Archived

This topic is now archived and is closed to further replies.

wyrd

What sort of AI demos would "wow" an employer?

Recommended Posts

It''s become aparent to me that the best way to get a game programming job is to concentrate on a specific area. I haven''t done much AI programming, but it certainly does appeal to me, far more than any other area. So I was wondering, what should I be working my ways towards in terms of a demo that would wow an employer?

Share this post


Link to post
Share on other sites
Flocking behavior of 1000''s of agents though a complex obstacle course. Coordinated behaviors of a squad of units toward a goal ( using as using overwatch, rear guard, and point ). Pathfinding of 1000''s of units in realtime across some complex terrain. You might want to combine the flocking with the pathfinding.

Those should be impressive enough. AI is an enormous field, but for games, mostly its state machines, flocking, pathfinding, and agent-agent communciation.

-ddn

Share this post


Link to post
Share on other sites
Why don''t you automate a virtual agent''s ability to scratch their ass? Oh sorry, just continuing the theme of waste my time posts.

Share this post


Link to post
Share on other sites
"Why don''t you automate a virtual agent''s ability to scratch their ass? Oh sorry, just continuing the theme of waste my time posts."

You mean like in farcry? lol i bet they thought that was hilarious to do.

Share this post


Link to post
Share on other sites
Guest Anonymous Poster
Try writting a small non graphics intensive rts. That way you can focus on your AI and show something complete.

Share this post


Link to post
Share on other sites
quote:
Original post by wyrd
How could you make a non-graphical RTS?


non-graphical _intensive_. he didn''t say without graphics, just without a focus on high quality graphics.

-me

Share this post


Link to post
Share on other sites
speaking of bots, do the most advanced lifelike bots (i.e. Counter Strike Condition Zero) use waypoints anymore? Or do they examine the world geometry in real time?

Share this post


Link to post
Share on other sites
Most games don''t look at the actual geometry, as the actual world geometry has much more complexity the AI needs. Additionally, it typically is optimized for rendering, not for AI use.

Several games do use ''nav meshes'' which are either generated automatically or is manually placed by a level designer. Waypoints are still very common, and are much simpler if the game can get away with them.

Share this post


Link to post
Share on other sites
The bots in UT2K4 are pretty impressive... i''m not sure how they work exactly (i.e. with waypoints or not) but i''m pretty sure they will work on any user created map. So if it does use waypoints they will be generated automatically...

I think if you want to make a bot, forcing the user to create waypoints for custom maps is not a good idea. Most people won''t have the patience to spend hours trying understand how to create the waypoints, they just want to play. Its also more impressive if they work on any map without any user guidance, that doesn''t mean automatically generated waypoints are not good.

Share this post


Link to post
Share on other sites
In my opinion the low-graphics RTS idea is the best way to show off an AI. Bots are good, but AI shines when you put it in an RTS.

I didn't know it was that easy to "wow" an employer...here I am trying to build an RTS with character based AI that causes units to order each other around and follow general orders that the unit decides how to follow(all in QBasic because I don't know C), and now I find that all I need to do is make a basic RTS AI.
Well, maybe not basic, but still...

Oh well, I like the project so much I think I shall finish it.

"The fell winds blow,
The darkness grows,
And I, I stand alone."

[edited by - KeylanFarr on April 7, 2004 5:38:17 PM]

Share this post


Link to post
Share on other sites
quote:
Original post by wyrd
It's become apparent to me that the best way to get a game programming job is to concentrate on a specific area. I haven't done much AI programming, but it certainly does appeal to me, far more than any other area. So I was wondering, what should I be working my ways toward in terms of a demo that would wow an employer?


You haven't done much AI programming, but you'd like to do it? Seems it would be best to do a some learning before trying to discover what would impress an employer NOW.

I did some kind of AI research for 7 years, with interrupts of course, and I don't think I would be able to impress employer. (The funny thing is you'd discover if that AI programmer was effective or not at the end of project, when it would be late)
You should decide if you'd like to learn a Strong AI, or a weak AI. There isn't market for the strong AI. Would you really think that University would try to use program that could remove 19/20 its researchers from its jobs? (AI for chiefs is much easier than AI for men under. So extremely low demand for it)
Or you could learn the weak AI. So you'd do web searching application, molecule binding research applications, path searching applications for bus routes. There are a really crappy jobs for weak AI programmers.


