Hacker News

I wonder if this is actually true in the long-term though. If they were to flood the market with lots of high capacity memory, then I think our programs would start using more memory too. As a result we might end up needing more memory faster compared to if they keep demand unmet.

Just consider that a chat application today takes more memory than a full 3D game with thousands of users (including chat) + the operating system used to ~20 years ago. If 128 GB of RAM were the norm then there's a chance we might expect to buy 128 GB of RAM too.

But I suppose it's really a question of how many dollars we expect to spend on memory rather than the specific amounts of memory.

I'm no expert analyst on this topic, but I'm worried this time might be different for them. China finally has the technological capability to challenge this time around and they will gobble up the unmet demand, allowing them to get more experience and more money to catch up.

> always leave demand unmet

And then allow small low cost manufacturers to get the rest of the market... like China has been doing this whole time.

Quite possible. He did say his “powerful” MacBook Pro CPU is faster than his iPhone.

I’m pretty sure even an iPhone 11 chip is more powerful than a 2019 MacBook Pro CPU in ST. An iPhone 15 is more powerful than the fastest 2019 MacBook Pro Intel CPU in MT.

I suppose he can be using a 2019 MacBook Pro or older and an iPhone 14 or older and compares only MT speeds.

> I still remember the huge DRAM shortage in late 80s forcing my startup at the time to delay launching our new product for a year.

That one was caused by manipulation by politicians, not market forces. Micron started a price war with Japanese memory manufacturers, the Japanese cut prices to compete, Micron sued them for "dumping". The saga ended with the 1986 U.S.–Japan Semiconductor Agreement, which, among other things, created production controls that limited the total dram supply. The level was set based on then current demand, and due to the rapid growth of demand at the time it almost instantly caused a massive global supply deficit.

The agreement also caused the rise of the South Korean memory industry, because the Japanese companies offloaded their now surplus equipment for cheap.

This has become like opec's best years. Only there's no cartel at least not openly.

But there's another key bottleneck. Even with all the money in the world, getting those machines that etch the RAM could be a multi year ration shop queue. And they're not making those companies every day!

To reframe your great comment:

Is it fair to conclude that DRAM is basically a commodity that can be specified well enough by a set of parameters?

If so it won’t allow you to get any competitive advantage in your products and thus wouldn’t be a business you want to be in as Apple.

The only reason I can see Apple do this is if it enables them to sell entry level devices with vastly more ram than the competition can afford. Say entry level MacBook Pro with 256GiB ram to facilitate running frontier level local models. If that is an edge they want to have.

And yet, Apple had to drop base configs from their lineup. They weren't selling $599 Minis at cost. They could take someone over and inflict damage on competition.

> And conversely, it may say something about their internal forecasts that they're not making the bet.

Idk if you can read into it that way.

All these companies have cafeterias but you don't see them investing into farmland so they can get their bananas a few cents cheaper.

But also why bother spending 20B on a fab when you can invest 20B into TSMC and let them build the fab?

> And conversely, it may say something about their internal forecasts that they're not making the bet.

It says they are no longer worried about being punished for monopolistic behavior and have bet a “ballroom donation” will exempt them from another round of punishment.

I feel like folks around here have already forgotten about the last time the memory suppliers quietly agreed to keep raising prices and stop competing with each other.

https://en.wikipedia.org/wiki/DRAM_price_fixing_scandal

Meta spent more on the Metaverse, it’s all the most certainly something they can afford to take a hit on and they won’t because memory and CPU usage is only going to go up from this point on.

Why should society care about people making profits? Society would greatly benefit from cheap abundant ram than FAANG shares going up. I'm kinda sick about only caring that some billionaire makes more money and would rather you know... actually improve conditions to better society.

This is why China is eating the West. Quite easy to start an electronics company when you have such an abundance of suppliers, compare this to America where there is maybe one or two players in the entire nation.

Quite pathetic, but we live in a pathetic world so it tracks.

China becoming self sufficient in chip fab could be the trigger point for WW3 so lets not root too hard

You needed to fill it constantly in the “old times” when memory was ubiquitous and cheap to justify the finance payments on the debt you incurred to build it.

When you’re Apple and can build it with cash, and ram is currently so insane it saves you a decade of memory price gouging in a single year, the math changes quickly.

