The future of AI lies in neuromorphic chips
ฝัง
- เผยแพร่เมื่อ 8 ก.ย. 2024
- Do you want to know what the future of artificial intelligence will be like? Take a look at neuromorphic chips! In this video, we explore the realm of neuromorphic computing, a novel approach to artificial intelligence that aspires to reproduce the structure and function of the human brain in silicon. The benefits of this technology, including as energy efficiency, real-time processing capability, and fault tolerance, are discussed, as well as the problems and restrictions that must be addressed. This movie gives an in-depth look at the intriguing realm of neuromorphic computing and why it is considered the future of AI, from its potential applications in domains such as robotics and autonomous cars to its ethical implications. Don't miss out on this thought-provoking and fascinating video!
Discord: / discord
![[TH] 2024 PMSL SEA Finals D3 | Fall | รู้ผลทีมแชมป์ พร้อมตั๋วสู่ PMGC](http://i.ytimg.com/vi/7NxsN0Ch6vk/mqdefault.jpg)
neuromorphic technology will be a game changer for the field of AI and will certainly develop new field of computing and a new possibility of may be human thinking...
Cool story chatgpt
Lmk when you stop censoring explanations on the ✡️
@@KA-vs7nl This channel's mission is to make videos about science using cutting-edge technologies, and every piece of information is checked before publishing.
Digital becomes conscious, or consciousness becomes digital ,
Brainchip Akida. Should be in our wearables next year.
the fact that this video is narrated by AI is creepy
I take this as feedback, as I used text-to-speech for the voice, but the script was written by me.
Just recently they have done photonic quantum chips which doesn't require the whole 0 Celsius states and big machinery. I think we will soon enter this mix of quantum photonic chips used with this neuron brain state
Yeah, I also think a dynamic chip system may be the better way to achieve this; I think combining them will be the more powerful solution, but it will require further research, which is not too long now.
Microcombs sound amazing.
Eventually it will be more efficient to replace all robots with organics. Because bones and skin regrow and bacteria and other viruses are easier to detect and fight off
Both have their own tradeoffs; a better way is always to create a dynamic, always-changing system that can run in a feedback loop that will help it be dynamic. because one of the key issues with organics is that often organic or biological stuff has very different priorities like survival and reproduction, so it is very important to make it dynamic, and genetic errors are also a very complex problem to solve.
Programmable matter.
Dynex is working on this
I should go back to college again.
Being interested in science and willing to learn new things is already good enough.
No need! *literally* everything you can learn in college, you can learn online/by yourself! Heck, use these upcoming new models to help tutor you, and you’ll be better than most college grads!
I get more idea using this video . thankyou. can you give me the references that you use .
Woah, this has finally been the video that makes me consider that digital will one day become conscious. A very freaky idea indeed. I hope to be alive during the singularity and pass away right before bots overpower us;)
Yes, it will come, and we may try to stop it out of fear, but that is the path of human intelligence.
when the robot kills you, you know its too late...
No technology is good or bad, it's us who make them look bad or good to us, our human infrastructure forces individuals to often use new technology that often makes them seem bad, we could have investigated why some one might do that, go into deeper psychologically of that mindset check its environmental or genetic veriable of that mindset and decide how can we minimise that mindset without harming or taking freedom from humans.
But we will not do that rather we will make rules and create an additional layer of infrastructure in an already super flawed system that creates problems at the higher speed then we can even imagine.
What would be the connections between neurons, that entire mesh of connections is still a bottleneck to implement. That's often addressed by increasing Sequential access speeds. It is not the straight solution
I think photonic and optoelectronic neuromorphic computers might be the approach that this needs; light is the fastest, and brain structure is very energy efficient, and we might find better ways to make efficient computers by using AI, and we humans also lack speed and technique to iterate on a faster scale and more efficiently, so using AI to accelerate AI will be the next best approach.
A computer with this chip will be better than a quantumm computer?
I would doubt it, as quantum computing architecture has better scaling potential, and the purposes of both are also quite different, like quantum computers are usually in demand in physics and nuromorphic in AI, where their efficiency is a lot better.
