As someone who's been a computer nerd my entire life and now over 60, I've seen tech hype come and go. When it comes to AI, companies like Nvidia and startups like Groq are really just selling faster chips-just like how they update phones or video cards (GPUs). It’s all part of the same cycle. But here’s the reality: Artificial Intelligence isn’t about how fast your hardware is. AI's real power comes from the large language models (LLMs) and how efficiently the software interacts with them. The hardware is just the vessel. The real breakthroughs are happening in the way we develop and access these models, not in speeding up chips. Don’t get caught up in the hype!
1NM chip refers to the size of the chip. The smaller the chip, the more expensive it is. However, a 1NM chip does not mean that it can run faster than a 5nm chip cluster. If the size is larger, more 5NM chips combined can run faster than a 1NM chip. This has to be considered from both cost and application aspects. For example, a missile is larger than a mobile phone, and a 14NM chip cluster is enough to develop AI weapons. In addition, weapon chips also need to be stable and resistant to high temperatures, etc.
@@lijackson-x6r While you're right that smaller chip sizes (like 1nm) don’t always equate to faster performance compared to clusters of larger chips (like 5nm), it's important to remember that AI performance isn't just about the hardware size. The real power in AI comes from how efficiently the software interacts with large language models (LLMs). Whether it's AI in weapons systems or consumer electronics, the key is how well the software leverages the model, not just how small or fast the chip is. The hardware is the tool, but the real intelligence is in the software.
@@Progressive_CanadianI completely agree with this view. AI requires 5G or 6G network infrastructure. Without network speed, no matter how fast the chip is, it will be useless. Moreover, the robotics industry is the foundation of AI applications. Without mechanical engineering and industrial foundation, AI cannot bring anything substantial to people. For example, chips alone cannot bring food to the table. At present, AI is in the early stages of development, but China's industrial foundation and network foundation make it easier for AI to become a reality.
@@lijackson-x6r I appreciate your perspective, but I’d push back a bit on the idea that network infrastructure like 5G or 6G is the critical bottleneck for AI advancement. While network speed is important for certain applications like autonomous vehicles or cloud-based AI processing, it’s not the main factor for AI progress. The real driver of AI's potential lies in large language models and software development, not just hardware or network speeds. As for robotics, you’re right that the industrial base is crucial, but AI’s impact can be felt even without physical machines-think of software-based solutions in healthcare, finance, or creative industries. China may have a strong industrial foundation, but AI is advancing globally, thanks to breakthroughs in software and model development, not just hardware infrastructure.
@@Progressive_CanadianAI software development is important, but AI models must collect a lot of data, all of which comes from the Internet. Only with data can AI models be analyzed and trained. AI language models are only a small part of AI applications. Many AI applications are in industry and medicine. For example, robot surgeons rely on supercomputers in the cloud to perform surgery, so Internet speed is crucial. In addition, remote mining and unmanned port loading and unloading of goods all require fast Internet connections. In addition, satellite image analysis, the data from satellites is very large, and it requires a fast network connection to transmit data and perform image analysis to identify farmland or oil fields or military missiles. Especially in military applications, AI can analyze the trajectory of the opponent's missiles and intercept them in time, which requires fast network transmission.
This report on the battle among AI chip makers for dominance is fascinating! It highlights the intense competition in the tech industry and how crucial these advancements are for the future. Thanks for shedding light on this important topic!
What a great report; they discussed technology, leading AI-Companies, their dependance on Taiwan, and economic and political risks associated with it. Comprehensive work. The compares are really well spoken!
There exist other much more common words pronounced dif between English speaking countries, let alone regions. NVDA and Kamala are not even English words, and even a linguist would suggest pronunciations and even meaning are fluid. That said, I had the same thought, but rather questioned the pronunciation I’ve come to commonly know. Bye
If groq is only focusing on AI inference, it does not compete with Nvidia as Nvidia GPU is more for AI training task. Nowadays more and more CPU has integrated NPU which is meant for accelerating AI inference as well. Groq seems to compete with these integrated NPU and trying to be a better NPU.
There are many AI inference hardware for different applications. Those hardware cannot use for training deep model.. Therefore, Nvidia is still a true king for training.... One DGX is almost equal to a house price..!!
Priya is really great, she has absolutely zero background in tech and she dared to dream big. Such an inspiration, especially to young woman. Need more role models like her to be pushed to the forefront of schooling.