Well, they acquired PA Semi in 2008, they used their first chip (the A4) in the iPad in 2010, and they progressively iterated using the chip in phones until they announced the switch from Intel in 2020 (concurrently, more or less, with the A14). So it took them 12 years and 10 different chips.

> I wonder how long it REALLY took them to move from intel to apple silicon, which they don't even make.

Apple has much, much more control over their silicon than they ever had with Intel.

With Intel, they were stuck with whatever Intel could supply - and that, particularly regarding power management, wasn't much.

In contrast, for ARM, Apple has been one of their longest-ever partners [1], with Apple holding one of the very rare and probably very expensive "Architecture Licenses" that allow Apple to not just build CPUs out of pre-made IP blocks, but also design their own cores and, and that is the crucial thing, extend the ARM instruction set on their CPUs.

The latter is what made the ARM move feasible, there would have been no way of getting high-performance emulation of x86 without a few specific extensions.

[1] https://www.macrumors.com/2023/09/06/apple-inks-new-deal-arm...

> The amounts of money circulating whilst some of us struggle to make rent ...

It may be expressed in terms of money, but these are assets and capital. You wouldn't find a fab a very comfortable or satisfying place to live.

The reason the numbers get so big is the gargantuan accumulated savings that are prepared to fund the similarly gargantuan ventures. The large numbers are supported by maintaining and sustaining the capitalized value by profitable investment and depreciation.

When you talk about rent, while it may be expressed in the same unit (money), the economic function played is very different. Here, it is consumption rather than saving/investment. If the dollar equivalent of all those assets went towards paying rent, while it would secure housing for some time, it would eventually become completely exhausted, and so too would housing security. Large pools of assets are simply not a way to get stable housing, it would only consume the large savings society has accumulated to be able to undergo those big ventures. It would be short-sighted and harm us in the future.

The cause and solution to rental instability is to be found in housing, building, and land policy that restricts the production of housing.

No, capitalism isn't a mistake, unregulated capitalism is a mistake. The solution is simple, don't let companies merge/acquire after a certain size. Capitalism works when there is a healthy competitive market. It doesn't work when there are 1-3 big companies all fixing prices.

> , is "man, capitalism is a mistake"

Man, we need a better term for this criticism.

The US right now is more harsh to live in than the US pre-Reagan, which was more harsh than many countries with a strong welfare state, but all of them are capitalistic.

China going capitalist ( while remaining authoritarian ) has helped lift 800 million out of desperate poverty.

I hope that India too can emulate this in my lifetime. I was born in Kerala and would love to see Indians live in a country that is as wealthy per capita ( PPP adjusted ) as Singapore or failing that even as wealthy as the USA.

Capitalism has it's problems. But you struggling with rent is entirely your self-inflicted problem...

Courtesy of TFA and capitalism:

"In 1985, if you were a reasonably affluent American, the best computer that you could afford was the IBM PC AT. The PC AT would cost you about $6,000—$19,400 in 2026 dollars—and thus represented about a quarter of the median American’s annual income; and it ran on an Intel 80286 processor, capable of something like 900,000 instructions per second. Today, if you find yourself in a market stall in Nairobi or Lagos, you’ll be able to find a cheap smartphone—like the Tecno Spark Go, manufactured by China’s Transsion—for somewhere between $30 and $120. That phone will run on a processor capable of billions of calculations per second."

> Nothing fair, or just, about this world we live in

We all make different choices, and so have different consequences.

For example, I squandered the proceeds of a business deal on a new car. If I'd bought MSFT instead, I could have bought 100 new cars.

fabs probably was low margin business compared to selling Ads before current AI boom.

30 years ago that's how the industry was structured. IBM was the largest memory manufacturer. MBAs spent decades undoing the arrangement you're now suggesting.

Well, isn’t Musk doing something similar with his new solar panels fab?

Frankly it’s ridiculous how we (the West) dropped renewables like a hot potato because it became synonymous with subsidizing China’s dominance in the field.

> Layoffs are being blamed on AI, but they are really a hangover of the covid hiring boom.

The world's most convenient virus.

