These forums have been archived and are now read-only.

The new forums are live and can be found at https://forums.eveonline.com/

EVE General Discussion

 
  • Topic is locked indefinitely.
Previous page12
 

Problems of EVE online

First post
Author
Previous Topic Next Topic
Ralph King-Griffin
New Eden Tech Support
#21 - 2016-09-05 17:48:53 UTC
Since the automated forum logic defense unit (codename:Tippia) was decommissioned there have been no thinking ai's

Murderers of Negotiable Motivations

=]|[=
Solecist Project
#22 - 2016-09-05 18:00:49 UTC
Andrea Cemenotar wrote:
I'm sorry to agree with solecist that your information on AI technology is sliiiightly outdated

Why do people always apologize when they agree with me... pffff....

That ringing in your ears you're experiencing right now is the last gasping breathe of a dying inner ear as it got thoroughly PULVERISED by the point roaring over your head at supersonic speeds. - Tippia
Keno Skir
#23 - 2016-09-05 18:10:45 UTC
Solecist Project wrote:
Andrea Cemenotar wrote:
I'm sorry to agree with solecist that your information on AI technology is sliiiightly outdated

Why do people always apologize when they agree with me... pffff....


... ah come on.

Black Lanterns Blog <- Read my ramblings -.-

250,000 Bonus SP when you start an Alpha Clone HERE <---
Andrea Cemenotar
Elena Minasse Operations
#24 - 2016-09-05 18:17:34 UTC  |  Edited by: Andrea Cemenotar
Dirty Forum Alt wrote:
Solecist Project wrote:
The difference between "looks like thinking" and "is thinking" ...
... was important back when we had bullshit AI that didn't deserve the name.

I would argue that the difference is still important regardless of how advanced simulated AI gets.

Ultimately the difference between "looks like thinking" and "is thinking" is the difference between a program that *does what its designers intended it to do* and one that *does things it was never designed or intended to do*.

The extreme example of a computer that "is thinking" would be for example the classic robot that decides to kill all humans, because it is better than humans are.

A less negative example would be a google search AI that developed a cure for cancer because it noticed a lot of people searched for it and figured it would help them out.

Neither is likely to happen any time in the foreseeable future - because we are nowhere near "true" AI where computer programs "actually think" - we are, as I said, just better and better at making it *look like* they are thinking.


Also I never mentioned chess program AI specifically at all - so not sure why you brought that up...


edit:
Also of course I'm not "up to date" as far as right this second - if anybody had *true* AI they'd have it classified so hard we'd never hear about it, unless you happened to work for the company. That would be a major breakthrough, and they wouldn't want their competitors to get ahold of it until it was finished and they had figured out a way to exploit it to the max.

edit #2:
Technology progresses a lot slower than people think in some cases. For example: 15-20 years ago I was 100% guaranteed I'd be using a quantum computer within 10 years. I'm still waiting...



as I mentioned in my edit - it really depends on how do you define "thinking"

if you use the definition I most commonly see [performing detailed analysis to determine further actions] then there are publically known AI softwares that would easilly classify as being able to "think"

please bear in mind we are talking about "thinking" not "wondering" :P


EDIT:
Solecist Project wrote:
Andrea Cemenotar wrote:
I'm sorry to agree with solecist that your information on AI technology is sliiiightly outdated

Why do people always apologize when they agree with me... pffff....


I was apologising that other person for supporting someone who was opo
Dirty Forum Alt
Forum Alts Anonymous
#25 - 2016-09-05 19:04:55 UTC
Andrea Cemenotar wrote:
as I mentioned in my edit - it really depends on how do you define "thinking"

if you use the definition I most commonly see [performing detailed analysis to determine further actions] then there are publically known AI softwares that would easilly classify as being able to "think"

please bear in mind we are talking about "thinking" not "wondering" :P

By that definition most computer programs, no matter how unintelligent they are, are "thinking". They take in inputs, perform analysis, determine what pre-programmed criteria (if any) are met, and use that information to determine which pre-programmed action to take.

This is why people strive for Artificial *intelligence* - not Artificial *thinking* - and in discussions about Artificial *intelligence* it is wise to take this into account, otherwise you risk rendering the entire discussion meaningless.

The dead swans lay in the stagnant pool. They lay. They rotted. They turned Around occasionally. Bits of flesh dropped off them from Time to time. And sank into the pool's mire. They also smelt a great deal.

Paula Nancy Millstone Jennings (Sussex)
Solecist Project
#26 - 2016-09-05 20:34:07 UTC
Keno Skir wrote:
Solecist Project wrote:
Andrea Cemenotar wrote:
I'm sorry to agree with solecist that your information on AI technology is sliiiightly outdated

Why do people always apologize when they agree with me... pffff....


... ah come on.