If you would like to impress employer, know well his sister, or be his brother. Even if you'd be psycho he might be able to get you in.


[edited by - raghar on April 8, 2004 3:02:23 PM]

Share this post


Link to post
Share on other sites
quote:
Original post by Raghar
quote:
Original post by wyrd
It''s become aparent to me that the best way to get a game programming job is to concentrate on a specific area. I haven''t done much AI programming, but it certainly does appeal to me, far more than any other area. So I was wondering, what should I be working my ways towards in terms of a demo that would wow an employer?


You haven''t done much AI programming, but you''d like to do it? Seems it would be best to do a some learning before trying to discover what would impress an employer NOW.

I did some kind of AI research for 7 years, with interrupts of course, and I don''t think I would be able to impress empoyer. (The funny thing is you''d discover if that AI programmer was effective or not at the end of project, when it would be late)
You should decide if you''d like to learn a Strong AI, or a weak AI. There isn''t marked for the strong AI. Would you really think that University would try to use program that could remove 19/20 its researchers from its jobs? (AI for chiefs is much easier than AI for men under. So extremely low demand for it)
Or you could learn the weak AI. So you''d do web searching application, molecule binding research applications, path searching applications for bus routes. There are a really crapy jobs for weak AI programmers.


If you would like to impress employer, know well his sister, or be his brother. Even if you''d be psycho he might be able to get you in.



What''s with the fatalistic attitude? So you have to sleep with an employer''s sister, or be buddies with his brother to get in... What do you mean by saying a strong AI has no ''marked'' (im guessing you mean market, but it still makes no sense)? And please do tell me where youve done research - im applying to grad schools now and I need to find out WHERE NOT TO GO!!

Share this post


Link to post
Share on other sites
quote:
Original post by SpaceDude
The bots in UT2K4 are pretty impressive... i''m not sure how they work exactly (i.e. with waypoints or not) but i''m pretty sure they will work on any user created map. So if it does use waypoints they will be generated automatically...[...]
In correct. You not only have to place waypoints for custom maps in Unreal Tournament 2004, but you have to specify the type of waypoint as well (regular bot path, ground vehicle path, flying vehicle path, ground bot path for use when ''extra jumping ability'' is abilable, jump pad path, etc).

Not only that, but you have to do a simmilar thing for rendering, telling it where zones end (like doorways, where the world on the opposite side doesn''t need to be drawn if you can''t see the doorway itself) and what geometry should occlude other geometry.

It works, but its a PITA

Share this post


Link to post
Share on other sites
quote:

You haven''t done much AI programming, but you''d like to do it? Seems it would be best to do a some learning before trying to discover what would impress an employer NOW.



Am I not entitled to show interest in a specific gaming field without being an expert in it? I''m interested in knowing what employers like to see so I have a goal to work towards. If I can "wow" an employer to show that I have potential to grow, then I''ve succeeded in my task.

quote:

I did some kind of AI research for 7 years, with interrupts of course, and I don''t think I would be able to impress empoyer.



Surely after 7 years in ANY field you''d learn some neat tricks to impress people, including employers.

Share this post


Link to post
Share on other sites
quote:
Original post by Raghar
You haven''t done much AI programming, but you''d like to do it? Seems it would be best to do a some learning before trying to discover what would impress an employer NOW.


"You can''t get a job without experience...
yet you can''t get experience without a job."?

-Predictor
http://will.dwinnell.com


Share this post


Link to post
Share on other sites
I''m no employer, but anything self-controlled would be impressive AI to me. For example, guards reacting to an assault, units self-defending themselves in battle, units & vehicles navigating around obstacles without getting stuck or trapped, a bot that is capable of using cover, or issuing legit orders, etc... Since it is geared for the game industry, the better graphics you use the more prestige I think it will earn. Simply because it shows that even in intense graphical situations, it still performs well. Plus, if the employer is a typical run of the mill HR person, you’ll want to keep it as idiot proof as possible (ie: no console demos!!).

Share this post


Link to post
Share on other sites
quote:

Since it is geared for the game industry, the better graphics you use the more prestige I think it will earn. Simply because it shows that even in intense graphical situations, it still performs well.



That''s what timers and benchmarks are used for. And unless you know a graphical artist and 3D modeler who can provide my demos with awsome graphics for free, it''s not going to happen.

Share this post


Link to post
Share on other sites