• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Gamer Nexus - BLACKLISTED by AMD | AMD's Dirty Tactics

Nvidia investement in CUDA/AI are more than all paid off already, we already know the answer, would they make $0 dollar starting tomorrow it was all more than worth it. Last year net income was 120 billions, that covered all the R&D they ever did in this, let alone the incredible 2023-2026 fiscal run combined, 77 billions in direct stockholder returns in the last few years plus built a 60 billions+ war chest despite that while removing all long term debt concern during it, ammassed giant R&D in things from networking, power, packaging, cooling, optic networking that will translate to many others things right away.

There is question and risk for many people in that space, not for Nvidia about having been worth it or not. As a bonus, no one has yet catched up fully to them in gaming GPU during that stretch anyway, not in market share or hardware or software, so not even much of any downside/sacrifice has been made yet with it much really anyway for it to be a question.
 
Last edited:
LukeTbk said:
Nvidia investement in CUDA/AI are more than all paid off already, we already know the answer, would they make $0 dollar starting tomorrow it was all more than worth it. Last year net income was 120 billions, that covered all the R&D they ever did in this, let alone the incredible 2023-2026 fiscal run combined, 77 billions in direct stockholder returns in the last few years plus built a 60 billions+ war chest despite that while removing all long term debt concern during it, ammassed giant R&D in things from networking, power, packaging, cooling, optic networking that will translate to many others things right away.

There is question and risk for many people in that space, not for Nvidia about having been worth it or not. As a bonus, no one has yet catched up fully to them in gaming GPU during that stretch anyway, not in market share or hardware or software, so not even much of any downside/sacrifice has been made yet with it much really anyway for it to be a question.
Click to expand...

I am not going to dispute that Jensen is very forward looking and that his ideas and approaches have paid off and will continue to pay off (over the long horizon). He sees Nvidia not just as a hardware company, but as a software developer enabling company that's focused on delivering infrastructure for hardware acceleration.

The risk is mainly short term over investment via cloud providers.
 
AVATARAT said:
Compiling shaders isn’t a big deal—we can just wait a little longer for it to finish, and ultimately, it doesn’t happen very often.
Click to expand...

Sadly, with UE5 and some other new games it becomes a daily or multiple weekly thing. Now, just check the shaders isn't too bad but some games if they patch twice in a week, that can be anywhere from 5-30 minutes per compile. Dune Awakening is crazy with checking every game start up and patches are worse. Marvel Rivals can take up to 25-30 minutes on my older quad core cpu.
 
Niner said:
I agree, but it's easier for GN to say they were blacklisted and create a whole video on it for clicks. In this day and age where it's expensive to build a PC and some people aren't interested in reviews on hardware they can't afford GN seems to be more interested in making videos like this which to me is a big turn off.

He has for the lack of a better term become more "arrogant" in his presentations.
Click to expand...
Then AMD shouldn't hold back samples for reviewers. There is a thing called the Streisand effect. To give you an idea, Linus Tech Tips video on the 9950X3DX2 got a little over 600k in views. No bad, not great. Gamers Nexus video on AMD black listing them got 445K in views. The AMD WTF video got 446K views. The 14 hour old 9950X3DX2 review from Gamers Nexus now has 210K reviews. I'm going to watch Gamers Nexus review just for the controversy. I didn't care to watch a 9950X3DX2 review because I'll never buy this CPU. Gamers Nexus already has more views around this CPU, before they even reviewed the CPU. That's good business for Gamers Nexus.
 
1.1.2.3.5... said:
It's risk.

We do not know the end state of the supply demand curves for AI. We don't for sure what profits will look like long term.

https://www.mckinsey.com/industries...-7-trillion-dollar-race-to-scale-data-centers

https://www.jll.com/en-us/insights/market-outlook/data-center-outlook

https://datacenter.uptimeinstitute.com/rs/711-RIA-145/images/2024.GlobalDataCenterSurvey.Report.pdf

Bottoms line we really do not know the real ROI due to so many factors so it is a massive risk with massive investment.

We have seen markets play out like this before.

Short term over investment and long term under investment. There's definitely going to be demand, but what's the fair market price to charge to cover the costs plus risk? You can't claim anyone know this for sure.

If there is a bubble that pops, and no one actually knows this either way, that's not going to be good for companies making these moves.

My preference for companies I invest is, is that that understand risk and have diversified supply chains and more than a couple of pillars they support.

If you're going to claim there's not substantial risk to AI over investment then your operating on narratives and not empirical evidence. I don't think there's a productive discussion under those circumstances. No offense intended.
Click to expand...

No offense taken, and none intended on my end.

I think you're misunderstanding a few things in your analysis:

1) The demand for compute does not start and stop with AI. AI is a significant source of the current demand, but it is not the only source of the current demand.