Untuk pemain pasar lama mempertahankan eksistensi perusahaan menjadi hal yang sangat efektif dalam menghadapi hadirnya para pesaing baru , disini naik turunnya tekanan pasar sudah menjadi pengalaman yang menjadikan perusahaan tampil semakin terpercaya oleh konsumen tinggal bagaimana pengembangan selanjutnya
It's a shame that here we have a video on AI, and they don't know the difference between AI and machine l;earning. What we currently have operational is machine learning. It is a start towards maybe one day having true AI. True AI is still only a goal, at this point.
First of all machine learning is a subset of the field of artificial intelligence. Also If you want to get more technical the subsets of machine learning being used for LLMs are Deep Learning and Reinforcement Learning. Using the words “Artificial Intelligence” is just fine. It makes sense to use those words since “machine learning” is not as familiar to the public because it’s relatively new in comparison.
@@2handsandwiches That assumes that computing will always be on/off constrained. However, you may be right. It could be that machine learning will get closer and closer, with advances taking longer and longer. And actual AI might not ever happen. But machine learning might get to the point where it's almost impossible to tell the differencew. I envision a new set of PHDs will emerge that deal specifically in how to define and measure when/if machine learning has finally become actual AI. As we have seen in the past, it is many times foolish to say something will NEVER happen. After all, we have only been doing eletronic sudo-brain type work for less than 100 years. What will happen in the next 1000? 10,000? Well, if we manage to not destroy ourselves first.
We want supersaturation. Supersaturation is a metastable state, meaning it is not in equilibrium and can easily revert to a saturated state. BTW thank you for reminding me this a Talent War & not the Pretentious Inn.They will leave you!
The AI chip wars are heating up! 🔥 Nvidia's dominance is being challenged by startup Groq! 💪 I love seeing innovation and competition in the tech space! 🤖 It's exciting to think about the advancements that will come from this battle for AI chip supremacy! 💻💸 Thanks for sharing, BBC News! 👏 Stay ahead of the curve! 🚀
One day my cloth broke zip handling part, so to close or to open the cloth, somehow the zip keep on zipping part of the cloth and get stuck and cant open, so i have to remove the stuck piece for i to proceed to open only for shortly for it to zip the cloth once again, what does this have to do with artificial intelligence, the cloth is analogy and doesn't have any brain chip. Be patient. Then one afternoon as i was rushing home under the sun, the temperature inside close cloth was at a boiling point so i struggle like usual to unzip the cloth, suddenly i hold both ends of the cloth and pull them away gentle, it open so smoothly without further struggle, few days latter in the morning i hard to unzip once again and i struggle as i was to struggling anyway for a long time as i can remember, suddenly i recall the pull away strategy i discover the other day and i did just that. From then hence forth i learn from nothing to how to unzip without struggling. Anyone agree am conscious and the way i learn is exert the same way deep learning algorithm of artificial intelligence learn. So many example feed on to the algorithm with slit changes each time on solving the issue at hand till this tinny solution amount to something slightly useful that the whole brain now moves towards this trend further in the solving of the issue at hand, from that we get AI chip being correct at solving things accurately at 5% on its own, latter it now at 98% accurate as of now deep learning system are at. Looking deeply why did my brain figure it out just the same way an artificial algorithm would? To sum it all the weights used on deep learning once mapped by the algorithm from sum less biases then name the object or the function will remain so as long as the weights and biases remain constant. That mathematical sequence remain constant indefinitely as long as the weights and biases remain constant. From that we can conclude the idea of our human brain has the element of weight and biases in its logic configuration. To that endeavour its valid to state we have a risk of underestimating what conscious is all about, i am conscious yet i see deep learning not conscious when the fact of the matter is we do share how we learn from being feed to much data and we learn from that, look how many things were discovered by accident from penicillin to aerial dynamic from the two brothers. We cant have two system performing sequence the same way then we go ahead and claim one is 100% conscious while the other 0% conscious, its logic to equate at list a 5% chance of conscious, to a maximum of at list 20% conscious. Once science community agree on this number bet it will grow each day each second to levels exceeding a million percent (1,000,000% conscious ) something we cant even dream of or speculate of.
I love you guys, but I found it hilarious that on a show about faster chips, the CEO’s film (zoom, I assume) is like on dial up. The BBC host noted it too. LOL. And yes, I realize that that’s due to a different thing rather than his type of chip. Still…hilarious.
Saudi Arabia might house the datacenters but if this video call's quality is the extent of their Internet connection, those datacenters won't be able to serve the world.
