![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Absolutely wrong...just because we have had fucktard political leaders doesn't make that "American". Classical liberalism (which is distinct from modern liberalism) is so fundamentally powerful that it has keep us afloat and made America the beacon it is despite everyones best efforts.
AI isn't and will never be used to for autonomy because it shouldn't be used for it. AI is only used for sensors and we should think of AI on cars ONLY as sensors. The automation is done by controls engineers, which is really at the heart of robotics. AI is merely a better sensor which helps controls.
The instruction set is X86_64 which AMD owns. Intel uses it in a cross license agreement between AMD and Intel. It is not transferable, but that could be negoitated. However it is not as easy as buying intel to get their IP for x86.
It's not like they haven't already done this for the Auto Industry, Lumber Industry, Banking Congolmerates, Boeing, Coal, ... it's a pretty long list.
We don't have to debate the facts, you can be assured that we likely agree on those. I don't agree with your conclusions. Likely i disagree with your definitions.
I am being cheeky and using the term American, as in the action taken by America, instead of the American used as a form of conceptual idealism.
Okay, well sure...you can do that
In any case, this doesn't change the fact that I can't come up with a reason why Nvidia would care to buy Intel.
Nvidia is 100% planning consumer ARM CPUs. They are fully intending to compete even more directly with Intel in the windows PC space then they already do in server. Taking 1/2 the x86 companies out out means the future of windows is Nvidia (Arm or x86) vs AMD. Or they could decide to couple "Intel" x86 CPUs with Nvidia GPU. If things like Strix point are the future, maybe Nvidia wants in on that market. Intel has no chance vs Strix... XE graphics suck. But Intel+Geforce SOC with a NV Tensor core for AI sure Nvidia could see that as good business... and some AI bubble protection. (if they believe their is a non zero chance of AI popping). I think we can all see how a Intel with a 20-30 CU Geforce GPU and a Tensor core NPU could sell a ton of laptops. Its something Nvidia could probably have wipped up within a year.
That would be crazy, but I don't think impossible
He doesn’t have to buy the company to do that. Offer them a cheque and they will come to Nvidia without the baggage.
Remember when many companies made x86 chips? Pepperidge farm remembers. If Qualcomm does buy Intel, I doubt they care to continue x86. If anything it'll be to cancel x86 to push for ARM. More specifically, their ARM design since nobody else can make a high performance ARM. I don't mean the agreement that Microsoft and Qualcomm have, but because Qualcomm bought Nuvia while ARM is trying to prevent bankruptcy, which likely means no new high performance upgrades for ARM. Considering that Snapdragon X chips barely scratched the laptop market, it would make sense for Qualcomm to buy out Intel and put a knife in x86. Not gonna happen, but I'm sure Qualcomm executives are rock hard because of this idea instead of Viagra.
Are you sure, this is really not what google say (they seem to have model not just for perception but for behavior and for simulation what others humans in the environment will do in the near future):
Not so sure if we can say that, an M4 ultra will be stronger than anything they do, same for the 128 cores Ampere or the google axion... or
It would give them everything they need to build a proper translation layer for x86 to ARM to facilitate that transition.
There is too much you have wrong for me to try to correct it all...
Ok, but am I wrong that it is at least the claim that driving decision, planning, and future decision made by other road users are coming from fundamental model not just used to find stop sign, road shape and others cars....
This is bad marketing....all of it is....there might be some predictive parts where they are doing path planning, but that isn't a hard problem to solve. AI is useful for filling in and solving information dense problems that lack complete information (like image classification). Using some kind of ML for picking an optimal route isn't really AI, and I don't know you would use AI for that. At NO point are they saying the AI is doing any of the driving...if you read carefully...this is just using buzz words. I don't feel like getting into details unless you can help me out and at least tell me you know something basic about controls (For example make a simple PID in matlab or python or at least help me understand that you understand it. Tell me you understand what a plant is. What the integration step is...) . If you don't know what controls are, then go read and or watch until you get that...then help me understand that you get that, and then we can talk about simple ML and where and why it fails, at which point I won't need to explain it because you will understand it. But I am not a creator, I am not good at condensing topics like this into short essays that are easily consumed by people that don't know that much about controls. I have avoided being an engineering manager because I don't want to make powerpoints explaining things to people....
And for what other vehicle and pedestrian will do in the near future, are you saying there is not ?
It is a very classic use of AI (and a good one because you do not need to have an exact response, good enough here work), one obvious use for those are predicting future traffic condition during the travel:
Yes in my computer science engineering program we did control grane to balance effect of winds, robots, a romba, etc... But I am supposing 20 years stuff as little to do with what is going now for something like self driving and using giant amount of data and training to influence the decision to change lane or not. I am not sure how saying driving decision could be more clear (I am assume that what we are talking about here, the important part driving decision, not the simple 4 wheel drive controls systems)
It would more than likely violate a lot of laws to do that and it would be exceptionally unethical...I work in a similar space and there is ZERO chance our lawyers would ever in a million years buy off on that. Never ever....because it is exceptionally irresponsible to do that.