2) I don't understand how the risks of this are borne by Nvidia, and how making gaming GPUs somehow mitigates this risk. Nvidia is the picks and shovel seller of the current datacenter gold rush. Their "diversification" in the form of gaming GPUs is already built in, because gaming GPUs are being made from technology they're also using for the datacenter. The guys you should be concerned about in the current situation are guys like Oracle. Most companies are paying for this out of cash flow, Oracle is borrowing money. Most companies are not vulnerable to a downturn or a "bubble popping" because they're not using leverage for these investments (Oracle notwithstanding). That's a key difference between today and 1999 or 2007.

3) People are throwing the word "bubble" around a lot without understanding what a bubble is. Bubble is often used to describe something that has ramped up quickly in value that the labelling party didn't invest in. AI might become a bubble. The current datacenter buildout might become a bubble. Projections from the companies doing the build outs are saying they are still short of compute they require REGARDLESS OF THE STATUS OF AI.

4) A lot of the ROI is known because tech firms are using the compute, and not just for toy apps and goofy fake cat videos.

5) There has been a huge correction in the value of many companies in the AI trade. This is not bubble behaviour.

You're free to place your bets and invest accordingly.
 
HockeyJon said:
No offense taken, and none intended on my end.

I think you're misunderstanding a few things in your analysis:

1) The demand for compute does not start and stop with AI. AI is a significant source of the current demand, but it is not the only source of the current demand.

2) I don't understand how the risks of this are borne by Nvidia, and how making gaming GPUs somehow mitigates this risk. Nvidia is the picks and shovel seller of the current datacenter gold rush. Their "diversification" in the form of gaming GPUs is already built in, because gaming GPUs are being made from technology they're also using for the datacenter. The guys you should be concerned about in the current situation are guys like Oracle. Most companies are paying for this out of cash flow, Oracle is borrowing money. Most companies are not vulnerable to a downturn or a "bubble popping" because they're not using leverage for these investments (Oracle notwithstanding). That's a key difference between today and 1999 or 2007.

3) People are throwing the word "bubble" around a lot without understanding what a bubble is. Bubble is often used to describe something that has ramped up quickly in value that the labelling party didn't invest in. AI might become a bubble. The current datacenter buildout might become a bubble. Projections from the companies doing the build outs are saying they are still short of compute they require REGARDLESS OF THE STATUS OF AI.

4) A lot of the ROI is known because tech firms are using the compute, and not just for toy apps and goofy fake cat videos.

5) There has been a huge correction in the value of many companies in the AI trade. This is not bubble behaviour.

You're free to place your bets and invest accordingly.
Click to expand...

My investment strategy on Ai is relatively simple.

If the value is there to justify the costs, then the real benefits are the end users. Pepsi, UPS, etc... and with the infrastructure builders. Buying large basket of those companies ensures my upside, if it's there, without absorbing the risks.

A bubble is when asset prices far and away exceed their intrinsic value. That's a very commonly used definition and the one I am applying here.

And we do not know, according to experts that are pouring over the data, if the ROI is there or not.

I am NOT claiming it's a bubble, I am claiming we don't know if it is and that's risk.

Im not going to argue if there is always more demand for compute. That was true in 1970. But that doesn't mean the ROI is there for the level of capex. And i haven't seen any argument that actually proves there isn't a lot of unknowns and associated risks because of that.
 
DukenukemX said:
Then AMD shouldn't hold back samples for reviewers. There is a thing called the Streisand effect. To give you an idea, Linus Tech Tips video on the 9950X3DX2 got a little over 600k in views. No bad, not great. Gamers Nexus video on AMD black listing them got 445K in views. The AMD WTF video got 446K views. The 14 hour old 9950X3DX2 review from Gamers Nexus now has 210K reviews. I'm going to watch Gamers Nexus review just for the controversy. I didn't care to watch a 9950X3DX2 review because I'll never buy this CPU. Gamers Nexus already has more views around this CPU, before they even reviewed the CPU. That's good business for Gamers Nexus.
Click to expand...
The only Streisand Effect will be against Gamers Nexus. He's bitching about nothing because he's butthurt and that's going to hurt him in the long run. People may be watching his whining videos but that doesn't mean they agree or support what he's saying and it definitely doesn't mean he'll garner any long term gains from this.

AMD simply didn't send out that many samples and he didn't get one. Not everyone gets one every time. AMD doesn't have some sort of obligation to send even a single review sample out much less one to every single person who thinks they are a reviewer. There won't be any long term consequences to AMD from this. Not even any medium term consequences. Hell, I don't see any short term consequences to AMD coming from this. It's just an entitled jackass complaining he wasn't invited to the party on a release that doesn't amount to much.

Sure, I'd like to have one of these CPUs because some of what I do would benefit from it a decent bit but I'm one of the few who would get a benefit. That said, I won't be getting one. To whine about not getting a free CPU that only a tiny fraction of his viewership could make use of is pathetic. It's even worse since we know that few review places got one. It's not a wide release nor an important one and AMD decided there was no reason to flood the review market with them.
 