Nothing works in heat. Spend an other billion dollars. Just like underwater sea resort and tunnels. Who has an ac to work below specifications. Cellphones are now in refrigerator calls. It is called fridge mobile.
AI robots are designed to do hazard or inconvenient jobs for human, and they turns out to be artists, music composer, photographic video creator, finance analysts , and eventually put human out of jobs. fortunately, the lawyers are not needed in the future because AI lawyers will be 10 times better than human lawyers to win the lawsuit, and politicians are not needed because AI will make wiser decision for mankind.
CEO of Groq has the worst connection or hardware ever? That's not a good sign . . . Didn't they test it before the interview? Heh. Remember dial-up? Er, yes. 9600, 14.4, 56 and we were told it wouldn't ever get faster, but they lied! Don't miss those days or the screaming modem sound! Other uses for these GPUs? As graphics cards for PCs? 🙂 Preferably with throttling or some hardware-based deterrent which can't be disabled to discourage cryptomining with them so the prices won't skyrocket after they're put on the market . . . Better auto-translate for TH-cam videos? 🙏 15:00 Talking about the United States supplying supercomputers to the Soviet Union. The cool thing about supercomputers that most don't know? Almost all of them are running on a Linux kernel! 🙂 15:35 Gregory Allen of CSIS (Center for Strategic & International Studies). Hypersonic, for those who don't know it, yet, is Mach five to ten. Anyone remember the X-15? The thing is with these higher-speed missiles is that detecting, targeting and taking them out is far beyond human capability (AKA the human operators or the human component), so that has to be done another way. Even Samuel Renshaw (he had some nice study tips in the end of the Psychological Optics paper he wrote) couldn't train 'ack-ack' (anti-aircraft) gunners to be that fast with his tachistoscope! (Fans of Robert A. Heinlein probably remember characters mentioning Renshaw in his works such as Stranger in a Strange Land (Jubal Harshaw's secretary, Anne, was a Fair Witness who was 'Renshaw-trained'), Gulf (VERY sad ending, but, those who read it will understand his character Friday a little better) and Citizen of the Galaxy.) 17:35 China engaging smuggling networks. Perhaps similar to multihopping with a VPN, but with more than one extra point, with the idea being the more used, the better, provided they're not the weak link types who snitch/squeal/rat you out/be a stool pigeon/sing like a canary/stool with the rogue, et cetera . . . 17:45 Samsung executives giving manufacturing secrets to others. Can you say NDA violation? I knew that you could! Another thing to be wary of is a brain drain scenario or 'poaching' by governments or competitors or reverse-engineering technology. Not to be confused with Operation Paperclip, which was a totally different can of worms . . . 19:50 Taiwan being a crucial resource for logic chips manufacturing and China claiming it as a territory (not in the way of a Risk game territory). Things like this are worrisome, as centralizing such valuable manufacturing in one place makes them highly vulnerable and might possibly lead to a scenario where someone might threaten it, stating that either they'll be allowed to have it and control it and how much they can export and to whom or they'll reduce or eliminate their manufacturing capacity completely (the 'if I can't have it, nobody will' bit). Example: Paul Atreides AKA Maud'Dib with melange AKA spice on Arrakis, threatening all the spice by using the changed Water of Life (via a Bene Gesserit 'witch' like his mother, Jessica) on a pre-Spice mass that would then continue on in a ripple effect to the rest of the planet). And, yes, I probably read too much, especially science-fiction! LOL
9:43 Wow!!! this company is modern! did you listen guys? they do speach to text as well, like what google did 10 years ago and also siri did few years ago!
It's, infact, very difficult to be great ! But it's much more difficult to be a good human being ! In these days of fast - degrading human values, the call of the hour is - a spurt in the number of sane and right - thinking people - in the greater interest of - sustenance of the Humankind and Humanity !
yeah yeah again some big companies will take control of everything, I don't want them to have control. we need to have the control, make chip which normal people can buy and do stuff, this shit is really like game of thrones
This is game changer I Love ❤️ AI the Ultimate equalizer between those who have and have not the vision of the future this is my religion with AI I'm the king of my castle if you are poor and marginalized just master the AI in any field you wish and you can lead the leaders.
More than the problem of processing power is the problem of intelligence. If you call pattern matching intelligence, I guess you are satisfied. If you expect brilliance from AI, you will be disappointed. When they suggest that this technology could write a book in seconds, I'm afraid you will be disappointed if you expect anything more than Dr. Seuss (i.e., see spot run). If it really could, then why did they not take the time to give an example? It will be some time before these AI systems can ever match human intelligence no matter how much processing power they are given. Consider this: AI systems are restricted on their ability to learn. Humans are very much afraid of building AI systems with true intelligence.