A few months ago I’d have been right there with you; however recently there have been a number of high profile multi-thousand-head layoffs attributed to budget pressure of deploying AI - and so technically it’s more of an indirect rather than “a robot stole my job” replacement, but still causal. Those tokens will be “doing” labor, rather than the human.

And building gas turbines (or even designing them) are probably not jobs that pay salaries developers got accustomed to, especially in Sillicon Valley, over the last 10-20 years.

Jobs at GE in Schenectady may not be what those developers are looking for even if they were qualified.

There are a bunch of factors that are intertwining. And unfortunately, some specific demographics are mostly those being most caught in the crosshairs.

The book "1177 B.C.: The Year Civilization Collapsed" essentially argues that as being the root cause of the Bronze Age Collapse. The eastern Mediterranean and middle east were sophisticated, highly integrated cultures with a large amount of trade. There was a series of droughts, earth quakes, fires, etc. that re-occurred over a period of time, causing agricultural output to drop, leading to rebellions, leading to wars, leading the trade routes to collapse, and everything collapsing roughly at the same time.

Coffee crops have already been failing the last few years due to global warming, and now with the fertilizer crisis that is hitting major coffee producers hard this year (Kenya, Ethiopia, Indonesia, all famous for coffee and all dependent on Gulf fertilizer).

> Climate change will accelerate that

The climate is changing in a positive direction for farming. Farming is easier now than ever before. There's literally no chance of food shortages. Unless, of course, there's another attempt at building socialism.

New Jersey just declared a state of emergency after an unprecedented cold snap wiped out crops across the state, up to 90+% for some farms.

they are not failing, our Socialist President provides billions in handouts for Farmers so they gonna be just fine

If Malaysia can't obtain 30% of its fuel because of the Hormuz closure, they don't just reduce their consumption of fuel by 30%, but make up for most of the deficit by buying it for a higher price somewhere else. But the fuel they buy to make up for the deficit simply means there's less fuel for other folks buying in that market to purchase, which drives up prices. This process repeats until prices stabilize worldwide, which ultimately results in higher prices for fuel in the UK.

refunded to whom? the businesses? great! where’s my refund as a consumer who paid those higher costs (passed along to me)?

Not to mention that some of the effects that OP cited are either deflationary (like layoffs and automation), or clearly cyclic (like the memory boom bust stages, experienced just a few years ago post COVID).

Ah, the ole’ “were doing better than Germany ca 1937” argument. /s

The Taiwan war that the world is sleepwalking into is going to wipe out so much chip capacity that the current inflation will look like a blip.

> You could literally rewrite the quote to be about iron and about building railroads for trains and passengers that don't exist yet. See how silly that would be?

Couldn't possibly happen with railroads!

https://libertystreeteconomics.newyorkfed.org/2016/02/crisis...

>Except the "profit that is mathematically impossible" part. That's just made up and false. It's entirely possible that we are actually underestimating demand, and there is going to be tons of profit.

JP Morgan says $650 billion in annual revenue required to deliver mere 10% return on AI buildout is equivalent to $35 payment from every iPhone user, or $180 from every Netflix subscriber 'in perpetuity.'

Very, very, very unlikely it makes profit, which why AI keeps getting overhyped by CEOs.

Uh... The history of railroads is littered with folks losing their investments for exactly this reason.

Was your point that owning the steel companies is the path to riches even when most of the railroads fail?

Where does the money for that profit come from?

Who buys anything made by AI if he could do it by AI themselves?

Who can afford AI if their customers do with AI what they do?

Who creates the next man made code needed as training data to prevent model collapse?

> No way one can comfortable run modern software on 2015 machine

My personal laptop is a ULV 6th gen i5 with 8GB of RAM. Gets the job done, slowly but usably. That's about 9 years old. It runs Ubuntu. A dozen or so Firefox tabs, Kicad, VS Code, Slack. Computer things.

I'm of the opinion that application developers should be given middle-tier consumer systems as work machines. The whole notion of "dogfooding" is "dog-something else" if your devs experience the product using different hardware from the customer.

My initial reaction was that old compute levels are probably good enough for modern applications, but after thinking through all of the common use cases you’re probably right.