  • Like
Reactions: Tsumi
like this
DukenukemX said:
Then AMD shouldn't hold back samples for reviewers. There is a thing called the Streisand effect. To give you an idea, Linus Tech Tips video on the 9950X3DX2 got a little over 600k in views. No bad, not great. Gamers Nexus video on AMD black listing them got 445K in views. The AMD WTF video got 446K views. The 14 hour old 9950X3DX2 review from Gamers Nexus now has 210K reviews. I'm going to watch Gamers Nexus review just for the controversy. I didn't care to watch a 9950X3DX2 review because I'll never buy this CPU. Gamers Nexus already has more views around this CPU, before they even reviewed the CPU. That's good business for Gamers Nexus.
Click to expand...
Exactly why we probably shouldn't be going to you tube for reviews at all.
 
HockeyJon said:
3) People are throwing the word "bubble" around a lot without understanding what a bubble is. Bubble is often used to describe something that has ramped up quickly in value that the labelling party didn't invest in. AI might become a bubble. The current datacenter buildout might become a bubble. Projections from the companies doing the build outs are saying they are still short of compute they require REGARDLESS OF THE STATUS OF AI.
Click to expand...
I believe I generally agree with you. The pressure on compute is not purely "AI" and much compute that is being labeled "AI" simply isn't. I mean IBM is pushing "AI" systems to banks to help credit card companies scan and track user trends to discover fraud. Technically they call that AI. Really though is it? Kinda. Lots of such examples. The pressure on hardware to do complicated database scraping and compute is higher then ever. We have spent the last 50 years building information databases, now we are starting to have the pure compute to do more interesting things with it. For the shareholders the are calling a lot of things AI, that are just information scraping, crosslinking and compiling routines.

Anyway on the idea of the bubble. Another important factor of bubble status is actual revenue creation. So far 90% of the companies making bank on the push for more AI compute are the hardware companies. Nvidia is the biggest winner but really so far the more general chat bot type AIs have yet to make any real + revenue. The AI companies like Tesla are investing in may possibly show some return maybe. In a lot of ways its much like the original .com bubble. It isn't that anything happening at that time was a bad idea. The people involved were CORRECT however they were also very early. People were not quite ready to buy all their pet food online yet. It also evolved from a 1001 .this and .that situation into the behemoth super winners like Amazon.
IMO AI is going to go the same way. There is going to be a period of POP if you will. Its not that AI is going to go away or compute is going to end. Just like the .com pop didn't end the internet. It is simply that winners will be chosen and there is going to be no doubt at least 1 or 2 massive face plants. My assumption would be OpenAI will be a forgotten footnote in a decade. The question right now is who is going to be sitting in the OpenAI chairs holding the $ bag when they tip over. (or maybe I'm wrong and they are the Amazon of AI down the road???)
 
I cant watch GN anymore...

Too much whining and drama. I could watch one of those whining women shows like sister wives with my wife if I wanted that...

If I want data driven content watching his is like watching paint dry. HUB is soo much better presented.
 
SmokeRngs said:
The only Streisand Effect will be against Gamers Nexus. He's bitching about nothing because he's butthurt and that's going to hurt him in the long run. People may be watching his whining videos but that doesn't mean they agree or support what he's saying and it definitely doesn't mean he'll garner any long term gains from this.
Click to expand...
People are watching him and his dislikes on YouTube are nearly nothing. As of writing this, only 127 dislikes out of 21k likes on his 9950X3D2 review, and his AMD Blacklisting video has 752 dislikes out of 40k likes. Seems like a lot of people are agreeing with him. He still got a 9950X3D2 thanks to Hardware Unboxed, which made his review a day late. Only effect I see here is a positive one for Gamers Nexus. Bonus points to Hardware Unboxed for being a good sport.
AMD simply didn't send out that many samples and he didn't get one. Not everyone gets one every time. AMD doesn't have some sort of obligation to send even a single review sample out much less one to every single person who thinks they are a reviewer. There won't be any long term consequences to AMD from this. Not even any medium term consequences. Hell, I don't see any short term consequences to AMD coming from this. It's just an entitled jackass complaining he wasn't invited to the party on a release that doesn't amount to much.
Click to expand...
Maybe no long term consequences, but absolutely short term ones. If AMD was trying to get nothing but positive PR for this $900 9950X3D2, then they failed. This is one of those situations where working against reviewers has made things worse for them. Keep in mind that this is likely AMD releasing this CPU because Zen6 won't be out until 2027, or maybe later this year. Short term fuck ups maybe a problem for a short lived product.
Sure, I'd like to have one of these CPUs because some of what I do would benefit from it a decent bit but I'm one of the few who would get a benefit. That said, I won't be getting one. To whine about not getting a free CPU that only a tiny fraction of his viewership could make use of is pathetic. It's even worse since we know that few review places got one. It's not a wide release nor an important one and AMD decided there was no reason to flood the review market with them.
Click to expand...
I'm not whining about the 9950X3D2. This CPU would have went past many peoples radars because it's just another very expensive CPU that I'll just ignore. It's not even a bad CPU. It's just that AMD selectively didn't give reviewers a sample in order to make sure only positive reviews were released. If anything the complaint is at AMD's PR department as they screwed up.
ChadD said:
Exactly why we probably shouldn't be going to you tube for reviews at all.
Click to expand...
Why? You think websites have more credibility? I know a lot of bad tech review websites.
  • Toms Hardware
  • CNET
  • The Verge
  • Gizmodo
  • PC Magazine

Those are the ones I remember to avoid. YouTube is just another media where shitty and good reviews can be found. People here hating on Gamers Nexus is nutty.
 