Lol, high tech big shot with slow internet and bad webcam. Once he opens his mouth it is only about money. It is alarming that he gives a feeling of “one blood can detect cancer” type of scam 😂
For a very short of period time, advanced AI itself can create its own choices for them very soon what showing in this is old compared with new fiber Chipes and infinity chips made from energy itself these guys are far behind
0:33. NVIDIA is making $120 Billion plus per year. And this Groc 'Raised' 1 Billion $ now. Nvidia's technology is 10-15 years in Future, and these guys are collecting pennies in hope of developing any low-quality AI-chip, which would dream to perform half as good as Nvidia's Chips, which they won't be able to do. Even NVIDIA's older Chip Hopper is 25x slower than the upcoming Blackwell architecture. No imagine this inferior chips from Groc? who is going to buy it and use it? Calculator making companies? Maybe phones and laptops at tops! Because they are definitely not going to be good enough for the data centers. And that all also only then, if they succeed to develop anything! These money hungry guys need to go a long way. it does not seem to be profitable what they are attempting to do. First come TSM and ASML, then come NVIDIA, then AMD, then All mega caps, then all other Broadcom, Micron etc. companies.. and in the end they come, and say, we will make a chip too. Good, do it ! good luck!
No, because peo0le are actively against that very integration, we don't want it on our phones, we don't want it to replace apps, it IS an app, and that's how it should be and should remain.
so the top tech CEO is using a dial up modem and webcam from the 90s 🧐
Correct , as he is using old microchip tech 14nnm😅
he prefers using the old Nokia function phone to safeguard his privacy.
This is a new ai chip.
Hahaha
Lol 😂😂😂😂
Maybe they talked with his microwave oven with a built-in new chip 😂
8:00 "wee'ree allllreaaddy thhhe wooorlds faaasteest" - worlds slowest speaker
😂😂😂😂 that part happened right on queue. you cant make this up LOL
Hilarious. I'm not convinced.
😂😂😂👏
had a great laugh :D
He looks and sounds like an AI bot
As someone who's been a computer nerd my entire life and now over 60, I've seen tech hype come and go. When it comes to AI, companies like Nvidia and startups like Groq are really just selling faster chips-just like how they update phones or video cards (GPUs). It’s all part of the same cycle. But here’s the reality: Artificial Intelligence isn’t about how fast your hardware is. AI's real power comes from the large language models (LLMs) and how efficiently the software interacts with them. The hardware is just the vessel. The real breakthroughs are happening in the way we develop and access these models, not in speeding up chips. Don’t get caught up in the hype!
1NM chip refers to the size of the chip. The smaller the chip, the more expensive it is. However, a 1NM chip does not mean that it can run faster than a 5nm chip cluster. If the size is larger, more 5NM chips combined can run faster than a 1NM chip. This has to be considered from both cost and application aspects. For example, a missile is larger than a mobile phone, and a 14NM chip cluster is enough to develop AI weapons. In addition, weapon chips also need to be stable and resistant to high temperatures, etc.
@@lijackson-x6r While you're right that smaller chip sizes (like 1nm) don’t always equate to faster performance compared to clusters of larger chips (like 5nm), it's important to remember that AI performance isn't just about the hardware size. The real power in AI comes from how efficiently the software interacts with large language models (LLMs). Whether it's AI in weapons systems or consumer electronics, the key is how well the software leverages the model, not just how small or fast the chip is. The hardware is the tool, but the real intelligence is in the software.
@@Progressive_CanadianI completely agree with this view. AI requires 5G or 6G network infrastructure. Without network speed, no matter how fast the chip is, it will be useless. Moreover, the robotics industry is the foundation of AI applications. Without mechanical engineering and industrial foundation, AI cannot bring anything substantial to people. For example, chips alone cannot bring food to the table. At present, AI is in the early stages of development, but China's industrial foundation and network foundation make it easier for AI to become a reality.
@@lijackson-x6r
I appreciate your perspective, but I’d push back a bit on the idea that network infrastructure like 5G or 6G is the critical bottleneck for AI advancement. While network speed is important for certain applications like autonomous vehicles or cloud-based AI processing, it’s not the main factor for AI progress. The real driver of AI's potential lies in large language models and software development, not just hardware or network speeds. As for robotics, you’re right that the industrial base is crucial, but AI’s impact can be felt even without physical machines-think of software-based solutions in healthcare, finance, or creative industries. China may have a strong industrial foundation, but AI is advancing globally, thanks to breakthroughs in software and model development, not just hardware infrastructure.
