it wont work

ez for ID Software, even for me. They just need a real player who say them quaker priorities on map, and when bot behavior not natural

We can get good training
Id can use this bot for their future titles
We can configure bot for different players behavior
We can grab movements from pro players demos and simulate them in game
We have ELO and we can configure everything

They can't get the netcode right, and you expect them to develop an AI system no game has ever had before.

Netcode was good in april, as i can judge. Later they just change some animation, for better feelings, so you feel like it's better, but in real it's just simulation. No one in game industry can won ping.

Also they made some netcode patches, but it's normal, coz game in development stage, you all just forget it everytime when you posting your whine shit

You're missing the point.

i'm not missing anything, you just not realize how it works

do YOU realize how it works? this is no simple feat.

Net code in QL already +-
Spiterbot done already

Where i'm wrong??????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????

"Ideally no one can detect it's bot or human "
spiterbot doesnt look very human.
what would actually be very awesome would be something like this:
https://www.youtube.com/watch?v=92tn67YDXg0
then again, its pretty hard to accomplish.
https://blog.openai.com/more-on-dota-2/

It's hard only in YOUR mind. Because you don't know how it works. Im a programmer and know what im talking about.

Spiter not looks like human, ok, then, ask yourself why.. and start to improve him. That's it, very simple.

xD

"ok"

American boy can does it all

then go ahead and sell the finished code to id

Spiter bot had godly strafe jumps, the aim was ok. Hence the modded side of things.

The maps and abilities are too complex to create an AI that will master the quake mechanics.

If you want to see god like AI

check out elon musks open AI that learned how to beat the top dota 2 players. If they could develop a self learning bot then yes, that would work.

We can get perfect bot without self-learning. I'm not talking about neural network.

We can make data tables for moving around map(like animation tables), we can copy movement traffic from players demos. And bot will choose. Something like this...

That would not work. There are too many variables. How will the bot determine what to do? It's not just movement and aim. It's decisionmaking. I have worked alot with machine learning and the only way to makr something close to human would be a deep neural network.

If you know game and how pro player thinks during game you can build algorithm, and i expect that cpu usage would be minimal, may be pentium 2 =), but a lot memory usage, may be 512-1gb of ram. Some variants(what to do) in ram.

Better would be to build a neural network based on the hundreds of thousands of demos out there of human players. Then set up it up on a public server 24/7 and let it learn by playing against human players. It'd probably attract a lot of players to queue up to play against a sophisticated AI bot 1v1.

what we need is more lootboxes and more attachments

xD

the water is bethesda money

Pay people to play QC

they're already doing that but they need to do it on a bigger scale

They could do it even without your faulty human gameplay to train the AI.

Alphago AI.
Perfect

Not really, that's simply a single-purpose neural net. It can't do anything else besides play Go.

Not an impressive feat when you consider HOW it became so good - playing more games of Go then an army of professional players.

ofc its an impressive feat. deep learning has been a revolution

neural network development costs a lot, noone will invest that money into quake ... also the neural network will at some point learn to aim "perfect", so it is already unfair ... developers will have to manually insert some aim-handycap, not sure how this will work and whatever is done will always be subject to controversy ... and also with a handicap its not really a man machine match on the same level anymore, so no real interest will arise ...
you will need a person to create again a bot in old school style with heuristics and a lot of manually programmed routines, but I think ids bot creator guy (Johannes Marinus Paulus van Waveren) died in 2017, not sure if they have a replacement for him, but they really should have one already for a long time ... and I also hope there will be some considerable improvement compared to the bots from 1999 .... :-)

I'm not talking about neural network. AIM handicap - no problem, easiest part . Not a man, but you can't feel it. ELO can be configured and we will get 800-3500 bots very easy. You can use bot for everything, just turn on your imagination :) You can change his aim and playstyle. Playstyle can be configured very easy by 20+ trackbars.

Main part - routes on map and time for each route(like GPS in car) - data table. After that configurable system of his priorities.

ok, a classic bot with routes and skill level, lets wait and see what id brings to the table ... but my guess is a bot is pretty low on their priority list and they dont have the expertise to build one either ... and since there is no modding a bot in qc probably is just wishful thinking

Yea, sad news about MrE. He left id to join Oculus in 2014 though.

There was another bot author who joined id on the dev team IIRC, although right now can't remember who, and obviously no idea if he was working with AI in any way or just was someone who happened to be familiar with how id's games are put together.

FWIW, https://jobs.zenimax.com/requisitions/view/1317 - probably doesn't relate to QC though and may well be talking single-player AI.

Artificial retardation

I remeber being impressed by spiterbots strafejumping around so fast. Of course, it wasn't a natural movement, didn't look natural, and was too fast for begginer players, but still infinitely better than what we have more than 10 years later.

Slap on it some weapon specific dodging patterns, limit the aim to 30-35 lg and 40/60 rg, and you're good to go.

But as always, Id noobs can't produce what a bunch of unpaid modders did 15 years ago in their free time, even with mommy bethesda pumping moneys in their asses.

The cpm bots are pretty good imo.

Its coming, just as any described it, and we are very close... Its obvious to anyone who is observant and knows how the internet worked in the past years, AI has swarmed the internet.

QC actually needs bots and campaign like q3 had for keeping the new players.

after they complete the campaign, they will have invest enough time to like and care about the game and will be able to endure the raping vs human players

IMO id have no incentive to include bots (and certainly not great bots which require a lot of work), they want you on the servers, playing other players and bolstering the community in a visible fashion.

Not locked away in an offline game, bots are not see as that important these days to virtually everyone having access to the net.
And good bots in an industry which is more interested in falsely making the players feel good for sucking, just goes against what a worthwhile FPS bot is about for all but the purists..

you hear a lot of stuff about how these google AI companies are doing amazing things with AI, but I doubt their billion dollar AI's could beat me at any traditional FPS without cheating and you have to consider how smart they really are, outside the realm of brute force.

So, your vote against progress. Well done

This seems like a solution that AI would propose

For offline play yes, giev!

For online play, no. The whole point is to play against human players? What good will bots do in this context, fill up the empty team game spaces? That sounds really boring already.


Anyways, on the base topic of AI and what it could have in store for us, there is this introduction for anyone interested. It's an entertaining read
https://waitbutwhy.com/2015/01/artificial-intelligence-revolution-1.html

Smart bot only for 1x1 offline and online.

TDM is more about fucking with your teammates.

they wont tell you you are playing a bot, thats the whole point of AI : )

just like they dont tell you reddit, youtube, 4chan are full of bots already having normal conversations with people

You're way off on your moderation of AI knowledge on those types of platforms.

im aware AI right now is more of a tool that expands the power of holding a narrative, making the process much easier.. other side is analyzing everything. the interaction part i mentioned will get more important with time as it gets connected into a continuous feedback loop with all the other elements

what do you think?

From a conversational point of view A.I would have to be able to understand the context and the emotional nuances of the English language. You'd have to have some massive neural network of monitor live chat where it could potentially learn how we communicate.

(Yeah I have no doubt that we'd see something like that in the future.)

Now having A.I respond to certain key phrases, that's already out there. A lot of the AI also respond in email marketing campaigns, there are bots you can pre-program on certain social media platforms that do respond to certain words etc. So it gives the general moderation of things a lot more hands off.

I personally wouldn't want to use anything of A.I to from a relationship point of view, if you're building a business, online and offline, the physical interaction you have with clients will never be able to be emulated from a robotic point of view.