DukenukemX said:
People are watching him and his dislikes on YouTube are nearly nothing. As of writing this, only 127 dislikes out of 21k likes on his 9950X3D2 review, and his AMD Blacklisting video has 752 dislikes out of 40k likes. Seems like a lot of people are agreeing with him. He still got a 9950X3D2 thanks to Hardware Unboxed, which made his review a day late. Only effect I see here is a positive one for Gamers Nexus. Bonus points to Hardware Unboxed for being a good sport.

Maybe no long term consequences, but absolutely short term ones. If AMD was trying to get nothing but positive PR for this $900 9950X3D2, then they failed. This is one of those situations where working against reviewers has made things worse for them. Keep in mind that this is likely AMD releasing this CPU because Zen6 won't be out until 2027, or maybe later this year. Short term fuck ups maybe a problem for a short lived product.

I'm not whining about the 9950X3D2. This CPU would have went past many peoples radars because it's just another very expensive CPU that I'll just ignore. It's not even a bad CPU. It's just that AMD selectively didn't give reviewers a sample in order to make sure only positive reviews were released. If anything the complaint is at AMD's PR department as they screwed up.

Why? You think websites have more credibility? I know a lot of bad tech review websites.
  • Toms Hardware
  • CNET
  • The Verge
  • Gizmodo
  • PC Magazine

Those are the ones I remember to avoid. YouTube is just another media where shitty and good reviews can be found. People here hating on Gamers Nexus is nutty.
Click to expand...
No one gives a shit about a guy whining about allegedly being blacklisted. And again, it's only alleged. Alleged by him on a release where few people got one of these to review. No one is going to care.

Do you think this is going to tank AMD sales somehow? Maybe you think it's going to lead to some sort of revolution from within AMD or a massive crowd with pitchforks and torches are going to siege AMD headquarters? It's not going to matter.

Look at the response to this "controversy" on this forum and in this thread. No one gives a shit outside of the fact he's whining. The overall reaction is negative towards him. When the most popular reactions are "meh" and "what a whiner" it's not good for him and it doesn't matter for AMD.

The real irony of the claim of Streisand Effect is that AMD didn't do anything. Literally, AMD did nothing. GN and plenty of other hardware reviewers weren't sent a CPU. AMD didn't send CPUs to everyone but him and put out a press release pointing out this fact. For this to be a Streisand Effect against AMD would require AMD to have made some sort of statement pointing out different treatment regarding him. The only one making a scene out of this is GN. That's the irony. You claim it will affect AMD when the only entity affected will be the one complaining about not getting a CPU with the worst part being that plenty of other reviewers didn't get one either so he obviously can't claim some sort of special discrimination. And yet again, being blacklisted for review samples is nothing new even if it has happened.

GN gets a small uptick in traffic in the short term but probably loses out in the long term once people notice the bitching and whining. If he had half a brain he wouldn't have said a damn word and put out a review with his borrowed CPU. Instead, he let his emotions and ragebaiting get the better of him.

If ragebait and pure sensationalism are what he wants to do now, he can go all in for I care. There are always people looking for that and he will find an audience. However, that will drive off much of his current/past audience in the process and he'll likely be much worse off in the long run.
 
SmokeRngs said:
If ragebait and pure sensationalism are what he wants to do now, he can go all in for I care. There are always people looking for that and he will find an audience. However, that will drive off much of his current/past audience in the process and he'll likely be much worse off in the long run.
Click to expand...
I mean he posted like an hour long rant video like a month ago accusing AMD of somehow being connected to Jeffery Epstein. And oh no no no they are friendly with the United Stated gov. You no the corporation incorporated in the United States, that operates in the United States, a company that has had major multi year multi billion dollar US Gov Contracts for at min 3 decades. If he was blacklisted... and ya who knows who cares. It could possibly very much be related to that video. I mean if I was the head of AMD, and saw clips of that video my reaction would be, well please stop calling and talking to that guy... lest he run some mis quoted insanity claiming we are in league with space aliens and the devil himself.

Steve has lost the plot.
Its either an act for clicks or someone might want to help the guy and get him a mental health hold. 🤣️
 
DukenukemX said:
People are watching him and his dislikes on YouTube are nearly nothing. As of writing this, only 127 dislikes out of 21k likes on his 9950X3D2 review, and his AMD Blacklisting video has 752 dislikes out of 40k likes. Seems like a lot of people are agreeing with him. He still got a 9950X3D2 thanks to Hardware Unboxed, which made his review a day late. Only effect I see here is a positive one for Gamers Nexus. Bonus points to Hardware Unboxed for being a good sport.

