As someone who works in the AI industry, it’s literally impossible to not use GPUs. I also think AI start ups are a terrible investment. It’s very difficult to monetize if you don’t have extremely large models like Meta or Open AI. Otherwise they’re just as useful as that rabbit thing that came out a few weeks ago.
Agreed, Any Investment needs to be planned and analyzed for cost reduction, potential profit. I understand if you fail to act your company will suffer, but target how AI can add most benefits and value. A company buying AI because of FOMO is reckless.
@@hwy9nightkid The key word here is "yet". I absolutely agree with you though... but if the stock price increases continue for 1 or 2 years lots of companies will be staggeringly overvalued
@@hwy9nightkid Agreed. That is the scary part. It may keep growing for a while. It took until March 2000 for that internet bubble to pop after several years of "growth", where we kept hearing that the previous valuation models no longer worked no more. Just take CSCO as an example.
It's not a mystery. Do your homework. AI is being used right now to improve weather predictions, drug discoveries and design, tremendous improvements in self driving cars, military applications, services, and the list goes on. Even NVDA uses AI to design and optimize their chips. Without AI, they would not have been able to design the latest Blackwell chip and system. AI will also be used to design the next generation chips too.
I get a feeling that AI is mainly being hyped up by the media more than anything. VCs are just trying to ride the hype train. There are very limited use cases of the current crop of “AI” models. They can surely perform very basic tasks but pushed too hard and they will falter.
It takes around a good two weeks to develop a good re-verse engineering AI GPU Nvidia chip for a startup company!😊 It takes Intel a good 3 years to do that so they alway 3-4 years behind all the time!😂 AI startup should just make their own!😊
Now, imagine having that training be done as an open source project instead of everyone doing them in their expensive garages... Your VCs reward idiotic business models
It’s led to a lot of advancements in medicine and chemistry, like the development of the cancer drug AOH1996 which is currently in human trials. Fights solid tumor growth. It’s also been used to find better battery chemistries that use less lithium.
I love the nay sayers. Instead of finding gold… we invented a type of gold that keeps growing spontaneously on its own, and you’re surprised at $50billion investments? lol
greed, ai scams, overinflated saleman for AI. so far AI has not lived up its hype. just an idea only and there are so many people want to share that limited & dwindling cake. spend $50 billions operating cost on some hardware and only profit $5 billions gross profit. 10:1 ratio ? bad bussiness model!
Jensen seems like a great ceo, but at the end of the day hes just out there saying whatever it takes to shill his wares . Disappointed to hear him claiming tesla is far ahead in autonomous driving tech considering reality.
Nvidia and Tesla were joined at the hip when the Model S came out. Right now no other car manufacturer is making their own AI neural network based chips. Everybody is relying on Nvidia, Bosch, Aptiv, MobileEye, Schaeffler and such. They've all been advertising cars having IOT 2-way communication for over 5 years but nobody had deployed it on a single production vehicle. I would love AMD to be able to get a leg in the door but Nvidia is running 2 lengths ahead with its CUDA marketshare/software.
@@Tokamak3.1415 I can't argue with Nvidia being ahead with their cuda adoption. My main issue was with Jensen's claim that Tesla is ahead in av tech when they have yet to deliver anything beyond L2 driver assist. Tesla's dojo chips utility is questionable considering they're loading up on Nvidia chips instead, and waymo utilizes Google's home grown tpu's so Tesla is not the only custom ic game in town. Obviously I don't know what musk is showing Jensen behind closed doors so maybe there's some kind of breakthrough coming, but given public data his claims are suspect.
@@PreschoolDropout673 as CEO he gotta shill his products even if they half working. it's also a reminder to never believe what companies CEOs says. lie for stocks money is all they care about.
You are not the one who own the company and some of the a.i like chatgpt already free it just that if u want to use the premium version u have to pay don't u know how much money they have to use to maintain those data center?
@@HungPham-zi9ks fraction of the profits they get. once you have a data centers you already foot the bill of the hardware and infrastructure. the only thing you will pay are electricity and technical team to operate the business, maybe some money here and there for replacement of broken equipments or installing something new.. etc keep in mind the data centers purchase are in bulk and the price for each piece of hardware is may may mire cheaper than if you buy it as consumer. these companies already have discounts too, a huge one for the bulk purchases of equipment.
If You can’t buy A5000 Nvidia cards for servers and hire people for maintenance then forget AI because it’s expensive and not everyone use it. It’s a leisure not a necessity.
As someone who works in the AI industry, it’s literally impossible to not use GPUs. I also think AI start ups are a terrible investment. It’s very difficult to monetize if you don’t have extremely large models like Meta or Open AI. Otherwise they’re just as useful as that rabbit thing that came out a few weeks ago.
Agreed, Any Investment needs to be planned and analyzed for cost reduction, potential profit. I understand if you fail to act your company will suffer, but target how AI can add most benefits and value. A company buying AI because of FOMO is reckless.
They are not all making LLMs and are working at specifically targeting single tasks which doesn’t take nearly the same power
Spend 50B on chips to make 3B Rofl..Ouch
🤡
That feeling that we are re-living the late 1990's internet bubble just keeps rising...
its no where close to that yet
@@hwy9nightkid The key word here is "yet".
I absolutely agree with you though... but if the stock price increases continue for 1 or 2 years lots of companies will be staggeringly overvalued
At least the hype about the capabilities of AI is mostly dud.
@@hwy9nightkid Agreed. That is the scary part. It may keep growing for a while. It took until March 2000 for that internet bubble to pop after several years of "growth", where we kept hearing that the previous valuation models no longer worked no more. Just take CSCO as an example.
yeah the internet never amounted to anything :)
Not more GPU, but smarter algorithm
So how exactly are companies making money from AI. Who are consumers buying AI generated products or services. It’s still a mystery.