@@Progressive_CanadianAI software development is important, but AI models must collect a lot of data, all of which comes from the Internet. Only with data can AI models be analyzed and trained.
AI language models are only a small part of AI applications. Many AI applications are in industry and medicine. For example, robot surgeons rely on supercomputers in the cloud to perform surgery, so Internet speed is crucial. In addition, remote mining and unmanned port loading and unloading of goods all require fast Internet connections. In addition, satellite image analysis, the data from satellites is very large, and it requires a fast network connection to transmit data and perform image analysis to identify farmland or oil fields or military missiles. Especially in military applications, AI can analyze the trajectory of the opponent's missiles and intercept them in time, which requires fast network transmission.
I love how people write on Facebook, Twitter or even youtube about their concerns about privacy. The irony is not lost.
This report on the battle among AI chip makers for dominance is fascinating! It highlights the intense competition in the tech industry and how crucial these advancements are for the future. Thanks for shedding light on this important topic!
This has been hugely important for like 2 years, very late to the party BBC but I'm glad you've finally chosen to attend.
What a great report; they discussed technology, leading AI-Companies, their dependance on Taiwan, and economic and political risks associated with it. Comprehensive work. The compares are really well spoken!
Jonathan Ross 😂😅 honesty thought ... never mind.
"Nah it couldn't be could it?" I thought to myself 😂
This is not how you pronounce NVIDIA :)
Yeah, some of the commentators on BBC also insist on using Trump's racist pronunciation of Kamala as well. They're not the best and brightest.
There exist other much more common words pronounced dif between English speaking countries, let alone regions. NVDA and Kamala are not even English words, and even a linguist would suggest pronunciations and even meaning are fluid. That said, I had the same thought, but rather questioned the pronunciation I’ve come to commonly know. Bye
is it cheaper? how about cost of operation as well as maintenance ?
I was hoping you were going to talk about the technical differences of the Chip designs. But I was disappointed.
If groq is only focusing on AI inference, it does not compete with Nvidia as Nvidia GPU is more for AI training task.
Nowadays more and more CPU has integrated NPU which is meant for accelerating AI inference as well. Groq seems to compete with these integrated NPU and trying to be a better NPU.
There are many AI inference hardware for different applications. Those hardware cannot use for training deep model.. Therefore, Nvidia is still a true king for training.... One DGX is almost equal to a house price..!!
The middle class is already artificial intelligence.
No chips needed.
Fr.
Artificial? Lol
Priya is really great, she has absolutely zero background in tech and she dared to dream big. Such an inspiration, especially to young woman. Need more role models like her to be pushed to the forefront of schooling.
NVIDIA CUDA platform reminds me of the history of RISC and CISC in computer organization and why ARM became so successful.
You should check out Cerberus also...
Please use fast internet before jumpping onto fast AI chips
Everybody talks about artificial intelligence nowadays.
This is embarrassing …. Liked the anchor’s comment “guessing you chips are not in this communication.. your face is frozen”😅
It is almost sad how they get scammed every single time.
Is this video a groq commercial?
Untuk pemain pasar lama mempertahankan eksistensi perusahaan menjadi hal yang sangat efektif dalam menghadapi hadirnya para pesaing baru , disini naik turunnya tekanan pasar sudah menjadi pengalaman yang menjadikan perusahaan tampil semakin terpercaya oleh konsumen tinggal bagaimana pengembangan selanjutnya
what the heck kind of camera did this CEO use a toaster 😂😂😂
8:00 has ot to be the funniest moment of irony I have ever seen, thank you, that made my day, I genuinely laughed out loud
Sounds like science fiction but yet it's something happening right before our very own eyes. Just think about the jobs🤦♂️
It's a shame that here we have a video on AI, and they don't know the difference between AI and machine l;earning. What we currently have operational is machine learning. It is a start towards maybe one day having true AI. True AI is still only a goal, at this point.
It's okay, it's BBC producing content for 60year olds.
It will never happen, on off switches will never out power the human brain.
First of all machine learning is a subset of the field of artificial intelligence. Also If you want to get more technical the subsets of machine learning being used for LLMs are Deep Learning and Reinforcement Learning.