Maybe no long term consequences, but absolutely short term ones. If AMD was trying to get nothing but positive PR for this $900 9950X3D2, then they failed. This is one of those situations where working against reviewers has made things worse for them. Keep in mind that this is likely AMD releasing this CPU because Zen6 won't be out until 2027, or maybe later this year. Short term fuck ups maybe a problem for a short lived product.

I'm not whining about the 9950X3D2. This CPU would have went past many peoples radars because it's just another very expensive CPU that I'll just ignore. It's not even a bad CPU. It's just that AMD selectively didn't give reviewers a sample in order to make sure only positive reviews were released. If anything the complaint is at AMD's PR department as they screwed up.

Why? You think websites have more credibility? I know a lot of bad tech review websites.
  • Toms Hardware
  • CNET
  • The Verge
  • Gizmodo
  • PC Magazine

Those are the ones I remember to avoid. YouTube is just another media where shitty and good reviews can be found. People here hating on Gamers Nexus is nutty.
Click to expand...
I don't buy this even a little bit. AMD's own slides showed minimal gains 1 - 7% or something like that. Then they priced it at $900. This means it's a marketing stunt more than anything and everyone that I've seen review it has said the same thing. Furthermore, if AMD was going to be mad HUNBX has done far more hit pieces on them than anyone else. The likely situation is that they wanted minimal people to review it so they sent out minimal samples for testing. That's it. Framing it like they were out to get Gamer's Nexus just doesn't even make sense considering Steve has been putting out hit pieces on nVidia for the last 2 or 3 months.
 
Ranulfo said:
Sadly, with UE5 and some other new games it becomes a daily or multiple weekly thing. Now, just check the shaders isn't too bad but some games if they patch twice in a week, that can be anywhere from 5-30 minutes per compile. Dune Awakening is crazy with checking every game start up and patches are worse. Marvel Rivals can take up to 25-30 minutes on my older quad core cpu.
Click to expand...
Bulldozer is a really good processor; I love my old one :) Multitasking on Bulldozer is simply flawless, yes.
I hope you're using a modified BIOS that unlocks the NVMe on the PCIe slot ;)
The truth is, every entry-level processor nowadays is much faster than it, and you can build a machine even from second-hand parts for a ridiculously low price.

DukenukemX said:
Maybe no long term consequences, but absolutely short term ones. If AMD was trying to get nothing but positive PR for this $900 9950X3D2, then they failed. This is one of those situations where working against reviewers has made things worse for them.
Click to expand...
No one is working against AMD because everyone knows that their marketing always works against them :D
 
AVATARAT said:
The truth is, every entry-level processor nowadays is much faster than it, and you can build a machine even from second-hand parts for a ridiculously low price.
Click to expand...
I dunno, with every price increase on brand new parts, I find the second hand market increases in proportion. Outside the USA a second hand 11700k is worth only $50 - $100.00 less than a brand new 12700k.
 
Mazzspeed said:
I dunno, with every price increase on brand new parts, I find the second hand market increases in proportion. Outside the USA a second hand 11700k is worth only $50 - $100.00 less than a brand new 12700k.
Click to expand...
Oh, I mean budget-tier processors... The 12400/5600 or even the 12100/5500 run faster even with the cheapest motherboards and memory.
 
Meeho said:
Apparently, some do, as the level of whining about GN from some in this thread rivals Steve's whining about AMD.
Click to expand...
Great way to miss the point and the content of the whole post.
 
ChadD said:
I mean he posted like an hour long rant video like a month ago accusing AMD of somehow being connected to Jeffery Epstein. And oh no no no they are friendly with the United Stated gov. You no the corporation incorporated in the United States, that operates in the United States, a company that has had major multi year multi billion dollar US Gov Contracts for at min 3 decades. If he was blacklisted... and ya who knows who cares. It could possibly very much be related to that video. I mean if I was the head of AMD, and saw clips of that video my reaction would be, well please stop calling and talking to that guy... lest he run some mis quoted insanity claiming we are in league with space aliens and the devil himself.

Steve has lost the plot.
Its either an act for clicks or someone might want to help the guy and get him a mental health hold. 🤣️
Click to expand...
I think it's 100% for clicks and only for clicks. Well, 90% clicks and 10% butthurt for not getting one. With the trends in the PC market for the past several years long with the current shape of the PC market it's likely his viewership is declining. Probably right along with almost all the others who mostly focus on mainstream, basic stuff. There's not much going on with hardware and what little there is tends to be too expensive to dabble in for most people. That kills interest which in turn kills views. He wants the views to increase so he's trying to generate fake outrage.
 
  • Like
Reactions: ChadD
like this
ChadD said:
I mean he posted like an hour long rant video like a month ago accusing AMD of somehow being connected to Jeffery Epstein. And oh no no no they are friendly with the United Stated gov. You no the corporation incorporated in the United States, that operates in the United States, a company that has had major multi year multi billion dollar US Gov Contracts for at min 3 decades. If he was blacklisted... and ya who knows who cares. It could possibly very much be related to that video. I mean if I was the head of AMD, and saw clips of that video my reaction would be, well please stop calling and talking to that guy... lest he run some mis quoted insanity claiming we are in league with space aliens and the devil himself.