It's not a mystery. Do your homework. AI is being used right now to improve weather predictions, drug discoveries and design, tremendous improvements in self driving cars, military applications, services, and the list goes on. Even NVDA uses AI to design and optimize their chips. Without AI, they would not have been able to design the latest Blackwell chip and system. AI will also be used to design the next generation chips too.
Apple fanboys/fangirls!😊
Sure sure, GPU alternatives. Yes, yes, AI that is less computational heavy. great ideas. yall know what you are talking about. for sure.
Deirdre Bosa
I get a feeling that AI is mainly being hyped up by the media more than anything. VCs are just trying to ride the hype train. There are very limited use cases of the current crop of “AI” models. They can surely perform very basic tasks but pushed too hard and they will falter.
When investors pressure, how much will security concerns be ignored eventually?
It takes around a good two weeks to develop a good re-verse engineering AI GPU Nvidia chip for a startup company!😊 It takes Intel a good 3 years to do that so they alway 3-4 years behind all the time!😂 AI startup should just make their own!😊
Her figures are off and severely inaccurate
oh, i'd say her figure is perfect....
Google cloud tpus
Why are you not considering Qualcom considering Microsoft push for the edge?
Qualcomm are not near the quality and powers of nvidia in hardware and software. nvidia has monopoly right now.
Revux is being mentioned everywhere - definitely a project to watch!
These scam bots
Now, imagine having that training be done as an open source project instead of everyone doing them in their expensive garages... Your VCs reward idiotic business models
explain?
@@maxrinehart4177 should I explain how open source works?
do we even need AI?
Where have been living in these past 2 weeks?
Massive announcements from giant tech companies
No, we don't. Don't use AI.
You can live without a car though.
It’s led to a lot of advancements in medicine and chemistry, like the development of the cancer drug AOH1996 which is currently in human trials. Fights solid tumor growth. It’s also been used to find better battery chemistries that use less lithium.
Dei. ❤❤❤❤❤❤❤❤❤❤❤❤❤❤
Ur lookin mighty fine Deirdre hun u goin on a date ❤
☺️☺️☺️
😀
I'm all in on Revux. Presales have the highest returns, and this one’s gold.
Scam…
This small startup is stupid and lies
You sound like a 6 year old lol
I love the nay sayers.
Instead of finding gold… we invented a type of gold that keeps growing spontaneously on its own, and you’re surprised at $50billion investments? lol
I never should have doubted Nancy Pelosi and bought Nvidia
greed, ai scams, overinflated saleman for AI. so far AI has not lived up its hype. just an idea only and there are so many people want to share that limited & dwindling cake. spend $50 billions operating cost on some hardware and only profit $5 billions gross profit. 10:1 ratio ? bad bussiness model!
Jensen seems like a great ceo, but at the end of the day hes just out there saying whatever it takes to shill his wares . Disappointed to hear him claiming tesla is far ahead in autonomous driving tech considering reality.
Nvidia and Tesla were joined at the hip when the Model S came out. Right now no other car manufacturer is making their own AI neural network based chips. Everybody is relying on Nvidia, Bosch, Aptiv, MobileEye, Schaeffler and such. They've all been advertising cars having IOT 2-way communication for over 5 years but nobody had deployed it on a single production vehicle. I would love AMD to be able to get a leg in the door but Nvidia is running 2 lengths ahead with its CUDA marketshare/software.
@@Tokamak3.1415 I can't argue with Nvidia being ahead with their cuda adoption. My main issue was with Jensen's claim that Tesla is ahead in av tech when they have yet to deliver anything beyond L2 driver assist. Tesla's dojo chips utility is questionable considering they're loading up on Nvidia chips instead, and waymo utilizes Google's home grown tpu's so Tesla is not the only custom ic game in town. Obviously I don't know what musk is showing Jensen behind closed doors so maybe there's some kind of breakthrough coming, but given public data his claims are suspect.
@@PreschoolDropout673 as CEO he gotta shill his products even if they half working. it's also a reminder to never believe what companies CEOs says. lie for stocks money is all they care about.
He said that, then a few days later, news come out Elon raised 6B, then nvidia stock goes up. it all makes perfect sense
The answer is Cerebras!
Apparently OpenAI is a big customer of Google cloud TPUs
Need the hardware to be more affordable, and no monthly subscriptions for ai, theres already enough bills as it is
You are not the one who own the company and some of the a.i like chatgpt already free it just that if u want to use the premium version u have to pay don't u know how much money they have to use to maintain those data center?
@@HungPham-zi9ks fraction of the profits they get. once you have a data centers you already foot the bill of the hardware and infrastructure. the only thing you will pay are electricity and technical team to operate the business, maybe some money here and there for replacement of broken equipments or installing something new.. etc
keep in mind the data centers purchase are in bulk and the price for each piece of hardware is may may mire cheaper than if you buy it as consumer.
these companies already have discounts too, a huge one for the bulk purchases of equipment.
If You can’t buy A5000 Nvidia cards for servers and hire people for maintenance then forget AI because it’s expensive and not everyone use it. It’s a leisure not a necessity.
Didn't Microsoft just come out with a new NPU Processor?
Only good for inference, not training.
Shifting my portfolio - heavy on BTC and Revux, with a sprinkle of DOT and ADA.
No one asked you
I see Revux doing 50x, maybe even 100x after it goes live on major exchanges.
revux is a scam being promoted by bots!
Is this the next scam?
lol
Got a strong feeling Revux will go 100x once it hits Binance.
Revux is dominating my crypto chats - seems like the next big thing!