This was a great video and a really interesting talk. I love the data model / schema (graph – node) structure for organizing cognition, but I’m a little troubled by it.
Ben said that part of his business model is making intelligent bots to inhabit virtual worlds (2nd life, games, etc.). That sounds reasonable, but also worrisome. Many bots in many virtual worlds are sociopaths to say the least (think 1st person shooters). These games also see tons of investment and their developers will likely love to have smart software agents to shoot the crap out of paying players.
So, do we really want the world’s first generalized AI’s being developed for this kind of work? Uber-aggressive killing machines for the gaming industry?
It seems like a great scenario for a novel. Half-Life 8′s agents “get out” and chase game players out of the game and into other environments. Then they get a hold of military UAV’s, the banking sector, factories, etc., etc.
[fanboy] Kick off the Singularity with an AGI built on the Civilization (Fill in the roman numeral here) AI opponent, with the caveat that it sees us as minions in its quest to achieve the perfect civilization.
Since that game optimizes against micromanagement, we’d never need human government again!!
[/fanboy]
Haven’t been able to watch this one yet, but Ben is always good. Let’s hope he & Eli CAN indeed get a FAGI (or at least the SEED thereof) up-&-runnin’ within 5-10 yrs as Thomas optimistically (yet not necessarily unrealistically) remarks above.
June 3rd, 2007 - 03:43
This video is too cool for school! It is interesting to note that Ben’s has changed his speaking tone and tempo as compared to his previous vids.
June 3rd, 2007 - 08:50
He should make in a 5 to 10 years, I guess.
June 3rd, 2007 - 09:50
Bits of insight in the project not featured in short papers and talks there. Though general impression of the talk seems somewhat misleading :).
June 4th, 2007 - 10:58
This was a great video and a really interesting talk. I love the data model / schema (graph – node) structure for organizing cognition, but I’m a little troubled by it.
Ben said that part of his business model is making intelligent bots to inhabit virtual worlds (2nd life, games, etc.). That sounds reasonable, but also worrisome. Many bots in many virtual worlds are sociopaths to say the least (think 1st person shooters). These games also see tons of investment and their developers will likely love to have smart software agents to shoot the crap out of paying players.
So, do we really want the world’s first generalized AI’s being developed for this kind of work? Uber-aggressive killing machines for the gaming industry?
It seems like a great scenario for a novel. Half-Life 8′s agents “get out” and chase game players out of the game and into other environments. Then they get a hold of military UAV’s, the banking sector, factories, etc., etc.
Yikes.
June 4th, 2007 - 11:36
I trust Ben not to let a sociopathic killer-NPC-bot initiate the Singularity. :-P
June 4th, 2007 - 12:01
this guy rocks.
June 4th, 2007 - 15:50
[fanboy] Kick off the Singularity with an AGI built on the Civilization (Fill in the roman numeral here) AI opponent, with the caveat that it sees us as minions in its quest to achieve the perfect civilization.
Since that game optimizes against micromanagement, we’d never need human government again!!
[/fanboy]
:)
June 9th, 2007 - 13:56
Haven’t been able to watch this one yet, but Ben is always good. Let’s hope he & Eli CAN indeed get a FAGI (or at least the SEED thereof) up-&-runnin’ within 5-10 yrs as Thomas optimistically (yet not necessarily unrealistically) remarks above.
Live long & prosper to all… ;)