I think it’s AV that makes the difference. The modern smartphone does a ton of processing on the image sensor, and the modern laptop is expected to output multiple 4k video signals (zoom camera + desktop sharing) while accepting multiple video signals, without dropping below 120hz.

I think the reasoning is incomplete.

RAM isn't the only reason to upgrade. CPU's are more efficient, screens are significantly better and support variable-refresh-rates (and high refresh rates), battery technology has not stood still: even if they didn't degrade over 10 years.

There's plenty of reasons to upgrade to a new system even if it has the same memory capacity on paper. The memory will probably be faster and lower voltage too.

Regulation is dead under this current US administration (unless you pay a bribe). It's so weird this is where we are right now and most of the tech leadership has to play along even though it's technically illegal?

very cool word, but I think the etymology stated in Wikipedia is wrong.

the second part of the word just means purchase (in a weird ancient Greek tense). there's no relation to fish whatsoever.

(I'm Greek)

Antitrust has been dead for decades. It spans administrations and parties.

To be fair electron to rust would probably work even with LLM ram overhead

What would have happened five years ago? I suspect this VP would have gone entirely without analysis tools for these datasets or would have a department of people slowly grinding away at his long list of tickets, tickets that were poorly specified and hard to get clarity on because by the time they were worked on, he’d lost some of the context of his original request.

Now, he makes small apps that scratch his own itches, while everything is fresh in his mind and he can clarify or learn “hmm, that’s not actually what I want” and the cost is some tokens and the occasional job that runs in 3 hours instead of not existing at all.

> It took me less than 10 minutes to get that processing time down to under a minute.

You could have done it with Claude in 1 minute :) It probably never occurred to the VP to ask Claude for performance optimizations

Oil companies were profitable.

Oil always burned provided there's oxygen.

Price of oil was crystal clear how much you'll get for what you're paying.

And nothing else was affected with supply of oil.

No harm in keeping comparisons inline with the subject matter at hand.

> time spent on reducing memory footprints is seen as wasteful by the business

I think that there is a way to change that.

If an application runs significantly better on lower end hardware while delivering the same results, the customers should prefer it. It is just a matter of promoting it that way.

Also 12 processes right now on my mac for some reason.

It's haaard to do state machines.

I always leave room open for failure, and that approach has generally served me well personally and professionally. I have never been punished for having an exit strategy.

Besides, the marketplace is still in its infancy for LLMs, with a lot of unanswered questions. A lot of those questions surround the commercial viability of frontier models on bespoke hyperscaler data centers with limited usage outside of LLMs specifically should those economics be non-viable. Since that's where the memory is being tied up into, that means it's a critical question to answer in order to determine long-term investment needs into further memory fabrication.

They work well functionally, but financially anything can still happen.

> Haven't they already proven to be extremely useful?

Most certainly not. The accuracy issues mean that they can't really be used effectively for coding or search, the two things you mentioned.

> For example, Apple might make fewer iPhone 18 and let it sell out frequently. They’ll use their RAM supplies mostly for the Pro phones.

Apple has a second option that may not be open to most other vendors - as they've just demonstrated with the MacBook Neo, they could cut the RAM in half on the budget models. One good cycle of optimising the hell out of their (almost entirely native) software stack, and iOS would once again sing on a 4GB SKU.

I mean, that's the correct short-term read, but if LLMs in hyperscalers remain commercially viable to the point of tying up memory for several years, and if that necessitates an expansion of memory fabrication to satiate unmet demand, and if that demand ends up getting hoovered up by AI companies again due to their unmet or delayed demand from technological adoption, then Apple et al may not have much of a choice but to adopt such a profound strategy change.

There's a lot of 'ifs' there to be sure, but they'd be fools not to at least discuss the possibility internally and understand their options.

It would make memory-poor phones more viable. Like why can't we have a 512MB, or even 256MB RAM phone. Although I doubt that the software effort would be cheaper than just buying the extra RAM. It's definitely much more uncertain.

That's newly fabbed memory vs. an existing stock. The stock is quite massive, so optimizing existing use and enabling it to be repurposed can be meaningful.

Very interesting and thoughtful comment. I agree. When demand from hyperscalars drops rapidly after the tide goes out and everyone realizes that they can't continue building out the data centers, then fabs will be left with overcapacity that will flood the market. Wondering what this will mean for local LLMs when good RAM is available for cheap?

