I am sooo happyyyy India me aise ideas, businesses and innovations ko push mil raha hai. Hats off to the team scaler for coming up with this! I am in same industry which revolves around Media and marketing, I will be an engineer by next year, I understand the number of resources it costs and it's not easy to bet on the outcomes in terms of no.s but it might change the course lives of many young enthusiasts.
As an AI engineer myself, consider this as an absolute joke .... deep distributed technology stack exists like nvlink, infiniband to reduce distributed latency on multi node clusters. AI is all about latency minimisation. even thinking about switching infiniband over a p2p distributed network of devices is an absolute joke. Forget 5G, 6G 7G it's never possible. Infiniband latencys are 400GB/s and the AI training intense to use Max performance of it while training any sort of AI models on a distributed node
Not saying the so called 'idea' is worth how much or feasible. But I love students thinking on several direction, good to see the buzz is helping some genzs to at least think rather than scrolling ig. Nevertheless, I guess this has a huge scaling issue, Not sure how they are coming up with a fed learning machine for secure scaling and what Babbar and Mohak is doing there
honestly, idea is feasible. If the user can decide whether they want the website to let use their computing power, or see ads instead of it. And about adblockers, many websites these days will prompt you to disable adblocker and not let you access the page. So it is viable
I was going to say this is a waste of time but I guess in the end it's a net positive for audience and the participant. Actually I am not sure if companies are really searching for ad monetisation alternatives. I think it's more about finding more subtle ways of showing ads. And the idea might have a better scope if it's used like a tool for developers, like distributed gpu computation.
Amazing, it is really useful when there is a silicone shortage, and I think it will be better if we allow the user to set the threshold of operations, capacity to be used, power, and stuffs by the GPU of the device. Great idea man. Dami xha brother ❤.
great that school kids are thinking in that direction .. but the difference in scale of openai's gpu and my phone gpu is just gigantic.. you cant compare the two. Also these models require a lot of ram Llama 7b takes 8gb ram if i remember correctly .. sure openai's models are a lot larger .. and also after that SPEED , sam altman will commit suicide waiting for computations to complete.
After watching the episode I got to know that Love Babbar doesn't know much about tech... Mohak Mangal has been called in for views. Subhash's knowledge is good and but communication skills are not. Anshuman Singh is good with a lot of things and is also a good human being
Subhash struts around RUclips like he's the king, giving off major Ashneer vibes with his 'too smart for everyone' attitude. Sure, you might get famous and even do some good work, but respect? That’s something you'll never earn with that arrogant demeanor.
✅ Apply Now to Join Scaler School of Technology!
👉 Click here to apply:scalerschooloftech.com/3Ziilx1
this concept already exists, its called federated learning, developed by one and only Google, Fl has limitations
I am sooo happyyyy India me aise ideas, businesses and innovations ko push mil raha hai. Hats off to the team scaler for coming up with this! I am in same industry which revolves around Media and marketing, I will be an engineer by next year, I understand the number of resources it costs and it's not easy to bet on the outcomes in terms of no.s but it might change the course lives of many young enthusiasts.
Bro is making a video against Google and uploading it to Google's RUclips platform 👌👌
That's what monopoly is
Superb, Program This kind of Insight really helpful in adding value in the area of Technology. Keep exploring
Google ko koi farq nhi padta inn chindi chijo se😂
As an AI engineer myself, consider this as an absolute joke .... deep distributed technology stack exists like nvlink, infiniband to reduce distributed latency on multi node clusters. AI is all about latency minimisation.
even thinking about switching infiniband over a p2p distributed network of devices is an absolute joke. Forget 5G, 6G 7G it's never possible. Infiniband latencys are 400GB/s and the AI training intense to use Max performance of it while training any sort of AI models on a distributed node
bro he is in 11th and look at his curiosity rather than criticise
Super confident speech 👍
Bro there is Cloud Computing
Not saying the so called 'idea' is worth how much or feasible. But I love students thinking on several direction, good to see the buzz is helping some genzs to at least think rather than scrolling ig.
Nevertheless, I guess this has a huge scaling issue, Not sure how they are coming up with a fed learning machine for secure scaling and what Babbar and Mohak is doing there
Background ki awaaz thodi or badha do bhai log
Amazing idea, I had a similar idea but for storage instead of gpu.
Nice Man, very good Solution❤❤
I didn't get my result yet ,my test completed date 19may
honestly, idea is feasible. If the user can decide whether they want the website to let use their computing power, or see ads instead of it. And about adblockers, many websites these days will prompt you to disable adblocker and not let you access the page. So it is viable
I was going to say this is a waste of time but I guess in the end it's a net positive for audience and the participant. Actually I am not sure if companies are really searching for ad monetisation alternatives. I think it's more about finding more subtle ways of showing ads. And the idea might have a better scope if it's used like a tool for developers, like distributed gpu computation.
mohak bhai ki 2 unnicorn startups hai and tech me toh bohot knowladge hai bhai ko.🤣🤣
Amazing, it is really useful when there is a silicone shortage, and I think it will be better if we allow the user to set the threshold of operations, capacity to be used, power, and stuffs by the GPU of the device.
Great idea man.
Dami xha brother ❤.
Marketing manager 😁😁
5 rupees ka ashneer Grover dekha 😂
Awesome idea, keep it up
Nice 👍🏻❤
great that school kids are thinking in that direction .. but the difference in scale of openai's gpu and my phone gpu is just gigantic.. you cant compare the two.
Also these models require a lot of ram Llama 7b takes 8gb ram if i remember correctly .. sure openai's models are a lot larger ..
and also after that SPEED , sam altman will commit suicide waiting for computations to complete.
They will be distributing these resources over seeds on a large scale
After watching the episode I got to know that Love Babbar doesn't know much about tech...
Mohak Mangal has been called in for views.
Subhash's knowledge is good and but communication skills are not.
Anshuman Singh is good with a lot of things and is also a good human being
Good idea but it's highly unreliable and the system design would be a walk through hell also one will choose google due to its reliability
Subhash struts around RUclips like he's the king, giving off major Ashneer vibes with his 'too smart for everyone' attitude. Sure, you might get famous and even do some good work, but respect? That’s something you'll never earn with that arrogant demeanor.
Sounds like pied piper
This is like shark tank for silicon valley
Why silicon valley? Why can't you people come out of that colonial mindset?
@@rosepainting8775 bruh what
@@theepicman8160 colonial mindset never heard of it?
@@sad_man_no_talent idk bro, I dont engage in these political stuff, I just said silicon valley coz it just came to my mind at that moment
@@theepicman8160 so ignorant, that's why u don't know why US kills innocent people
silicon valley webseries dekh ke aya hai ye banda
New face of utorrent