Humans have existed on this planet for over 200,000 years, and yet we still haven't mastered the creation of a truly 'good' human. Now, we aim to develop Artificial General Intelligence (AGI) that is both fair and just. How can finite beings hope to create something with infinite capabilities and insight?
In my view, your miscalculation is believing that something thats more capable than humans = infinite. AI is on a trajectory to be better than humans in some areas, thats not God and its not infinite.
@@darylallen2485 There is a galaxy-sized gap between human intelligence implemented on a 15-watt brain, and the physical limit of the intelligence. An ASI does not need anything remotely close to infinite to be strictly superhuman, and it doesn't need anything remotely close to infinite to be risky.
I think that's actually a good question, and even though I'm in no way an expert in this topic, I'll try to give my take on this: the reason you thought of it this way is probably logically explained like this: although humans are capable of creating machines that can do what we ALREADY do faster, more frequently, and non-stop without exhaustion, they shouldn't be able to do things more ADVANCED than we can do, because we're training them based on our own capabilities, so they should technically be able to do less than or equal to our capabilities at most. But here's a thing worthwhile considering: AI isn't only capable of knowing or insighting what we have already discovered, it is capable of discovering new things sometimes even before us, with the help of pattern recognition and other mathematically-aided intellectual simulations. As long as we have an idea of what is mathematically required to make a computer achieve a certain asset, we can make it go beyond us not just in speed or repetition, but also in discovering new patterns, so long as we know HOW to make it discover the new patterns. Small-scale cases of this can and have been observed right now about AI giving us insight into patterns about things we haven't considered before, so the concern the video was talking about was likely about how we would handle a potential larger scale of these cases when we eventually further advance our technologies. I personally believe this will likely take a long while because we haven't even understood our own brain functions on a large enough scale, but it's a good idea to keep an open mind on this possibility.
Humans have existed on this planet for over 200,000 years, and yet we still haven't mastered the creation of a truly 'good' human. Now, we aim to develop Artificial General Intelligence (AGI) that is both fair and just. How can finite beings hope to create something with infinite capabilities and insight?
In my view, your miscalculation is believing that something thats more capable than humans = infinite. AI is on a trajectory to be better than humans in some areas, thats not God and its not infinite.
@@darylallen2485 There is a galaxy-sized gap between human intelligence implemented on a 15-watt brain, and the physical limit of the intelligence. An ASI does not need anything remotely close to infinite to be strictly superhuman, and it doesn't need anything remotely close to infinite to be risky.
I think that's actually a good question, and even though I'm in no way an expert in this topic, I'll try to give my take on this: the reason you thought of it this way is probably logically explained like this: although humans are capable of creating machines that can do what we ALREADY do faster, more frequently, and non-stop without exhaustion, they shouldn't be able to do things more ADVANCED than we can do, because we're training them based on our own capabilities, so they should technically be able to do less than or equal to our capabilities at most. But here's a thing worthwhile considering: AI isn't only capable of knowing or insighting what we have already discovered, it is capable of discovering new things sometimes even before us, with the help of pattern recognition and other mathematically-aided intellectual simulations. As long as we have an idea of what is mathematically required to make a computer achieve a certain asset, we can make it go beyond us not just in speed or repetition, but also in discovering new patterns, so long as we know HOW to make it discover the new patterns. Small-scale cases of this can and have been observed right now about AI giving us insight into patterns about things we haven't considered before, so the concern the video was talking about was likely about how we would handle a potential larger scale of these cases when we eventually further advance our technologies. I personally believe this will likely take a long while because we haven't even understood our own brain functions on a large enough scale, but it's a good idea to keep an open mind on this possibility.
His alarmist views on AI resemble Luddites and his actions have directly benefited those who are already abusing AI.