There is better option to predict what future traffic would like using the current data and feeding it with the current time of the day-day of the week to a trained model, I doubt it and seem way more complicated. And google lie on their website about doing so... why ?
You are saying to me that
That the car flashing will turn left or that the pedestrian that just looked both ways and seem to walk in a direction will traverse the road is not so much a physic problem and AI is not that bad a solving physics problems already.
Yeah, I gotta check out. Good luck....
I am not sure the tone is necessary, special when AI is so much used in fluid dynamic and elsewhere:
If intel fails and there is no government intervention, it would be picked apart by vultures and its parts run by other tech giants. The collapse of a company in an industry that's still growing is an entirely different proposition from the collapse of an entire industry where all competitors are in heavy waters.
I get the 11th off. I plan to spend the day with no computer, no phone...just my dog and field
.![[Spectre]](/data/avatars/m/89/89728.jpg?1455675927){kind=avatar}
Here's the thing, but most of those chips are not something you'd use in laptops and desktops with the exception of the M4. Even the M4 has it's problems in terms of usability. Before ARM you had UltraSPARC which used SPARC V9 architecture as were many other chips that were specifically made for the server market which I would even include Intel's Itanium. None of these were going to replace the CPU's used on consumer products because they sucked at it. As for the M4, what makes you think it's stronger? Geekbench and Cinebench? This is why I'm looking forward to CP2077, because that's a real application. None of those synthetic benchmarks are serial work loads which a lot of daily computing is done on. I can almost guarantee you that M4's won't be any better at Baldur's Gate 3 than competing x86 machines. I'm sure they'll do fine in Cinebench and fantastically better in Geekbench, but nobody runs them. It's funny because John Carmack predicted this years ago.
You act like ARM is doing just fine. Everyone who makes ARM chips are fine, but ARM the company is not. Even the ones who make ARM based chips aren't going to compete against Apple and Qualcomm. As it stands right now, nobody else can compete against Qualcomm and Apple in therms of making a fast ARM chip. The chips you find in Nvidia, Ampere, and Axion are not built for serial work loads which is why none of these are desirable for anything outside of servers. If ARM isn't making the fastest ARM chips then nobody else will either. You also already have an interesting market for ARM and it's called Android. Except the Android software market kinda sucks.
Nvidia had already made a good ARM based chip with their graphics and it failed. The chips ended up in Nintendo Switches because no phone or tablet manufacturer would take those over Qualcomm. A good GPU helps but it's the software that's the problem, and it's massive. It's the reason why Apple is losing Mac sales and why Qualcomm still can't penetrate the Windows market. If you're making an ARM server then software is probably less of a problem, but the same can't be said for home and business users. There's a lot of legacy software that either won't work or works very slowly on ARM. A lot of new software is still not being made for ARM. Even with that all good and done, you still have the problem where ARM is not unified like x86 in terms of UEFI and bootloaders and etc. As much as everyone wants to hate on AMD and Intel, but their software support for their hardware on Linux is top notch. No ARM cpu even comes close to the level of commitment that AMD and Intel provide for Linux support. I'm not surprised most people overlook software as a problem because everyone imagines their use case would also apply to others. Doesn't everyone just use a web browser and Microsoft Office and Teams? The elephant in the room is gaming and even with native ports you still see Apple lag massively behind. You could attribute that to the GPU, but Apple's GPU isn't exactly slow as it does well in synthetic tests. The problem maybe that the CPU just isn't meant for gaming like work loads. You can see with AMD's Zen5 and Intel's Arrow Lake that they are more built for servers but not so much for gaming.
The chips weren’t the problem it was licensing costs for the wireless modems needed to operate in Cellphones and Tablets.
Alternatively.... I would say there is a non zero chance that after their ARM chip mostly failed. They just start making x86, slapping their name on Intel chips. Tell ARM to take their threats of canceling licenses and their royalty requests and shove em. I know AMD has some power over a sale... but a merger might give them some more wiggle room to ignore AMD. I mean ARM has accused them of being bad chip designers anyway and just buying their ARM cores, doing the same with Intel seems crazy cause its Intel. Still its just a larger version of something they already did.
I really wouldn't know. I do know from working in aviation that until relatively recently Boeing's aircraft had "hardened" 286 CPUs running their data collection systems on their 737s, and then they upgraded to i5 CPUs (and wanted to charge us a gazillion dollars for the privilege of getting our own flight telemetry data streamed instead of copied via USB stick. Fair enough, it was more data then a typical USB stick would have held back then, but still.)
As an African looking in, looking at your history the most American thing to do would be to enrich the already rich and keep the company afloat until a European company buys it, like Chrysler.
Okayyyyyyyy. Did you forget your Xanax this morning or something?