I wouldn't be so pessimistic, Intel and AMD aren't going to stop making CPUs, and if their integrated graphics adds AV2 it will be motivation enough for others to follow.
Ram and Ssd is just the start capitalism dictates profits will suck all production to AI. CPu have 3-5 year production timeline ie cpu prices will start going up a lot more as previous production contracts expire you might even get shortages new consumer cpus won't be affordable or spread as fast as you think.
That it's not possible for him to love her back as much as she does. "Requite" is quite an obscure word, I've only ever seen it used in the phrase "unrequited love", which means a love which isn't returned (in quite a different sense than what is used here, since I assume that the author didn't mean that he didn't love his wife, only that his love didn't measure up).
The certificate issuer doesn't have access to the underlying private keys, so while getting a fake certificate may be useful for MITM [0], undermining the certificate authorities doesn't actually allow spying on traffic that uses the genuine certs, no matter how corrupt the CA is.
There is such a thing as overestimating the power of the NSA, if the spooks actually had undermined the system to that degree they wouldn't need to lobby for all the surveillance bills that keeps popping up.
[0] And you can't get a fake certificate either without it being visible in the certificate transparency logs, or being an obvious fake since it is absent in those logs.
That Musk couldn't even get over the first stumbling block which is the statute of limitations, does not make the win any lesser. It makes it more decisive, since Musk now has to overcome that hurdle before even having a shot at the meat of the case.
I think parent probably means winning on merit or a sense of justice. As in won for a deserved reason rather than a technicality. The technicality here is exceeded the statute of limitations.
A win in any manner isn’t landing the same for observers as winning for a just reason.
There is no "sense of justice" that will sway an appeal, it's all about the law. And this jury just found that there was zero merit under law.
Appeals are for finding legal technicalities or edge cases. They do not overturn findings of fact from a jury.
That is, it used to be that way in the US, when the courts were ruled by law. In the modern US, the Supreme Court is a partisan political body, so perhaps people are confident it will get overturned because Musk is now political enough for the Supreme Court to give Musk personal favors for all his massive political contributions.
That sort of rank corruption is the only reason to be confident that Musk could ever win this silly case.
I'm clarifying and drawing the distinction related to the following because the responder wasn't responding to what was meant, but what they heard:
> Sam didn't "win" the case in the sense that most people will think of when reading this headline.
There are 2 senses of "win" here. You're talking about the "win" (A) where it is achieving victory regardless of reason.
The second type of "win" (B) that is being called out in quotes is one based either on the merits of the law (verses a technicality) or a sense of justice.
I'm not highlight the first sense of win (A) which parent of my comment and you seem to be talking about I'm pointing out grandparent is talking about (B).
Juries issue findings of fact, they don't issue rulings about the law. The issue of whether the statute of limitations applies to any given factual situation is one of law.
This doesn't compare like for like, since its comparing the total cost for the local machine with the usage cost for the cloud service, despite the cloud service also needing a local machine to be useful.
> But then Apple can negotiate on another basis and say, well, if you don’t do us a favor here and give us a better rate, then maybe we won’t work with you when all this settles down. You know things are going to settle down. These things are always cyclical. There’s never been a semiconductor boom that’s not followed by a semiconductor bust. Never. And they know it.
I have to think that the RAM suppliers wouldn't be that easy to intimidate with threats, since they know perfectly well how few alternatives Apple has. And they are also perfectly aware that Apple will play hardball with them when the market turns, regardless of whether they were nice to Apple now.
Apple bought PA Semi as the starting point to getting off of Intel. Theoretically, memory seems like something Apple could figure out how to fab. And it's not like they don't have any capital reserves.
They bought P.A. Semi, but it was for their design capability; they never had fabs anyway, and Apple still depends on TSMC and others for manufacturing chips. Apple building fabs to ensure a guaranteed supply of memory (or logic) chips would be an unprecedented level of vertical integration, even for them.
No RAM, no profits. Apple has vertically integrated in the past for less reason than this.
Moreover it's a massive economy of scale, while their consumer electronics competitors are busy fighting a losing battle against the server market for chips, Apple can undercut them, grow their market share and get even more service revenue.