Steve has lost the plot.
Its either an act for clicks or someone might want to help the guy and get him a mental health hold. 🤣️
Click to expand...
You're every politicians wet dream. The perfect voter.
 
Meeho said:
I got the point, just found it a bad one and not worth a response.
Click to expand...
The basic point is: Long haired dude is all pissy that he didn't get a free CPU, bearing in mind that apparently long haired dude wasn't the only tech reviewer to miss out on a free CPU, so it's not like AMD targeted long haired dude specifically despite his click bait claims. Which is the crux of the matter regarding his spike in views - people want to hear his click bait rants based on forum chatter, but in doing so people loose respect for a guy that talks so fast he obviously needs to cut back on the caffeine ranting at a million miles an hour about conspiracy theories and his self inflicted butthurt as a result of his ego which appears to be the size of Jupiter.

"How dare AMD not send me a free CPU".
 
ChadD said:
Anyway on the idea of the bubble. Another important factor of bubble status is actual revenue creation. So far 90% of the companies making bank on the push for more AI compute are the hardware companies. Nvidia is the biggest winner but really so far the more general chat bot type AIs have yet to make any real + revenue. The AI companies like Tesla are investing in may possibly show some return maybe. In a lot of ways its much like the original .com bubble. It isn't that anything happening at that time was a bad idea. The people involved were CORRECT however they were also very early. People were not quite ready to buy all their pet food online yet. It also evolved from a 1001 .this and .that situation into the behemoth super winners like Amazon.
IMO AI is going to go the same way. There is going to be a period of POP if you will. Its not that AI is going to go away or compute is going to end. Just like the .com pop didn't end the internet. It is simply that winners will be chosen and there is going to be no doubt at least 1 or 2 massive face plants. My assumption would be OpenAI will be a forgotten footnote in a decade. The question right now is who is going to be sitting in the OpenAI chairs holding the $ bag when they tip over. (or maybe I'm wrong and they are the Amazon of AI down the road???)
Click to expand...

Short term over investment and long term underinvestment phenomenon has been a fairly consistent pattern for transformative technology over the ages. I think it's true here too (which is what you're saying).

On the point of a bubble or not...here's the only correct way to think about it. What's the definition of a bubble? It is when market prices exceed intrinsic value.

Homes, land, gold, ram, any trade-able commodity the buyers and sellers generally try to get the best prices for the market at the moment. The bid to ask spread is usually pretty small.

The intrinsic value of a commodity like compute at scale is relative to how its tied to profits for the user base. Mostly corporate users. Opening new technology no one actually knows what that is. Is it valuable? Of course. Long term will it create solid ROI? Of course, just like the internet did. If you held qqq before the dot com crash you had to wait about 15 years for it to recover. So it was a smart move over a long enough horizon and i am sure that's true here. And maybe there's no crash on these hyperscalers. But saying there no risk is not in the data.
 
  • Like
Reactions: ChadD
like this
ChadD said:
I believe I generally agree with you. The pressure on compute is not purely "AI" and much compute that is being labeled "AI" simply isn't. I mean IBM is pushing "AI" systems to banks to help credit card companies scan and track user trends to discover fraud. Technically they call that AI. Really though is it? Kinda. Lots of such examples. The pressure on hardware to do complicated database scraping and compute is higher then ever. We have spent the last 50 years building information databases, now we are starting to have the pure compute to do more interesting things with it. For the shareholders the are calling a lot of things AI, that are just information scraping, crosslinking and compiling routines.

Anyway on the idea of the bubble. Another important factor of bubble status is actual revenue creation. So far 90% of the companies making bank on the push for more AI compute are the hardware companies. Nvidia is the biggest winner but really so far the more general chat bot type AIs have yet to make any real + revenue. The AI companies like Tesla are investing in may possibly show some return maybe. In a lot of ways its much like the original .com bubble. It isn't that anything happening at that time was a bad idea. The people involved were CORRECT however they were also very early. People were not quite ready to buy all their pet food online yet. It also evolved from a 1001 .this and .that situation into the behemoth super winners like Amazon.
IMO AI is going to go the same way. There is going to be a period of POP if you will. Its not that AI is going to go away or compute is going to end. Just like the .com pop didn't end the internet. It is simply that winners will be chosen and there is going to be no doubt at least 1 or 2 massive face plants. My assumption would be OpenAI will be a forgotten footnote in a decade. The question right now is who is going to be sitting in the OpenAI chairs holding the $ bag when they tip over. (or maybe I'm wrong and they are the Amazon of AI down the road???)
Click to expand...

People are also forgetting how many services are going online. Everything is “smart” now. People’s fridges are now connected to “the cloud” for Lord knows what reason. This all necessitates additional compute infrastructure. I sell automation equipment. It’s all going online now. Programmers want to remotely program a PLC. That functionality wasn’t always there, but it is now, and you need compute infrastructure to support that. And oh yeah TikTok, YouTube, Reels, people flooding the cloud with pictures of their food at a restaurant they’ll never look at again but think the entire planet needs to know about. All of these things are demanding more and more compute and datacenter space. The demand is not slowing.