Andrea Cemenotar wrote:
I was apologising that other person for supporting someone who was opo



I wasn't 100% serious. :p

That ringing in your ears you're experiencing right now is the last gasping breathe of a dying inner ear as it got thoroughly PULVERISED by the point roaring over your head at supersonic speeds. - Tippia
Shallanna Yassavi
qwertz corp
#27 - 2016-09-05 21:36:39 UTC
The AI which could beat a chess grandmaster was impressive in its day. It worked by selecting the most likely next moves from all possible game moves.
The more recent one which could beat the best Go players in the world couldn't work the same way. They had to use more modern machine learning.
They made one which beat the best Starcraft players in the world.
They made one which beat the best fighter pilots in simulators regularly.

EVE could be another such challenge. How well would a proper, modern AI FC, with either human or AI fleet members? How well would it be able to come up with and execute a doctrine, and how effective would said doctrine be? How well would it be able to do logistics?

A signature :o
Dirty Forum Alt
Forum Alts Anonymous
#28 - 2016-09-05 21:55:23 UTC
Shallanna Yassavi wrote:
The AI which could beat a chess grandmaster was impressive in its day. It worked by selecting the most likely next moves from all possible game moves.
The more recent one which could beat the best Go players in the world couldn't work the same way. They had to use more modern machine learning.
They made one which beat the best Starcraft players in the world.
They made one which beat the best fighter pilots in simulators regularly.

EVE could be another such challenge. How well would a proper, modern AI FC, with either human or AI fleet members? How well would it be able to come up with and execute a doctrine, and how effective would said doctrine be? How well would it be able to do logistics?

It would certainly be an interesting test - however during the early phases of data collection such programs tend to be incredibly clumsy and obvious... So it would be hard to get it through undetected. and I suspect that if CCP openly allowed such an experiment they would lose a lot of players over the idea of willingly letting a bot play EVE.

Also I tend to suspect that an "AI" fleet commander would struggle to maintain control over real players in EVE - so it would likely be significantly more effective with an entire "AI" fleet.

The dead swans lay in the stagnant pool. They lay. They rotted. They turned Around occasionally. Bits of flesh dropped off them from Time to time. And sank into the pool's mire. They also smelt a great deal.

Paula Nancy Millstone Jennings (Sussex)
Ezin
Doomheim
#29 - 2016-09-06 02:30:01 UTC
Alexander Bor wrote:

Ganking possibility must be seriously restricted


Would you agree on mining restrictions as well?
lets say you may only mine 50K of cubic meters per day.
Infinity Ziona
Sebiestor Tribe
Minmatar Republic
#30 - 2016-09-06 09:31:37 UTC
Mara Pahrdi wrote:
Alexander Bor wrote:
Ganking possibility must be seriously restricted

Why?

EvE is supposed to be a harsh and difficult game. EvE is supposed to reward effort. EvE is supposed to be a non-consentual pvp game.

Sitting safely in highsec in a super cheap ship that can be replaced in a single tick of anom ratting and killing empty freighters costing billions of isk is not harsh. Sitting safely in highsec where anyone attempting to preemptively attack you is insta-gibbed by NPC protectors is not non-consensual pvp.

Ganking is the ultimate in carebearing because the risk is so very low. Compare risking a 500 million isk battleship when you undock to do a mission vs risking a 10 million isk catalyst. Compare risking a 2 billion freighter and cargo or a 6 billion isk jump freighter and cargo to risking a 10 million isk catalyst.

Its so cheap and easy if you check Baltec1's last session of ganking in highsec you'll see he and goons killed 17 ships, 14 were freighters, 13 were empty. When ganking is so cheap you can kill around 40 billion worth of empty ships for around 150 million per pilot there is a serious imbalance.

It should never be so cheap that you can gank random empty ships.

CCP Fozzie “We can see how much money people are making in nullsec and it is, a gigantic amount, a shit-ton… in null sec anomalies. “*

Kaalrus pwned..... :)
Nalia White
Tencus
#31 - 2016-09-06 17:29:11 UTC
Solecist Project wrote:
Dirty Forum Alt wrote:
So as of now at least, machines *can not* think... But people are getting much better at making them *look* like they can think.
Common ChessAI isn't thinking, but looks like it.
Calling a chess program an AI nowadays is actually insulting to modern AI.

I kind of doubt you're up to date, unless the "last time" you mention was last week...
Progress is at an insane rate.
Free AI tools, devkits, SDKs are spreading everywhere.

AlphaGo on the other hand got actually creative.
Creative! It made actually brilliant moves no one thought about!
AlphaGo wasn't just a pattern machine, but a highly trained AI.

The difference between "looks like thinking" and "is thinking" ...
... was important back when we had bullshit AI that didn't deserve the name.

Nowadays the difference is completely useless, because the end result is the same.


Hell, Google made an AI that - all by itself - managed to identify a paper shredder ...
... on a picture of a full office room ...
... without ever getting taught what a paper shredder is.

And they had NO idea how that happened!


Let that sink in for a minute...


and all will happen again... :) (don't remember the exact phrasing)

Syndicate - K5-JRD

Home to few, graveyard for many

My biggest achievement
Previous page12