RAM prices surging in the AI hype era does not mean they'll stay there for decades (see xAI already letting one data center go), and it would take a long time for Apple to become competitive.
Should they also start CPU fabs? Batteries? Lithium mines?
The risks are not symmetric. If the RAM crisis becomes the new normal it threatens Apple's business model which requires large quantities of RAM.
On the other hand, if Apple invests in RAM production and prices fall, it's not like the investment is wasted, RAM is a commodity. They lose at worst the opportunity cost of deploying the capital inefficiently, but they have so much that it hardly matters.
Apple should take this crisis as a warning that they aren't vertically integrated enough to protect their business model.
As for batteries, Apple is not even close to the largest consumer of batteries. If they were an electric car company then yes they should be making their own batteries.
That’s a decision for the new CEO (thank God he’s a technocrat), more than likely Tim Cook, and John Ternus) probably have already decided on what they’re gonna do long-term, from the outside looking in Apple has already replaced five companies? In recent times.
Memory is well within Apples design and Engineering capability. Long-term, Apple has to think about the Chinese getting a bigger part of the market in memory because they can undercut the three company cartel worldwide in time with this fake AI memory crisis.
Apple in the same timeframe also bought Intrinsity, and Anobit (a flash/SSD memory) company the Apple Silicon design group probably can do the design and engineering in house and we know they have the money the question is do they have the will their history says they do.
>It makes no sense for xAI to make their own chips.
The initial investment in chip fabs is so big it can't be justified when the established players already make enough to satisfy demand, but right now they don't so there's an opportunity.
It's still risky for sure but it makes some sense that it happens now. Hyperscalers spend 100s of billions yearly, at some point the amount given to TSMC gets larger than starting your own fab.
If success was guaranteed (it's not, as AMD and several others have learned) I think many more co's would start their own fabs in the current market.
As for why xAI, well why not - many of the others who can afford a fabbing attempt can't risk getting on TSMC's bad side even for a year or two.
The Chinese will be the one slipping in because of this opportunity. The question is is whether or not you still want to be dependent on outside memory when the Chinese takeover a larger part of the worldwide market?
Operating a FAB requires employing PhDs that are willing to work 8 hours shifts with no breaks (each removal of a bunnysuit is an expensive exercise), and there’s no reason to believe SpaceX is capable of hiring such people.
There was a point made recently by Musk that the whole clean room idea is outdated if you can just ensure the path the silicon takes from wafer to lidding is clean. Seems solvable to me, but leaves me wondering why it hasn’t been done before. I assume there is no human handling of raw/etched silicon now anyway, so why does the whole room need to be clean?
The semiconductor fab process changes dynamically to manage yield. It is not a static environment, automating with robotics is fine when things are static like a automotive assembly line, but high end semiconductor fabs are a different beast (The analogy I heard was repairing a plane while in flight).
Robots are not purely clean as well they shed contaminates as well, which must be managed too. Entropy is the reason why we still need humans in the loop.
You could probably apply that logic to any innovation in any industry no?
Reusable rockets likely got the same ridicule, as did fast satellite internet, self driving and fully electric vehicles.
I can understand that Musk does not have the most palatable personality, but floating ideas and at least attempting innovation regardless of outcome over a long time is a net positive for society and should not be discouraged.
Reusable rockets likely got the same ridicule, as did fast satellite internet, self driving and fully electric vehicles.
In those areas, Musk successfully leveraged government largesse to compete with fat, lazy incumbents who had either coasted for decades (rockets and satellite Internet) or who didn't bother to show up to the game (EVs, self-driving and otherwise.)
That does not describe the semiconductor industry.
Musk has never beaten anybody who actually put up a fight, as far as I'm aware. I guess Blue Origin technically counts, but again that's not exactly TSMC.