Regarding the current AI landscape and the dot com parallels, I don’t entirely agree. A lot of dot com companies were ahead of their time and ahead of the available tech of the day, I think we can all agree on that (pets.com is basically Chewy), but there was also significantly more leverage involved, and despite the cars, the infrastructure buildout was later taken advantage of. Fast forward to today, the companies building the infrastructure for this are largely doing it with cash flows, not leverage (Oracle and a few others notwithstanding). Investors in the AI sector are far more disciplined than the ones in dot com ever were (look at Oracle’s chart, this is not bubble behaviour). Most importantly, the companies doing it are reporting real increases in real revenue (see Anthropic’s latest for further questions).

I think a lot of tech enthusiasts want this to be a bubble because it’s making our hobby unreasonably expensive, and a lot of investors who missed the ramp want it to be a bubble so they can be like “see, called it, I’m so smart”, but wanting something and reality are not always the same thing.
 
Mazzspeed said:
The basic point is: Long haired dude is all pissy that he didn't get a free CPU, bearing in mind that apparently long haired dude wasn't the only tech reviewer to miss out on a free CPU, so it's not like AMD targeted long haired dude specifically despite his click bait claims. Which is the crux of the matter regarding his spike in views - people want to hear his click bait rants based on forum chatter, but in doing so people loose respect for a guy that talks so fast he obviously needs to cut back on the caffeine ranting at a million miles an hour about conspiracy theories and his self inflicted butthurt as a result of his ego which appears to be the size of Jupiter.

"How dare AMD not send me a free CPU".
Click to expand...
I actually watched the video and saw a recap of valid criticism after the initial report of not getting the CPU. Not getting the CPU was the smaller part of the video. But, hey, who cares, right? It's just expected corporate behavior, why should anyone talk about it. Just consume product and get excited about the next product.

And criticism here is really rich. Complaining about GN whining with whining about them whining. The only valid criticism was about them calling it blacklisting. Even that is more about semantics. It wasn't without merit, because AMD made other moves beside not giving them a sample, nor were they Pikachu face shocked, as they themselves stated this wasn't the first time this happened, and they certainly weren't surprised. Other criticism boils down to "I don't like how they present", "I don't like they talk about corporate-political aspect", "I don't like Steve's ego", "It's just the way it is, why mention it", "They should just do reviews" and similar vapid responses. Nothing about any factual errors or invalid critique in the content itself.
 
HockeyJon said:
People are also forgetting how many services are going online. Everything is “smart” now. People’s fridges are now connected to “the cloud” for Lord knows what reason. This all necessitates additional compute infrastructure. I sell automation equipment. It’s all going online now. Programmers want to remotely program a PLC. That functionality wasn’t always there, but it is now, and you need compute infrastructure to support that. And oh yeah TikTok, YouTube, Reels, people flooding the cloud with pictures of their food at a restaurant they’ll never look at again but think the entire planet needs to know about. All of these things are demanding more and more compute and datacenter space. The demand is not slowing.

Regarding the current AI landscape and the dot com parallels, I don’t entirely agree. A lot of dot com companies were ahead of their time and ahead of the available tech of the day, I think we can all agree on that (pets.com is basically Chewy), but there was also significantly more leverage involved, and despite the cars, the infrastructure buildout was later taken advantage of. Fast forward to today, the companies building the infrastructure for this are largely doing it with cash flows, not leverage (Oracle and a few others notwithstanding). Investors in the AI sector are far more disciplined than the ones in dot com ever were (look at Oracle’s chart, this is not bubble behaviour). Most importantly, the companies doing it are reporting real increases in real revenue (see Anthropic’s latest for further questions).

I think a lot of tech enthusiasts want this to be a bubble because it’s making our hobby unreasonably expensive, and a lot of investors who missed the ramp want it to be a bubble so they can be like “see, called it, I’m so smart”, but wanting something and reality are not always the same thing.
Click to expand...
https://www.foley.com/insights/publ...s-and-considerations-for-market-participants/

https://debtexplorer.whitecase.com/...ebt-markets-evolve-as-data-center-boom-builds

https://www.ernestchiang.com/en/pos...g-the-data-center-boom-while-hiding-leverage/

https://m.economictimes.com/news/in...ls-borrowing-frenzy/articleshow/125228221.cms

Not sure how you're justifying that these hyper scalars are not being finaced?

Here's a 2025 article laying out the similarities.

https://techblog.comsoc.org/2025/09...-optic-buildout-during-the-dot-com-boom-bust/

Again, I don't know the future, but i know risk when I see it.
 
