Update 6.001: Odessa and Leningrad Early Access

[kraut1]

For your information: EMG Odessa Modification v1 is completed

https://forum.il2sturmovik.com/topic/91755-emg-odessa-modification-v1-update-13062025/

@LukeFF

[kraut1]

 

[kraut1]

current status of new EMG front lines:

(targets for new front lines have to be defined after last front is completed)

to do: early 1941 front lines 02.07.41-25.07.41

[kraut1]

https://forum.il2sturmovik.com/topic/91755-emg-odessa-modification-v2-update-04082025/#findComment-1356948

 

EMG Odessa Modification v2 Update 04.08.2025

 

Update v02 04.08.25 NEW

2.) EMG Modification Odessa 1941-1944 :

with many new 1941, 1944 Full-Map-Front Lines 

Features:

-modified 1941 Odessa Generic template

   -with modified Ship convoys with russian destroyers

   -additional torpedo boat = gun boat units

     -changed patrol ships (destroyers, gun boats, torpedo boats as patrol ships)

   -1941 tank colums / tank concentrations (new, activation with batch file)

   -1944 tank colums / tank concentrations (new, activation with batch file)

    -random activated ambient flights to have a variation for each mission start

   -ambient flights activated in increased distance

   -ai artillery opens fire only when enemy units are nearby

   -5 optional Odessa map templates, to be activated with batch files for optimized mission generating time

   

[Ace_Pilto]

Where does the content spam end and the AI improvement begin? Every single branch of information technology is obsessed with AI right now, but not a single simulator developer pays the slightest of interest. Strange to be caught so far behind the curve.

[AndyJWest]

2 hours ago, Ace_Pilto said:

Where does the content spam end and the AI improvement begin? Every single branch of information technology is obsessed with AI right now, but not a single simulator developer pays the slightest of interest. Strange to be caught so far behind the curve.

They are ahead of it. The obsession with AI is driven almost entirely by marketing hype, like every other fad (blockchains, the dotcom boom etc) that IT has gone through. The current 'AI' fad - the LLM (ChatGPT etc) - is nothing but a glorified next-word-predictor. It isn't 'intelligent'. When it looks that way, it is because it is being asked to do something for which it has multiple examples to plagiarise and mash together amongst the humongous dubiously-sourced dataset it has been trained on. When it doesn't have anything to plagiarise, it invents. It makes stuff up that looks superficially convincing, but is unambiguously fake ('hallucination'). There is no 'aircraft simulator' dataset to train LLMs on. And simulation requires deep knowledge the thing being simulated. Actual knowledge, not random factoids dredged from Reddit or Wikipedia. Simulator developers aren't using AI because they know it is useless to them. 

[Aapje]

@AndyJWest

 

The problem with your argument is that it can just as easily be applied to people, who also constantly 'plagiarise' others and make up stuff that is wrong, but that sounds good to them.

 

Note that the entire question of what counts as intelligence, is one that is extremely subjective and subject to redefinition. For example, once many people thought that chess requires advanced intelligence and that a computer that could play chess, would beat humans at any intellectual endeavor. Yet when computer programs were made that could beat people at chess and not do much else, that idea was abandoned.

 

Quote

And simulation requires deep knowledge the thing being simulated. Actual knowledge, not random factoids dredged from Reddit or Wikipedia. Simulator developers aren't using AI because they know it is useless to them. 

 

This is just you taking sides in the classic intentional intelligence vs emergent intelligence debate.

 

Ultimately, the proof of the pudding is in the eating, and LLMs are way more capable than what the intentional intelligence side has been able to produce.

[LukeFF]

And let's make that the last off-topic post in this discussion.