techlegals
2 views
Skip to first unread message
Marcus Roy
Jul 27, 2023, 8:11:56 AM7/27/23
to wikitechblog
techlegals
ARTIFICIAL INTELLIGENCE, MYTHS AND REALITIES
Artificial ingenuity Does it help us to have a machine that behaves just like someone with billions on our planet? ... In specific, technological expertise fiction and Hollywood have given us many futuristic and often apocalyptic visions, in which scenarios are set in which machines are so superior that humans end up client objects or are changed.
Obviously there is lots of proof that machines have a few talents that improve ours, but there may be additionally masses of others that we do not see machines come close to our skills.
But even if we hit this clinical milestone, we have come to the notion that there may be no point in having a machine that behaves like a human with so many billions on our planet.
Our vision is that we are increasingly depending on our era, but instead of striving for an "empire" of machines, we usually generally tend closer to a symbiosis in which we are capable of extend our talents. But there's no question that our technology requires the durability of our abilities.
In this section we will attempt to research the real extent of the three phenomena in recent times, eagerly awaited through the general public and intently associated with the sphere of AI improvement.
A futuristic situation of technological uniqueness became proposed.
The technological singularity basically defines that there'll come a time when generation advances so unexpectedly that machines can beautify and increase an ingenuity that a long way exceeds our highbrow and physical competencies and then our own. Control.
This debatable concept stems in huge component from Moore's law, proposed in 1965, which says that the variety of transistors in a microprocessor doubles more or much less each years. In different words, computing power is developing exponentially. Since then, his prediction has come real with some accuracy.
We presently have extraordinary thoughts as to whether or not or not this law may additionally even observe inside the destiny and whether or not or no longer this boom in computing power ends within the singularity noted above.
On the only hand, some authors like Ray Kurzweil (modern CTO of Google) have advised that the tech singularity will occur in the center of the twenty first century.
But even Moore himself declared in 2010 that his law have become obsolete, pointing out that there are bodily limits to the compactness of chips and the speed they might gain without melting.
Intel, a leading manufacturer of microprocessors, confirms it keeps to conform with Moore's Law to in recent times and plans to hold to obtain this, albeit with a few caveats. By 2020 they want to have 7 nm processors, it truly is considered the smallest plausible period.
The future of Moore's law
Throughout this time, many believed that reality would quickly surrender, even though till now they were all wrong.
In any case, a number of these physical limitations imposed on us with the aid of hardware have been offset by way of way of various technologies, collectively with:
A) GPU
When using multiple processors, cores, or snap shots processing devices (GPUs), are essentially coprocessors designed to address photos or floating component operations to lessen the CPU workload and increase processing power.
B) Parallel computing
We moreover solved part of the trouble with parallel computing; H. With multiple computing sources at the identical time to remedy a computing trouble. Resources are processors or cores, they are able to are dwelling on a computer with multiple cores or multiple computer systems connected to a network, or maybe in a combined surroundings.
We divide the problem into many elements that may be solved at the same time, this is, on the equal time. A separate CPU is assigned to each part. That manner, you could remedy troubles faster or treatment troubles that require more reminiscence than a unmarried laptop.
But right here too we find large barriers, on occasion the issues are not clean to parallelize, in truth there are a few problems which are inherently sequential, consisting of linear programming, iterative numerical strategies, in quick, there are many troubles which can be difficult to clear up are. Parallelize.
Obviously there is lots of proof that machines have a few talents that improve ours, but there may be additionally masses of others that we do not see machines come close to our skills.
But even if we hit this clinical milestone, we have come to the notion that there may be no point in having a machine that behaves like a human with so many billions on our planet.
Our vision is that we are increasingly depending on our era, but instead of striving for an "empire" of machines, we usually generally tend closer to a symbiosis in which we are capable of extend our talents. But there's no question that our technology requires the durability of our abilities.
In this section we will attempt to research the real extent of the three phenomena in recent times, eagerly awaited through the general public and intently associated with the sphere of AI improvement.
A futuristic situation of technological uniqueness became proposed.
The technological singularity basically defines that there'll come a time when generation advances so unexpectedly that machines can beautify and increase an ingenuity that a long way exceeds our highbrow and physical competencies and then our own. Control.
This debatable concept stems in huge component from Moore's law, proposed in 1965, which says that the variety of transistors in a microprocessor doubles more or much less each years. In different words, computing power is developing exponentially. Since then, his prediction has come real with some accuracy.
We presently have extraordinary thoughts as to whether or not or not this law may additionally even observe inside the destiny and whether or not or no longer this boom in computing power ends within the singularity noted above.
On the only hand, some authors like Ray Kurzweil (modern CTO of Google) have advised that the tech singularity will occur in the center of the twenty first century.
But even Moore himself declared in 2010 that his law have become obsolete, pointing out that there are bodily limits to the compactness of chips and the speed they might gain without melting.
Intel, a leading manufacturer of microprocessors, confirms it keeps to conform with Moore's Law to in recent times and plans to hold to obtain this, albeit with a few caveats. By 2020 they want to have 7 nm processors, it truly is considered the smallest plausible period.
The future of Moore's law
Throughout this time, many believed that reality would quickly surrender, even though till now they were all wrong.
In any case, a number of these physical limitations imposed on us with the aid of hardware have been offset by way of way of various technologies, collectively with:
A) GPU
When using multiple processors, cores, or snap shots processing devices (GPUs), are essentially coprocessors designed to address photos or floating component operations to lessen the CPU workload and increase processing power.
B) Parallel computing
We moreover solved part of the trouble with parallel computing; H. With multiple computing sources at the identical time to remedy a computing trouble. Resources are processors or cores, they are able to are dwelling on a computer with multiple cores or multiple computer systems connected to a network, or maybe in a combined surroundings.
We divide the problem into many elements that may be solved at the same time, this is, on the equal time. A separate CPU is assigned to each part. That manner, you could remedy troubles faster or treatment troubles that require more reminiscence than a unmarried laptop.
But right here too we find large barriers, on occasion the issues are not clean to parallelize, in truth there are a few problems which are inherently sequential, consisting of linear programming, iterative numerical strategies, in quick, there are many troubles which can be difficult to clear up are. Parallelize.
Read MORE : techlegals
Reply all
Reply to author
Forward
0 new messages