Using the words “Artificial Intelligence” is just fine. It makes sense to use those words since “machine learning” is not as familiar to the public because it’s relatively new in comparison.
@@2handsandwiches That assumes that computing will always be on/off constrained.
However, you may be right. It could be that machine learning will get closer and closer, with advances taking longer and longer. And actual AI might not ever happen. But machine learning might get to the point where it's almost impossible to tell the differencew.
I envision a new set of PHDs will emerge that deal specifically in how to define and measure when/if machine learning has finally become actual AI.
As we have seen in the past, it is many times foolish to say something will NEVER happen. After all, we have only been doing eletronic sudo-brain type work for less than 100 years. What will happen in the next 1000? 10,000? Well, if we manage to not destroy ourselves first.
@@timetraveler_0 Oh, good, it doesn't apply to me, then. I am 70. ;)
Laughs in open source. No one wants this stuff on their personal devices. The battle is for which big tech company will fail first.
This
SBGAI
What about arm,apple,qualcom or cerebras ❤
We want supersaturation. Supersaturation is a metastable state, meaning it is not in equilibrium and can easily revert to a saturated state. BTW thank you for reminding me this a Talent War & not the
Pretentious Inn.They will leave you!
The AI chip wars are heating up! 🔥 Nvidia's dominance is being challenged by startup Groq! 💪 I love seeing innovation and competition in the tech space! 🤖 It's exciting to think about the advancements that will come from this battle for AI chip supremacy! 💻💸 Thanks for sharing, BBC News! 👏 Stay ahead of the curve! 🚀
So Groq doesn't do AI learning, they only do AI inference, Interesting so do they just use publicly downloadable AI and automate it?
lmao, the saudis can't even provide a dial up quality internet connection and they think that they can compete in the AI race... xD
One day my cloth broke zip handling part, so to close or to open the cloth, somehow the zip keep on zipping part of the cloth and get stuck and cant open, so i have to remove the stuck piece for i to proceed to open only for shortly for it to zip the cloth once again, what does this have to do with artificial intelligence, the cloth is analogy and doesn't have any brain chip. Be patient.
Then one afternoon as i was rushing home under the sun, the temperature inside close cloth was at a boiling point so i struggle like usual to unzip the cloth, suddenly i hold both ends of the cloth and pull them away gentle, it open so smoothly without further struggle, few days latter in the morning i hard to unzip once again and i struggle as i was to struggling anyway for a long time as i can remember, suddenly i recall the pull away strategy i discover the other day and i did just that. From then hence forth i learn from nothing to how to unzip without struggling.
Anyone agree am conscious and the way i learn is exert the same way deep learning algorithm of artificial intelligence learn. So many example feed on to the algorithm with slit changes each time on solving the issue at hand till this tinny solution amount to something slightly useful that the whole brain now moves towards this trend further in the solving of the issue at hand, from that we get AI chip being correct at solving things accurately at 5% on its own, latter it now at 98% accurate as of now deep learning system are at.
Looking deeply why did my brain figure it out just the same way an artificial algorithm would? To sum it all the weights used on deep learning once mapped by the algorithm from sum less biases then name the object or the function will remain so as long as the weights and biases remain constant. That mathematical sequence remain constant indefinitely as long as the weights and biases remain constant.
From that we can conclude the idea of our human brain has the element of weight and biases in its logic configuration. To that endeavour its valid to state we have a risk of underestimating what conscious is all about, i am conscious yet i see deep learning not conscious when the fact of the matter is we do share how we learn from being feed to much data and we learn from that, look how many things were discovered by accident from penicillin to aerial dynamic from the two brothers.
We cant have two system performing sequence the same way then we go ahead and claim one is 100% conscious while the other 0% conscious, its logic to equate at list a 5% chance of conscious, to a maximum of at list 20% conscious. Once science community agree on this number bet it will grow each day each second to levels exceeding a million percent (1,000,000% conscious ) something we cant even dream of or speculate of.
Is that guy they’re interviewing an AI of the human owner?
I thought that was a pic of Sam with Grok 😂😂
Looks like he is calling via dial up internet while explaining groq to be broadband of chips. 😂
I love you guys, but I found it hilarious that on a show about faster chips, the CEO’s film (zoom, I assume) is like on dial up. The BBC host noted it too. LOL. And yes, I realize that that’s due to a different thing rather than his type of chip. Still…hilarious.
Nuh-vidia is crazzzy
Never any AI in my traditional town.
only on your smartphone. And tablet.....and youtube,,,
Ahkmed and Imran if it's Bradford.