So what? Maybe a hand full of full bunnies per shift, and another dozen or two half-bunnies. There aren't more. This can be seen/validated by some older yt-videos, where something went wrong in the fab, for instance a https://en.wikipedia.org/wiki/FOUP ejecting a wafer in wrong ways into a machine, then being ejected by that onto the floor, and shattering. Causing all systems to stop, and all the warning lights beginning to blink in an expanding cascade. At about 4:30AM. Maybe 20 seconds later two half bunnies with face masks appear, another 10 seconds later a full bunnie. Some gesticulating ensues, full bunnie opens his suit, gets his flip phone, half bunnies downing their masks. All looking very concerned and exasperated. Having a really bad day. No more bunnies appear over several minutes. Video ends.
In the Tim Cook era when Apple needs to lock down the supply of a commodity part, they have a history of buying a dedicated manufacturing line for a manufacturing partner.
There's a bunch of chinese DRAM companies currently playing catchup to get closer to modern densities. Could Apple buy one of those? I'm guessing there would be regulatory hurdles to that on both sides of the pond.
Can a US company by a Chinese company... no. Number one, China won't let them. Number 2, China is building up these companies as a strategic reserve against the US/Korea for when they eventually go to war. So, yea, eventually the US will ban any imports of memory from those companies which would turn it into a toxic asset for Apple.
Yes, the author knows very little about the industry or how Apple operates. Fanfiction indeed.
They book manufacturing capacity often years in advance. Samsung is their majority RAM supplier and they reportedly agreed to doubling their price a few months ago.
> Yes, the author knows very little about the industry or how Apple operates.
Hardly. While it may be fan fiction, or speculation, Horace has been researching and writing about Apple's operations for decades. I tried listening to his podcast years ago and the discussion at the time of Apple's supply chain movements was extremely detailed to the point where it wasn't even listenable for me.
"Our team has over 25 years of daily research on Apple Inc"
Ask Intel, Broadcom, AMD, Nvidia, Samsung chip division and soon to be replaced Qualcomm, Apples SOC designs, probably meant memory was coming in house at some point down the road anyway. The present market conditions will probably just hasten the inevitable move.
After all, how does one miniaturize future SOC devices if you don’t bring memory in the house eventually?
> It's not possible to learn anything about other elements when performing binary search, _except_ the only thing there is to learn: if the target is before or after the recently compared element.
You have another piece of information, you don't only know if the element was before or after the compared element. You can also know the delta between what you looked at and what you're looking for. And you also have the delta from the previous item you looked at.
Assuming your key space is anything like randomly distributed.
Thinking about it--yeah, if you can anticipate anything like a random distribution it's a few extra instructions to reduce the number of values looked up. In the old days that would have been very unlikely to be a good deal, but with so many algorithms dominated by the cache (I've seen more than one case where a clearly less efficient algorithm that reduced memory reads turned out better) I suspect there's a lot of such things that don't go the way we learned them in the stone age.
Is the disconnect here that in many datasets there is some implicit distribution? For example if we are searching for english words we can assume that the number of words or sentences starting with "Q" or "Z" is very small while the ones starting with "T" are many. Or if the first three lookups in a binary search all start with "T" we are probably being asked to search just the "T" section of a dictionary.
Depending on the problem space such assumptions can prove right enough to be worth using despite sometimes being wrong. Of course if you've got the compute to throw at it (and the problem is large) take the Contact approach: why do one when you can do two in parallel for twice the price (cycles)?
You can to vary the split of the output by cracking heavier hydrocarbons into lighter. So it's not a fixed fraction, but driven by both demand and cost of processing.
And also by isomerization and alkylation. Some of the processes involved in that involve rather heroic chemistry, using things like superacids to ionize hydrocarbons.
I find all those arguments unconvincing. The right 10,000 lines of code can be worth a billion dollars. The idea that it would be somehow uneconomical for me to take the time to get it right feels like utter nonsense. I don't have to have much of an edge over an LLM to come out on top once you start to distribute the resulting product. Three months of my time costs $25,000 or so (hey, I'm in Europe, adjust as you see fit), if I can make something just a little bit better than AI Albert who can whip something together for a tenth of the price, my time will pay for itself once you have modest amounts of revenue from it.
And I'm fully convinced that what I do will not just be a little bit better than what AI Al makes. It will trounce it in all quality criteria. But of course, coincidentally with the rise of AI assistance, software quality has completely disappeared from the conversation. I wonder why.
reply