Neuromorphic and Neural Networks are much better, cheaper, with much more similarities to the human brain and much more compact in size for AI at the edge than huge and ultra expensive hardware needed to run quantum superconducting transmons.
@@azhuransmx126 thanks for your answers
If technology is mimicking the brain it will fall into the same "mistake".. a person uses as much as possible that can be understood from life. Which means we understand life as we understand it from there wwe grow..gaining space within our own brain.. why is that so since we have so much more brain capacity than we actually need? WHY DON´T brains compute stuff and get us results effortlessly ? like my computer is doing right now for me?
Sorry to say, but I didn't completely get what you were saying. The brain is efficient because evolution made it that way; it uses way less energy than any modern home computer. The brain is a very dynamic system; it rewires its structures and constantly adapts as it needs to incorporate new information within the existing framework.
I think direct replication of brain structure may still pose some limitations, especially in speed, as our brain is very slow but dynamic. I think combination, fast solutions like photonics, and efficient solutions like nuromorphic may be better solutions.
This was funny when it was happening to artists
Initially, all living creatures have inherent limitations that are fundamental to their survival and reproduction. In humans, for example, we require specific abilities and capabilities to thrive in our environment. However, we also seek to do more than just survive and reproduce; we want to solve problems, improve management, and achieve other objectives.
One of the fundamental flaws that we face is the limited scope of our language and core concepts. Many of these are based on ancient ideas and are not aligned with modern science. This limitation can impede our progress and prevent us from developing new solutions to complex problems.
However, with the advent of advanced technologies such as artificial intelligence and neuromorphic chips, we have the potential to create creations that can surpass our limitations and help us solve problems in ways that were once impossible. By creating machines that can adapt and learn, we can accelerate our progress and achieve our objectives more efficiently. Ultimately, the purpose of these creations will be to surpass us in both capability and priority, enabling us to solve complex problems and achieve our goals more effectively than ever before.
chips that can recombine their processing routes physicly ? i am more worried about the software engineer well being. that is a specialisme not everybody is being able too master. does the dude have enough energy drinks and pizza ?
That was meant to say that, like how our brain dynamically manages the neural pathway, it is much more dynamic in how much resources it has and what it needs to do.
Nuromorphic chips try to do the same, and programming wise, I think we will only need to do the base; other dynamic processing might get managed by the chip itself.
neuromorphic
agi
Yeah, it can possibly take the ai to a bigger more capable state.
why do you have a robot narrator? Just ask a human to read this. It would be much better. I'll read your next script for you.
I agree that it would have been significantly better, but I don't feel confident about my voice being suitable for these kinds of videos. It might irritate people and discourage them from watching the video. Since this channel is one of the few science-first channels on TH-cam and is not yet large enough to afford paying someone to narrate the videos, I think using AI voices in the future may produce less robotic sounds. Many people also mistakenly believe that my video scripts are generated by AI, but that is not true. I conduct my own research, and the information in most of the videos on this channel comes from my articles on my website with the same name. As a non-native English speaker, I use AI tools to make my words and sentences clearer and more refined.
@@scifilogic I understand, but I feel that the world would be a better place if we heard your real voice. I don't know how severe your accent is, but more diversity is better, in my opinion. I'll happily audition your next script, fwiw.
I am about to make a Discord server for this channel; after making it, I will send you the invitation link. After that, we can decide what we should do.
@@scifilogic sounds great, thanks
@@nobody-u-know discord.gg/WBKGvnNH
Thank you AI generated video 😂
Thanks Biological intelligence 😊
This is BS! No manufacturer and materials can produce this kind of chip:(
Evolution created it, and humans are not far from it. Humans have already become a major factor in the evolution of life.
Understand this: AI will soon be helping to accelerate its own development.
Intel is already doing it. Plenty of research and implementations are already in place.
One Google search is always worth it. 👍
Brainchip company - Akida chip. Already taped out, commercialising now. BS you say?
$DNX @Dynex