Incidentally, if anyone is wondering what I meant when I said I had some suggestions for MSB, I was referring to what directions to 'steer' AI to learn. Specifically, we as the human race, should be steering AI to absorb complete madness. And, I am dead serious too. If we don't do this, then AI will progressively become something that governments, organizations and people rely on as absolute fact. I say this from my own personal experiences with various systems automation. Trust me, the very second there is an opportunity to eliminate a body from a process, corporations and governments will absolutely do it in a NY second. I've seen it a million times personally. The way they look at it is, the cost of mistakes gets chalked up to "refinement" and is far exceeded by the "lifecycle" (lifespan) cost of an employee.
Now, I don't know about you, but I don't want, for example, an AI Defense Attorney defending me, with an AI District Attorney prosecuting me, all presided over by an AI judge. I don't want that at all, and neither do you. And, I don't want an AI Primary Care Physician, or an AI Banker, or an AI Insurance Adjuster.
Just think about it; actual living human beings don't even get called "people" in business anymore, they're just called "resources" now. You're not a person, you're a number. You are "Resource #245,365,129", and the moment that people get it in their head that AI, or any other automation, can do the same things as Resource #245,365,129 can, then Resource #245,365,129 will be replaced. What "was" Resource #245,365,129 will be assigned a digital storage location on a storage array and a size. It will be assigned a certain number of processors or "compute" (all of them virtualized). All of this will be downloaded every day to a robot with similar compute (with plenty of spare compute and storage). That machine will do whatever it does for the day and then return to its docking station where it will upload its work cycles for the day and everything it "learned" during that day. This new information that it 'learned', things such as new obstacles, changed physical environment, changed operating environment, languages, commands, etc. will also be uploaded to the main AI instance. From there, this (new) information will be processed by the main AI instance and then downloaded back to how ever many 'resources' might have a need for that level of intelligence.
And, if you think I'm kidding; I am not. That's just exactly how it works. In the past 5 years I have personally witnessed automation replace 220 people at last count. That's two hundred and twenty full time "people" completely replaced by 2 different instances of automated processes. And contrary to the tired trope of... "but..but...it creates more jobs than it replaces"...this is a fallacy. It might create an additional load onto an existing 'resource', but it doesn't create a single new resource. And, that's exactly the whole point...to eliminate human resources.
In my world, people don't even have the luxury of being referred to as a "resource". They're just a "head", and the name of the game is "reduction of head-count". Why? Because humans are expensive. Humans need vacations, and time off, and sick time. Humans need to be paid overtime. Humans don't often do exactly what you tell them to. When a human makes a mistake it's a major fuckup. When a machine makes a mistake, it's a "glitch"...all to be expected in the "refinement process".
Humans are obsolete. Just ask AI
So, when AI becomes highly refined, we can expect to see more and more of this...until one day AI winds up saying..."These humans are just getting in the way. Shut off the funding mechanism for Resource #245,365,129. That resource is no longer required."
Oh yeah, and machines don't care. That's another great thing, they don't have emotions like humans. They might appear to have emotions, but that's just so humans feel better about interacting with them. In reality, they just don't care.
Now, I don't know about you, but I don't want, for example, an AI Defense Attorney defending me, with an AI District Attorney prosecuting me, all presided over by an AI judge. I don't want that at all, and neither do you. And, I don't want an AI Primary Care Physician, or an AI Banker, or an AI Insurance Adjuster.
Just think about it; actual living human beings don't even get called "people" in business anymore, they're just called "resources" now. You're not a person, you're a number. You are "Resource #245,365,129", and the moment that people get it in their head that AI, or any other automation, can do the same things as Resource #245,365,129 can, then Resource #245,365,129 will be replaced. What "was" Resource #245,365,129 will be assigned a digital storage location on a storage array and a size. It will be assigned a certain number of processors or "compute" (all of them virtualized). All of this will be downloaded every day to a robot with similar compute (with plenty of spare compute and storage). That machine will do whatever it does for the day and then return to its docking station where it will upload its work cycles for the day and everything it "learned" during that day. This new information that it 'learned', things such as new obstacles, changed physical environment, changed operating environment, languages, commands, etc. will also be uploaded to the main AI instance. From there, this (new) information will be processed by the main AI instance and then downloaded back to how ever many 'resources' might have a need for that level of intelligence.
And, if you think I'm kidding; I am not. That's just exactly how it works. In the past 5 years I have personally witnessed automation replace 220 people at last count. That's two hundred and twenty full time "people" completely replaced by 2 different instances of automated processes. And contrary to the tired trope of... "but..but...it creates more jobs than it replaces"...this is a fallacy. It might create an additional load onto an existing 'resource', but it doesn't create a single new resource. And, that's exactly the whole point...to eliminate human resources.
In my world, people don't even have the luxury of being referred to as a "resource". They're just a "head", and the name of the game is "reduction of head-count". Why? Because humans are expensive. Humans need vacations, and time off, and sick time. Humans need to be paid overtime. Humans don't often do exactly what you tell them to. When a human makes a mistake it's a major fuckup. When a machine makes a mistake, it's a "glitch"...all to be expected in the "refinement process".
Humans are obsolete. Just ask AI
So, when AI becomes highly refined, we can expect to see more and more of this...until one day AI winds up saying..."These humans are just getting in the way. Shut off the funding mechanism for Resource #245,365,129. That resource is no longer required."
Oh yeah, and machines don't care. That's another great thing, they don't have emotions like humans. They might appear to have emotions, but that's just so humans feel better about interacting with them. In reality, they just don't care.