(08-29-2025, 08:02 AM)GK3 Wrote: (08-28-2025, 11:33 PM)BostonCard Wrote: Are you saying that there may be some sort of “fog” that descends during war that prevents us from knowing everything that is going on, even in our modern day with cameras everywhere? ;)
BC
A good question, but I would say, yes the "fog" of war still exisits, maybe in a mutated form. Homo Sapiens are still involved and the old addage "stuff happens" still applies. The day may come when all wars are fought by drones and robots, but I do not think we are there yet. The big question that has yet been asked or answered will be are we better off running a confilict by Homo Sapiens or simply turning the whole thing to an AI which is of course designed by Homo Sapiens. There are so many variables it is almost impossible to answer this question. Terrain, weather, ground cover, enemy capabilities all play a role in what you decide to do and how you decide to do it.
I keep wondering how it might have been back in the day in the jungles of SE Asia if we had the tech then we have now.
IMO the fog of war that descends on the general public always has been worse than that of the military that is actually running the war. Before WW I the general public knew almost nothing about what was happening in the war. Communication was so slow it might be weeks before reports of major battles were publicly available. These reports were influenced by propaganda and by the fact the few reporters had a very limited view of the action. It was only what they could physically see. Virtually never would the mauling of a 6 gun artillery battery make the papers. We see such incidents on X all the time now. By WW I and certainly by WW II the reporting was much more timely. However, it was still heavily censored/propgandized and still limited by what the reporter could actually see for themselves. Of course the major military commands were by that time making regular press releases that delivered the "big picture", but most of these reports were heavily slanted (MacArthur) or total fiction (Goebbels et al).
Jumping forward to today we now get first hand reports with drone video of many, many individual actions. There are also many more "professional" reporters out there reporting on what they see. These reports are not "censored", but they are often misleading, either intentionally or unintentionally. Of course, the combatant military units release their own reports and videos that provide even more data. We also have groups like the ISW that attempt to sort this situation out. IMO the civilian fog of war indeed still exists. In many ways it is worse than ever. It is TMI situation that we can't in most cases properly evaluate or understand. Confirmation bias runs amok. I know, I have been guilty of it myself :-). "I expect X, so this video shows X". How well the various state actors and their military actually understands the "real" situation is a subject into which we have virtually no visibility.
IMO John Keegan's book Intelligence in War still has the situation exactly right, particularly Chapter 5 about Crete. It is important also IMO to remember that the course of the Franco-Prussian war was charted mostly by commanders who in different ways did not follow their orders exactly. Part of that was Auftragstaktik by individual commanders, but the key part was major commands moving forward quicker than the plan called for, resulting in the world's largest meeting engagement. The US civil war battle of Antietam was made possible by a major intelligence coup, but was indecisive because "stuff happens". These results are important for the use of AI in war.
At present, AI is "trained". It will always produced the "school" solution. This means AI is fundamentally incapable of executing Auftragstaktik unless it is very similar to a situation it has been trained upon, actually a "school solution". Against an intellectually agile opponent AI is at a tactical disadvantage. If and when AI achieves "General Intelligence" this limitation can be mitigated. At present, it would be possible to observe the opponent's previous responses to a certain threat and know exactly and reliably what your AI opponent will do. That can be fatally exploited. It is also unclear how and to what extent AI hierarchies can function. AI can only evaluate the data it has in the context of its training. If some of the "vital" data is missing, a situation that is common in war, no telling what the AI will do. The Borg solution requires lots of communication where the various parts of the entity "knows" a totality of what is transpiring. At present, such technology is well beyond the state of the art. The AI 11Bs available in the near future won't know much and are incapable of "figuring out" what to do and especially deciding to do it. Nobody is going to stand up and say "follow me" :-). If one did, none would follow because it isn't in their training set :-).
Therefore, IMO the most important use of AI in the near future is in intelligent weapons, specifically anti-drone weapons. It is a "detect, recognize, attack, pursue and destroy". For that, the school solution is fine. We have had wake-following weapons that could recognize specific ships and specific ship types for a long time. Super AI anti-drone weapons will IMO be in general use very soon.