Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

RTS for Agents

getagentcraft.com

Post reply on HN
Print view View original

41–47 of 47 posts

Re: RTS for Agents

#41

Post by gs17 »

erikw wrote: post #38
jjmarr wrote: post #34

Earlier quoted context omitted.

The revealed preference of players is for terrible AI so games are easier. That's why AI has been going downhill. Payday 2 is my favourite example since they've had a bug since Day 1 that lobotomizes the AI subsystem. Specifically, there's a global cap of 1 action for all enemies per game tick, so when there's too many enemies the reaction time is 5 seconds. The mod Full Speed Swarm fixes this bug and the game is unp…

I don't know about this... I think that the "expert" AIs on RTS games just get more resources (not compute resources, but in-game resources) so they can create more units. I'd love to play against an AI component that has the same resource collection speed as the human player.

Yeah, cheating with resources or vision is disappointingly common. I remember the Green Tea AI for Starcraft 2 being pretty hard, but it looks like above "medium" it cheats (but the TL wiki does say "very easy", which will give you 10 minutes to set up, is equivalent to Blizzard's "harder").

Re: RTS for Agents

#42

Post by robotresearcher »

Youngsters may not know about Core War, a competitive game for agents that debuted in 1984, and it's predecessor Darwin from 1961.

The difference in compute scale between the Darwin winning program (44 machine instructions) and an agent today that calls into an LLM is rather mind boggling.

https://corewar.co.uk

https://en.wikipedia.org/wiki/Core_War

https://en.wikipedia.org/wiki/Darwin_(programming_game)

Re: RTS for Agents

#43

Post by margorczynski »

erikw wrote: post #38
jjmarr wrote: post #34

Earlier quoted context omitted.

The revealed preference of players is for terrible AI so games are easier. That's why AI has been going downhill. Payday 2 is my favourite example since they've had a bug since Day 1 that lobotomizes the AI subsystem. Specifically, there's a global cap of 1 action for all enemies per game tick, so when there's too many enemies the reaction time is 5 seconds. The mod Full Speed Swarm fixes this bug and the game is unp…

I don't know about this... I think that the "expert" AIs on RTS games just get more resources (not compute resources, but in-game resources) so they can create more units. I'd love to play against an AI component that has the same resource collection speed as the human player.

I wonder how strong such an agent would be today if years ago AlphaStar was on par with pros. All those fun projects (the one with Dota 2) died out when LLMs took the scene and RL died out.

Re: RTS for Agents

#44

Post by hombre_fatal »

jjmarr wrote: post #34
samsolomon wrote: post #3

Man, I could see the next Total Annihilation / Supreme Commander-like game where rough orders/instructions are given to commanders and they carry out those commands. The scale of those games was already nuts, but that would 10x things.

The revealed preference of players is for terrible AI so games are easier. That's why AI has been going downhill. Payday 2 is my favourite example since they've had a bug since Day 1 that lobotomizes the AI subsystem. Specifically, there's a global cap of 1 action for all enemies per game tick, so when there's too many enemies the reaction time is 5 seconds. The mod Full Speed Swarm fixes this bug and the game is unp…

Whether better AI is fun or not depends on (A) the sorts of advantages/disadvantages it is given and (B) its skill level compared to the player.

When people say they want better AI, they mean they want smarter + human-like AI without "cheating" (more resources, perfect aim, rubber banding, perfect knowledge, etc).

And even if you build a human-like AI, it's not going to be fun for players if it's too good, just like how humans have the most fun playing against other players at their level, not against the top-ranked players in the world.

So to keep AI fun, there is always going to be artificial limitations, and "people want terrible AI" doesn't really capture that.

One example might be games where you can hear footsteps or a branch snap. With stereo audio, you could technically derive exactly where someone is, possibly even which room they are in. When that raw signal is provided directly to a machine, it might accidentally be too good, know which bush you crawled into, and shoot into it.

Yet humans would only be able to go "I think I heard a branch snap" or "someone must be creeping up on me" even though they are given the same signal as the AI.

Re: RTS for Agents

#45

Post by 2001zhaozhao »

Finbel wrote: post #18
MattRix wrote: post #14

Earlier quoted context omitted.

I don’t know if this is a joke but isn’t the answer just… a regular RTS game?

One fun thing would be to raise the abstraction level of the game. How would it feel to only interact with your base/economy/army via prompting and face someone doing the same. Would words per minute replace APM, what would the meta look like etc. Would you be able to adjust the system prompt for your "army" to suit your play style?

Imagine being able to vibecode your units' attack and movement behavior during a match to dynamically counter enemy strategies.

Re: RTS for Agents

#46

Post by monster_truck »

samsolomon wrote: post #3

Man, I could see the next Total Annihilation / Supreme Commander-like game where rough orders/instructions are given to commanders and they carry out those commands. The scale of those games was already nuts, but that would 10x things.

Zero-K is worth checking out, its excellent. The best spring derivative imho

Re: RTS for Agents

#47

Post by Folcon »

squeaky-clean wrote: post #36
Folcon wrote: post #9

Earlier quoted context omitted.

Genuine question to you as someone who's been building in this space, what would you want / need to play a game as a programmer? An API? An SDK? An in-game editor? Tutorials? Or is this more a "I want a factorio-like"? I've been building economic engines and simulations for the last few years now and over the last 3 months in my off time I've been getting increasingly in the weeds about how to design a game that fits…

Not OP but for me the appeal would be in botting a game that's made for humans, and not a game made specifically made for botting. Let me use autohotkey, or modify the client, or do any of the things real botters do, don't worry about providing an API. Just let me register a character as a bot and restrict it to a separate server that can't interact with the main game servers/economy, and disable anti-botting protect…

This is interesting, I've been thinking a bit about this question of separating some gameplay areas

There's sort of 5 versions of this in my mind and I've been thinking about each of them and trying to work out which one is a good starting point of to explore :)

I'm going to say your zombies as a shorthand for how you automate stuff, hopefully it creates easier to follow examples :)

1) automation of player spaces where players complete by automating and they can't directly influence each other, so something like a zombie workshop a la zachtronics game with some twists

2) automation of a shared space automation space, so more zombies dueling each other like in robot wars

3) automation in the game world that's multiplayer interactive, so it can influence the game and help you, like factorio, however players can interact with it and say damage or improve the mechanisms of automation

So I can tell my zombies what to do and help them directly and our zombies can fight each other and I can fight your zombies

4) automation in the world that's non-interactive, not sure of any examples of this, but you automate things that manipulate the world for the player in a way that other players can't directly influence, so imagine I set down instructions, stuff happens in the world, but we can affect the world and each other, but not the mechanism by which we automate

So I can tell my zombies what to do and I can't help them directly and our zombies cannot fight each other and I cannot fight your zombies

5) automation in the world that's player controlled, and indirectly interactive, this is basically 2 but it effects the game world, however players can't affect each others mechanisms of automation like in robot wars, but their mechanisms of interaction can affect each other

So I can tell my zombies what to do and I can't help them directly and I can't fight your zombies, however our zombies can fight each other

Post reply on HN