He looks like an AI generated image
Saudi Arabia might house the datacenters but if this video call's quality is the extent of their Internet connection, those datacenters won't be able to serve the world.
Nothing works in heat. Spend an other billion dollars. Just like underwater sea resort and tunnels. Who has an ac to work below specifications. Cellphones are now in refrigerator calls. It is called fridge mobile.
This guy looks and sounds like he is an AI.
AI has yet to be anything more than a waste, inconvenience, novelty or annoyance. Technology was meant to free us but it is doing the opposite.
Tell that to those that have lost their jobs recently, due to AI.
@@StuartJ that sounds like an inconvenience and annoyance that didn't free those people.
@@saintsalieri yeah, but it wasn't a waste, inconvenience, novelty or annoyance to the company that's saving those salaries.
@@StuartJ but does that company know how to ligma
AI robots are designed to do hazard or inconvenient jobs for human, and they turns out to be artists, music composer, photographic video creator, finance analysts , and eventually put human out of jobs. fortunately, the lawyers are not needed in the future because AI lawyers will be 10 times better than human lawyers to win the lawsuit, and politicians are not needed because AI will make wiser decision for mankind.
According to my shares the market is shit mow 😂
CEO of Groq has the worst connection or hardware ever? That's not a good sign . . . Didn't they test it before the interview?
Heh. Remember dial-up? Er, yes. 9600, 14.4, 56 and we were told it wouldn't ever get faster, but they lied! Don't miss those days or the screaming modem sound!
Other uses for these GPUs?
As graphics cards for PCs? 🙂 Preferably with throttling or some hardware-based deterrent which can't be disabled to discourage cryptomining with them so the prices won't skyrocket after they're put on the market . . .
Better auto-translate for TH-cam videos? 🙏
15:00 Talking about the United States supplying supercomputers to the Soviet Union. The cool thing about supercomputers that most don't know? Almost all of them are running on a Linux kernel! 🙂
15:35 Gregory Allen of CSIS (Center for Strategic & International Studies). Hypersonic, for those who don't know it, yet, is Mach five to ten. Anyone remember the X-15? The thing is with these higher-speed missiles is that detecting, targeting and taking them out is far beyond human capability (AKA the human operators or the human component), so that has to be done another way. Even Samuel Renshaw (he had some nice study tips in the end of the Psychological Optics paper he wrote) couldn't train 'ack-ack' (anti-aircraft) gunners to be that fast with his tachistoscope! (Fans of Robert A. Heinlein probably remember characters mentioning Renshaw in his works such as Stranger in a Strange Land (Jubal Harshaw's secretary, Anne, was a Fair Witness who was 'Renshaw-trained'), Gulf (VERY sad ending, but, those who read it will understand his character Friday a little better) and Citizen of the Galaxy.)
17:35 China engaging smuggling networks. Perhaps similar to multihopping with a VPN, but with more than one extra point, with the idea being the more used, the better, provided they're not the weak link types who snitch/squeal/rat you out/be a stool pigeon/sing like a canary/stool with the rogue, et cetera . . .
17:45 Samsung executives giving manufacturing secrets to others. Can you say NDA violation? I knew that you could! Another thing to be wary of is a brain drain scenario or 'poaching' by governments or competitors or reverse-engineering technology. Not to be confused with Operation Paperclip, which was a totally different can of worms . . .
19:50 Taiwan being a crucial resource for logic chips manufacturing and China claiming it as a territory (not in the way of a Risk game territory). Things like this are worrisome, as centralizing such valuable manufacturing in one place makes them highly vulnerable and might possibly lead to a scenario where someone might threaten it, stating that either they'll be allowed to have it and control it and how much they can export and to whom or they'll reduce or eliminate their manufacturing capacity completely (the 'if I can't have it, nobody will' bit). Example: Paul Atreides AKA Maud'Dib with melange AKA spice on Arrakis, threatening all the spice by using the changed Water of Life (via a Bene Gesserit 'witch' like his mother, Jessica) on a pre-Spice mass that would then continue on in a ripple effect to the rest of the planet). And, yes, I probably read too much, especially science-fiction! LOL
jimmy carr?
9:43 Wow!!! this company is modern! did you listen guys? they do speach to text as well, like what google did 10 years ago and also siri did few years ago!