Last edited:
HockeyJon said:
Regarding the current AI landscape and the dot com parallels, I don’t entirely agree. A lot of dot com companies were ahead of their time and ahead of the available tech of the day, I think we can all agree on that (pets.com is basically Chewy), but there was also significantly more leverage involved, and despite the cars, the infrastructure buildout was later taken advantage of. Fast forward to today, the companies building the infrastructure for this are largely doing it with cash flows, not leverage (Oracle and a few others notwithstanding). Investors in the AI sector are far more disciplined than the ones in dot com ever were (look at Oracle’s chart, this is not bubble behaviour). Most importantly, the companies doing it are reporting real increases in real revenue (see Anthropic’s latest for further questions).
Click to expand...
The investing is heavily circular in many cases, based on growth projections that are unsustainable and are already reported as falling behind schedules (datacenter construction delays, energy unavailability), and Anthropic is scaling back their plans because the revenue is not there to support the service.
 
Meeho said:
and Anthropic is scaling back their plans because the revenue is not there to support the service.
Click to expand...
And because it got so good and usage exploded so much that people are ready to pay more, Anthropic ARR looking like this, is not a sign of their spending being unsustainable:

es%2F3b8a6e69-4912-4f95-ab30-ac98b9a241cc_1200x828.jpg


we went from people saying MIcrosoft and other $15 billion in Anthropic was too much for that 15 billions being just half their annual revenues rate real quick.
 
LukeTbk said:
And because it got so good and usage exploded so much that people are ready to pay more, Anthropic ARR looking like this, is not a sign of their spending being unsustainable:

View attachment 798897

we went from people saying MIcrosoft and other $15 billion in Anthropic was too much for that 15 billions being just half their annual revenues rate real quick.
Click to expand...
Sure, they are drowning in cash and just decided to lower limits and remove existing features from plans just because.
 
LukeTbk said:
And because it got so good and usage exploded so much that people are ready to pay more, Anthropic ARR looking like this, is not a sign of their spending being unsustainable:

View attachment 798897

we went from people saying MIcrosoft and other $15 billion in Anthropic was too much for that 15 billions being just half their annual revenues rate real quick.
Click to expand...
Yeah, that tracks. Ive been using Anthropic at a lot as have most people I know at work. It's just faster than I am.

But my company always does a "go all in" like this to start and then the bills hit and they scale it way back. Lets see where the spending is in a year and what prices are for their services.
 
Meeho said:
Sure, they are drowning in cash and just decided to lower limits and remove existing features from plans just because.
Click to expand...
that because they stopped offering $20,000 of tokens for $200 that they are, do you doubt Anthropic ARR numbers going around or ... what do you mean exactly ?

That really underrating how much tokens an openClaw type machine running many agents 24 hours a day can consume to conclude anything special about anthropic removing that use case from a fix and low monthly payment type of deal, people can use for $200,000 a year, that a good sign that they can force people toward a clean API cost model for that use case, which show right away if there is real value and not just because it is 90% subsidized.

They are like OpenAI right now, quite restricted in the amount of compute available and they need to restrict usage, they need years of datacenter buildout just for the current demand there is outthere.
 
ChadD said:
Exactly why we probably shouldn't be going to you tube for reviews at all.
Click to expand...
Unfortunately the number of written articles on a subject where you can quickly skim through to the part that you want are decreasing at faster and faster rates, its all about view times these days not clicks
 
https://www.theverge.com/ai-artific...tion-anthropic-openai-token-economics-revenue

According to the article investment was about 6.3 trillion.

The target is ROIC of about 25%.

That means investors are looking for about 1.575 in net operating profit.

If they're running at 10% operating margins that means the revenue has to be 15.75 trillion a year.

Which means the users of that service have to see more than 15.75 trillion in returns to for it to make sense for them.

Which means there has to be that much slack in either the growth opportunity or cost savings opportunity.

For scale the US GDP for 2025 was about 30ish T, but a lot of that was specifically from that AI boom. Realistically I would claim 28 trillion sans the AI numbers.

Those are insane economy scale numbers.

Quick add here: global GDP growth is around 3% or somewhere between 3 and 4 trillion. So AI needs to add 3-4 times global GDP growth every year for 4-5 years. And those can only really come from saving or growth.
 
Last edited:
LukeTbk said:
that because they stopped offering $20,000 of tokens for $200 that they are, do you doubt Anthropic ARR numbers going around or ... what do you mean exactly ?

That really underrating how much tokens an openClaw type machine running many agents 24 hours a day can consume to conclude anything special about anthropic removing that use case from a fix and low monthly payment type of deal, people can use for $200,000 a year, that a good sign that they can force people toward a clean API cost model for that use case, which show right away if there is real value and not just because it is 90% subsidized.

They are like OpenAI right now, quite restricted in the amount of compute available and they need to restrict usage, they need years of datacenter buildout just for the current demand there is outthere.
Click to expand...
Precisely. AI usage is mostly subsidized now and is not sustainable, both financially and compute resource wise. At the same time, investment is collected on the same empty promise of both user and compute expansion. Which are based on unrealistic assumptions if you want to be generous, outright lies if you want to be truthful. The whole model as it is currently set and advertised is unsustainable. Hence, VC money pouring in, appreciation explosion, market caps balooning, GDP growth out of thin air on "trust me, bro" hopium... A baloon waiting to pop, leaving some very rich groups and a sea of dead bodies (figuratively...mostly).
 
You must log in or register to reply here.
Back
Top