This doesn't make sense. If you could scale up limitless numbers of virtual Senior/staff level software engineers, the whole career would die assuming their quality of output was similar to the real thing. Sure maybe some career would come along to replace that, but I'm not seeing what that would look like.
Human level means at the level of a human. Replace your 6 senior software engineers with 6 random humans, heck, replace them with 600 random humans. Let them lose on a problem and report back the results.
At the same time it seems to ignore the automatic absence of certain human features that are clearly net-negative for the human when doing "level" comparison with the bots.
I am thinking of things like: need for sleep or rest, need to eat, need to tend to relations and activities outside of work etc.
When he calls it "human-level" he is ignoring this. Perhaps intentionally.
Yes, it certainly means that there should be some structure in the work, but do we need 6 or 600 human level AI agents or instead all can be done by just 1 agent that would do the human intelligence level work just faster?
Imagine the same problem that happened with computation. One human can compute, but slowly, now this can be scaled up by multiple human computers, doing organized task of computing - but even the best large scale human computing efforts can be simply replaced by a single electronic computer.
Human-level AI will radically change everything, and superhuman AI would probably come right after, so he’s using his own definition of “change” here.
A virtual human that doesn’t need a salary? Doesn’t need breaks? Can work 24/7 at superhuman speed? Can multitask with ruthless efficiency?
That…would change the world much less than we think?
Intelligence as a bottleneck will be eliminated overnight, simultaneously, in all domains.
But sure, some things that we don’t like such as conflict and greed will still be around, and hence things will change less than expected if that’s what you’re hoping will vanish.
AI could develop VR for people to either keep the World seemingly as is for people, or it could just not affect any change to the shape of the World, as Sam mentioned, unless it for whatever reason wanted to please flat earthers, but what are the odds?
The only credible reason I would see for AGI not radically changing the world is if it were limited by available power because of thermodynamic limitations.
Maybe less than we think for the rich and privileged. They still make money. THe rest of us? Human-level AI will probably put a lot of people out of white-collar jobs and into manual labor.
16 comments
[ 4.3 ms ] story [ 46.0 ms ] threadAt the same time it seems to ignore the automatic absence of certain human features that are clearly net-negative for the human when doing "level" comparison with the bots.
I am thinking of things like: need for sleep or rest, need to eat, need to tend to relations and activities outside of work etc.
When he calls it "human-level" he is ignoring this. Perhaps intentionally.
Imagine the same problem that happened with computation. One human can compute, but slowly, now this can be scaled up by multiple human computers, doing organized task of computing - but even the best large scale human computing efforts can be simply replaced by a single electronic computer.
Step 2: Use super AGI to make you own strategically important parts of the economy.
Step 3: Do as you please.
A virtual human that doesn’t need a salary? Doesn’t need breaks? Can work 24/7 at superhuman speed? Can multitask with ruthless efficiency?
That…would change the world much less than we think?
Intelligence as a bottleneck will be eliminated overnight, simultaneously, in all domains.
But sure, some things that we don’t like such as conflict and greed will still be around, and hence things will change less than expected if that’s what you’re hoping will vanish.
Adding 1 trillion more human minds to the cloud would obviously change everything