But seriously, my theory on the absolute state they left tactical vehicular AI is that they have a plan laid out to build a new framework that will now and forever allow superb AI, modular, customizable, madly detailed. It’s just not quite done yet.
But while you’re building the ritz, would you invest manhours into laboriously patching up the worn through bits of the rug in your rickety old hotel? No, fork it, let it rot for a couple weeks more, we got the ritz in… two weeks. Right? right.
To paraphrase Mr Spock, when it comes to DCS AI there are two possibilities:
They are unwilling to alter it
They are unable to alter it
The first would include insufficient resources to handle it plus other priorities, so they have pushed it to the back burner/worked on it very slowly while prioritizing “revenue generating” coding.
The second would include legacy code that is intertwined to the extent that fixing it will break other things, perhaps catastrophically, so they have to find work arounds that are little more than a bandage and not the surgery needed.
I don’t know which of the two possibilities it is, although I infer it is mostly the first with a bit of the 2nd mixed in (ie it’s so deeply embedded that it will be a LOT of work to correct it without adverse effects and they haven’t been able to devote enough to it).
Regardless, if it was easy it would’ve happened long ago.
THIS.
It still surprises me that the base sim is free. A lot of people complain that they focus on areas that create revenue which should be no great surprise. At the end of the day, you need revenue to keep the lights on.
I would charge for the base sim if I was king, and put the generated funds back into development.
I haven’t looked into this at all, but what came to mind was, it’ll be interesting to see if the large language model AIs will play a part - you’d think you could use them to come up with fluid narratives in dynamic, evolving situations - say a close air support situation, where the AI was able to come up with the live point of view of the forward air controller / infantry guy under fire. Could be a great immersion assist.
As long as I don’t need another GPU to run both the graphics and the AI…
Actually I take that back.
I’m already running two PCs to record and occasionally stream my flights, and I’ve been thinking about running SP missions on an MP server on the streaming PC to reduce the load on the gaming PC… if I can get a better experience with an old (but still CUDA enabled) GPU running in another PC then I suppose that makes sense!
Nobody said this was a cheap hobby 
That is also my impression on ED’s general policy of operations. They always have this grand plans for total revamps somewhen in an undefined future so refrain from any practical improvements right now.
I forget when DCS World itself became free to play with Su-25T and TF-51D, but it wasn’t at first.
It was after the release of FC3, which would have included the 25T, and the P-51D.
They did it to increase the reach of the game instead of making some demo that would take even more work. As the TF-51 only lets you sight-see and the 25T is 95% air to ground, and fairly slow, they’re hardly the best advertisement for what DCS can do. However, the F-15C and MiG-29 are only $15 each or whatever so people can spend fairly little to try it out.
You forget anyone can free try almost every module for two weeks. Hardly anything better than that.
The Su-25T can be a bit buggy, and is weird with its Russian avionics to boot. The TF-51D is fine for introducing people to warbirds, but I feel the FC3 A-10A would have been a better choice for the free base module.
Hmm. There should be a “$5 starter pack” with your choice of one of the FC3 aircraft (OK, it’d probably be $20 these days) as a way to get people into the game. That way they can start with a jet that they actually want. If they had that, I wouldn’t complain about the base game costing money … but that’s probably not enough money to do significant development off the back of it …
The two main flaws I see with the 2-week trial are:
It takes too long to be applied on newer releases. Customers may find interest in DCS over a recently released module, but because it’s not available as a trial they’re forced into either dumping a lot of money into something they aren’t sure they’re going to like, or waiting x-months until it comes to trial, by which point they’ve probably forgotten about it and ED lost a potential customer.
It takes longer than 2 weeks for newbies to learn basic flying much less anything else unless they’re unemployed students with oodles of free time and money. 2 weeks is only valid evaluation time for people who already have considerable experience with sims and/or DCS.
I would say the 2 week trial is ok for a terrain, but for anything you fly it only works if you’re a DCS vet who knows what it takes and has time off from work to devote to it! 
I personally used the free trials only back in 2020 when ED had the COVID event and I tried the JF-17. I liked it, but not enough to pay the price for it. It also has had the weakest sales of all DCS modules to date, so it has yet to drop to the range where I would be willing to buy it.
I think many of you mistake the purpose of the two weeks trial.
It’s not to gain proficiency- but just to give you enough to whet your appetite for it.
They give you a free bite so you want the whole thing OR just enough time to find out it really doesn’t do it for you.
That could work for something simpler, but I don’t think 2 weeks is even enough time to learn how to take off in an AH-64! How are you supposed to know if it’s worth the money to enjoy the sneaking and popping up and taking out armor and then sneaking away again if you’re continuously crashing inside the confines of your base?
Of course, if you’re someone who can dedicate 2-3 hrs per day to it over those 2 weeks, maybe you can. I only get 2-3 hrs per week for DCS so my brain has a hard time contemplating how it could work for those with more time to devote to it.
I honestly disagree on basically the whole line but… You know, opinions and all that.
Appetites can’t be whet with zero proficiency. To get to a level of proficiency where it can be whet requires more than two weeks.
Ideally you’d have quickstart missions already in the air and near the combat zone for people to have a play with… but I hear you on the basic proficiency issue. They perhaps need a helo in the base game as well?
I’ve long been telling people the first thing they need to get is the Community A-4, and now I’m wondering whether I should also recommend either the Blackhawk or Loach as well?
I tried the Two Week trial only once, a ways back to see if the reported issues had been fixed with the F-86 at the time. They had not, so no purchase (they may have now).
No clue what their market research tells them about the two weeks thing.