

      Delete
  24. Being on a receiving end of US military might in the past, I have a lot of respect for it, so it hard for me to comprehend the lows that it had reached.
    This approach of "let's invent the non-existing problem, so we can justify the solution, which happens to be our product" is resulting in goofy weapons systems.
    Why is US military unable to produce doctrine that has impact? If doctrine says something, and military is asking for 10 of tools to do that, how is it possible that then parliament ADDS money for 3 more?! This moronic procurement system is so toxic, it will eventually grind us military to a second rate fighting force. You see, it’s not about the money (United States prints it anyway), it’s about the personnel that is forced to do something that’s not helping in case of future conflict. Like LCS fiasco, which wasted generation of navy officers, having had to deal with “figuring how to make this work along with civilian contractors” instead of honing their seamanship skills.
    Until US military is able to be a real doctrine-based force, every awkward technology peddler will have its attempt at taxpayer’s money, and the outcome will eventually be that of USSR – a paper dragon.

    ReplyDelete
    Replies
    1. "Until US military is able to be a real doctrine-based force, every awkward technology peddler will have its attempt at taxpayer’s money, and the outcome will eventually be that of USSR – a paper dragon."

      I think that you unfortunately hit the nail on the head (hope that idiom has meaning in your home country, since your comments suggest you are from somewhere other than the USA). What are the problems we are apt to face?

      1) Peer war with China
      2) Peer war with Russia
      3) Threat from rogue nation or irregular terrorist actor(s)

      What is our doctrine to win each of those? I don't know if we have a doctrine to win 1) or 2), and the way we have been fighting 3) with our hands tied by absurd ROEs suggests that we don't know how to win that one either.

      Delete
    2. USA can't win war against Russia, and vice versa due to mutually assured destruction by the nuclear weapons. Probably could win against China, if allied with Russia, India and Japan at the same time. But that is pointless to debate.

      The point is to have military that is scalable (in case of all-out war), and affordable in peace. The doctrine should reflect this, and procurement too. Robot drones supplying warriors under fire are simply not scalable at this time. As a former soldier, i'd rather have mules than robots.

      The way i see US military (regardless of what it says in doctrinal documents) is that it's built to fight Vietnam/Iraq type of campaign. This means plenty of in-theater supply bases that are invulnerable to enemy, so complicated weapons can be maintained. It is not built to take punch, just to give one. Given that, it follows that US politicians seek fight with enemy that is compatible with that (Iraq/Serbia/Afghanistan/Libya/Syria...).
      While it's good for standing army to have experience in local conflicts, i think it's stupid to scale a lot of systems to that kind of conflicts.

      Delete
  25. I'm telling you, the LHA/LHD is the root cause of all this. The Marines are stuck with a platform from which conventional amphib assaults are not possible, so they are grabbing at straws to try to find a mission and method of accomplishing it. And of course, their friends from the industrial side of the military-industrial complex are all too happy to offer assistance.

    I worked on End Sweep, the mine clearance operation at the end of the Vietnam war. You wouldn't believe the hare-brained schemes that everybody and there brothers were trying to foist upon us. Right now the Marines are very vulnerable.

    ComNavOps, as you have correctly noted more than once, this comes from not having a CONOPS, so you latch onto whatever shiny object comes around. And you can't have a CONOPS without a grand strategy, and right now I don't see one. We may have our disagreements about the grand strategy, or the CONOPS that flows from it, but I think we both agree that you have to start with strategy, derive CONOPS to accomplish that strategy, and then develop systems to achieve that CONOPS. We're doing it backwards. We are spending billions on platforms that are cute technologically but don't fulfill any particular CONOPS (Fords, LCSs, LHAs/LHDs, Zumwalts), and then trying to derive a strategy from what those platforms can do. Since they can't do much, there isn't much of a strategy that can be derived.

    I've discussed my strategic concept and how to build a Navy and Marine Corps to execute it. I don't go the way the Navy is going, except I do include something similar to the FFG(X), though I frankly think the Italian version is overall better than the one we are building.

    ReplyDelete

Comments will be moderated for posts older than 7 days in order to reduce spam.