It's, infact, very difficult to be great ! But it's much more difficult to be a good human being ! In these days of fast - degrading human values, the call of the hour is - a spurt in the number of sane and right - thinking people - in the greater interest of - sustenance of the Humankind and Humanity !
yeah yeah again some big companies will take control of everything, I don't want them to have control. we need to have the control, make chip which normal people can buy and do stuff, this shit is really like game of thrones
"A.I" aka Automated Stupidity.
Congrats worlds half ai chips and data centers to come out of arabia❤❤🎉🎉
En videa dude.
This is game changer I Love ❤️ AI the Ultimate equalizer between those who have and have not the vision of the future this is my religion with AI I'm the king of my castle if you are poor and marginalized just master the AI in any field you wish and you can lead the leaders.
I wonder if the news reader likes kids 😂
Forget chip. WHY AREN'T YOU REPORTING ON ROTHERHAM?
More than the problem of processing power is the problem of intelligence. If you call pattern matching intelligence, I guess you are satisfied. If you expect brilliance from AI, you will be disappointed.
When they suggest that this technology could write a book in seconds, I'm afraid you will be disappointed if you expect anything more than Dr. Seuss (i.e., see spot run). If it really could, then why did they not take the time to give an example?
It will be some time before these AI systems can ever match human intelligence no matter how much processing power they are given.
Consider this: AI systems are restricted on their ability to learn. Humans are very much afraid of building AI systems with true intelligence.
Lol, high tech big shot with slow internet and bad webcam. Once he opens his mouth it is only about money. It is alarming that he gives a feeling of “one blood can detect cancer” type of scam 😂
😂😂😂😂this matrix is crazy 🤣🤣
Non of this is real
It's a battle for who's going into trash cans everywhere tbh.
For a very short of period time, advanced AI itself can create its own choices for them very soon what showing in this is old compared with new fiber Chipes and infinity chips made from energy itself these guys are far behind
0:33. NVIDIA is making $120 Billion plus per year. And this Groc 'Raised' 1 Billion $ now. Nvidia's technology is 10-15 years in Future, and these guys are collecting pennies in hope of developing any low-quality AI-chip, which would dream to perform half as good as Nvidia's Chips, which they won't be able to do.
Even NVIDIA's older Chip Hopper is 25x slower than the upcoming Blackwell architecture.
No imagine this inferior chips from Groc? who is going to buy it and use it? Calculator making companies? Maybe phones and laptops at tops! Because they are definitely not going to be good enough for the data centers.
And that all also only then, if they succeed to develop anything! These money hungry guys need to go a long way. it does not seem to be profitable what they are attempting to do. First come TSM and ASML, then come NVIDIA, then AMD, then All mega caps, then all other Broadcom, Micron etc. companies.. and in the end they come, and say, we will make a chip too. Good, do it ! good luck!
AI is becoming more and more integrated into every aspect of your smart device, is there a point where AI will replace Apps?
No, because peo0le are actively against that very integration, we don't want it on our phones, we don't want it to replace apps, it IS an app, and that's how it should be and should remain.
Monday 🌎hello❤️
Oichi mama channel
Not trusting my money with a grown man who uses filters for a interview
lol doesn’t matter about u, ur not buying an AI mega server in the of the desert, u will use the ai tho ;) or, failing that, it’ll probably use u
@@Macarite you aren’t using AI to spell check and you seem to be doing fine.
It's worthless and we must get rid of it.
Whys this greg down playing the danger of AI. It's a lot more then running outlook. ridiculous.
Martinez Donald Miller John Lee Amy
Hey wanna make money, i have stock on ai firms...😐😐😐
I have never heard anyone pronounce NVIDIA like that, you can tell this guy had zero exposure to tech 😂
Speaking of AI… AI should not be regulated, regulations do more harm than good AI should remain unregulated.
🎉🎉🎉🎉🎉
Here to leave my dislike
this guys internet connection is a Troll 😂
They can't compete with China, they have Sarah AI : younger and beautiful!
so asian music for jensen. racist buggers
My sincere thanks for sharing it.
Peace prosperity and progress are people's pride and property, please participate.
ಜೈ ಹಿಂದ್ ಜೈ ಭಾರತ್, ಶ್ರೇಷ್ಠ ಭಾರತ್.
❤❤❤❤❤❤❤
Hahaha! FAaaaaaaaaassssttttttttteeeessssttttt,
ai is trash!
Absolutely right ..
..
Its just the connection of a.i and stock market..
Manipulating it in some way or the other 😂😂😅
I was hoping you were going to talk about the technical differences of the Chip designs. But I was disappointed.