Calin said:
well for starters you could emulate the multiplayer experience in games. In FPS games bots are pretty good at imitating a human player however RTS games don`t have anything close to a realistic human like bot.

(… phew, it was hard work to tickle that out out finally)

Sounds like a specific problem and goal. So why don't you start a thread about just this? You would get some real arguments or proposals from people that worked on things like that. Probably nobody here ever tried to simulate a human (or animal) brain, so all you get is broad discussion of general AI limitations and possibilities in the best case, which leads to nowhere.

Did you follow Alphastar AI beating more than 99 percent of human players in Starcraft? Is there a match to watch, and does this AI feel like a human opponent?
If so, you could start to research how it works, which may be a task for years but it's a plan.

Do you know how RTS AI works in general? There surely is path finding, but what else? I assume it's a lot of combinatorial problems like this (which is no easy stuff): https://en.wikipedia.org/wiki/Weapon_target_assignment_problem

And finally, why exactly falls RTS AI short in comparison to humans? Is it just junky paths the AI units take? Is it that behavior in certain situations feels illogical? What are such situations?
If you could list some examples, a person like me could make some proposals for inspiration, or tell about technical reasons so defining the problem better, etc.

making a terminator or making a game AI that could match a human no problem boils down to a common problem how do you make something with exhaustive thinking.

Calin said:
making a terminator or making a game AI that could match a human no problem boils down to a common problem how do you make something with exhaustive thinking.

Sure, but we do not know how to make something that thinks.

You can try to do it and fail, or you can focus on smaller problems like RTS AI, solve them, learn new things and put them into your bag of tools.
With a larger bag of tools, you may eventually be able to generalize some tools so they are applicable to solve more problems.
Also, you will encounter many challenging but solvable problems on your way, and you will see there is more than enough work to do on them. You will see this has progress and value, and maybe, in the future we might be able to make thinking machines building on top of all the things we have done and learned.

In contrast, just sitting there and dreaming of general intelligence applicable to RTS games without any clue about how this could work will bring you nowhere.
Your proposal is to solve the much harder and abstract problem of AGI just for the application in a RTS game, although we can create game AI already and it works. (Ofc. there's always room for improvements, but you do not care about this room.)

From the outside, there can be only two reasons for such unrealistic proposal:
* Knowing nothing about how games work, you assume game developers to be unable to do it 'proper and finally right', ignoring that nobody (including yourself) knows HOW to do so.
* You are unable or too lazy to learn how current RTS games work, and to distract yourself from admitting this, you come up with fantasies for justification.

It appears naive and childish, tbh. You do not know where you aim to, but without any doubt you aim much too high.

JoeJ said:
Sure, but we do not know how to make something that thinks.

problem solving = thinking

1+1 = 2 is just a small thinking process

we can break this barrier of sound called human thinking

Calin said:
1+1 = 2 is just a small thinking process

No it's not. Similar to plants solving to find paths through a labyrinth is no intelligence either. Draining dumb fluid into it can do it too.

Ability to see two stones and than counting their number is neither problem detection nor solving.

But keep dreaming. Nothing wrong with that. While working on your RTS game you will see i'm right anyways ; )

Any life is intelligence. If it has parts that move following an algorithm it`s intelligence.

counting might seem trivial but it`s complicated to your mind: move numbers to a register, add the two registers etc.

Well, maybe any matter is intelligence, because it arranges in ways so life can evolve… just to end up at creatures that fail to see ability to make calculators ≠ ability to make thinking machines.

We are able to build machines that can store and replicate our ‘thoughts’, e.g. how we do an addition. But because we do not know how our own thinking works, we can not make machines to think on their own.
And we did not yet come up with an alternative thinking process (not inspired by our own workings) either.

Deal with it, or work on it.
If you would work on it, you would not ask 'does it make sense to study the brain?'. You would ask ‘how does the brain work?’, or ‘how does thinking work?’.
But you know there are no answers, so you dodge asking the proper questions for whatever reason.

I don`t have all the answers but there are things thought to be impossible that are actually within reach.

JoeJ said:
Ability to see two stones and than counting their number is neither problem detection nor solving.

https://en.wikipedia.org/wiki/Bird_intelligence#Studies

If we include ability to count to our loose definition of intelligence does not matter - counting won't help us to improve current RTS AI.

I'm out. It's hopeless. During this thread, we brought up quite a few topics that would be valuable to discuss in more detail. (Your slime molds being one of them because it's related to flow fields for movement and path finding, just to say.)

But there is no interest. Discussion always falls back to broad definitions, oversimplifications, wishful thinking, fiction, and other plays of thought.