I'm yet to see any convincing argument that inference is subsidised in any substantial way. Training and speculative expansion are where the spend is from what I can see.

I have the same opinion, these AI companies can't even work financially with the cheap ram prices, they certainly won't with high prices.

And ram producers are betting on it, they will just milk the AI companies until they collapse.

Can a recipe be acquired and reproduced to perform correctly from the start? Say one line is performing well, can its parameters be copied over to a new one and start producing?

My head annoyingly refuses to accept the non-deterministic outcomes when applied to electronics. In other disciplines, let's say bow or musical instrument making, you need to harvest the wood, age it, dry it, pray to a deity, work with its natural im/perfections etc, but this is "just" bits.

I don’t. Also from my optimist side :-)

You are underestimating these phones. Transsion is in the top 5 smartphone shippers worldwide together with Samsung, Apple, Oppo and Xiaomi. There is nothing Frankenstein about them: https://techeconomy.ng/transsion-ranks-in-top-5-worldwide-sm...

Care to share your constructive rebuttal with us?

From an outside view OpenAI buying all that raw RAM certainly looked like they were buying it mostly to prevent anyone else from using it more than that they had an immediate need. I recall a Microsoft exec complaining that many of the GPUs they purchased were just sitting in a warehouse as they didn't have the power supply to actually do anything with them.

It will result in more people having to "rent" their computer from a corporate cloud. An extension of what already exists for poor gamers who stream their own games from a cloud renderer.

> The only thing I've really been a bit tempted by is Lidar on Apple

Maybe in that vein, one thing I wish phones would do to differentiate themselves is just add more sensors. I want my phone to be the tricorder from star trek. iPhones should have first party support for generating point clouds and measuring distances using lidar. Their microphones are probably already calibrated, why not expose that as a decibel meter. Same for light sensors. Phones used to have IR emitters, why not add those back in?

Also the iPhone still only has 240fps slow motion, I've found samsung's 960fps really useful in capturing transient phenomenon or even measuring mundane things like LED flicker.

Conversely the Pixel seems to be the only one shipping with an IR thermometer, and they'll probably remove it given most people don't seem to care. That's something I would've found useful in ad-hoc situations where I've had to make do with the back of my hand.

Air quality detection (especially pm2.5, CO2, and CO levels) would be great but I don't know if those sensors can be miniaturized enough to fit.

Improvement in what way and over what previous phone? The parent mentioned a number of metrics.

Near the end of the article it says:

> We’re already at the point where marginal buyers in the poor world are getting priced out of the smartphone market. We’re rapidly approaching the point where buyers in the rich world feel the same thing.

So it predicts that phones we buy are next.

What does it improve, in practice? For myself, I need my phone to make phone calls, take photos/videos, occasionally run apps, and to be a wifi hotspot. My iPhone 6S did all of these well enough that I only upgraded it recently because I dropped it and bent the power button. My new phone has a slightly nicer camera and better battery life, that’s about it.

About 7.6%.[1] That's Changxin Memory Technologies. That makes them #4 in DRAM, after Samsung, Hynix, and Micron. They're still only a third the size of Micron, the last US manufacturer. They have a lot of scaling up to do, but they can make current-generation DRAM.

Changxin just became very profitable, due to high DRAM prices. They're planning an IPO. And, of course, building more fabs. Yangtze Memory Technologies isn't as far along, but they're building a new fab and planning an IPO.

Both of these companies were created in 2016 as part of a national effort to become a force in the DRAM industry. It took a while to get the technology working. Now it works, and they can scale up.

[1] https://chinabizinsider.com/how-the-economics-of-the-global-...

Processors are currently becoming the next bottleneck on the server component side of things. If consumer SKUs start seeing even less volume than they are today, manufacturers will simply produce more for the datacenter.

Certain popular AMD SKUS are already 120 days lead time and growing. If a vendor will talk to you at all.

I expect motherboards to get cheap until they get to be really expensive niche products should the current situation last a few more years. At least as we know them today as PC enthusiasts and/or prosumers. I could very much see that market evaporating entirely and moving towards large integrators instead. Gamers and such were already a rather small niche to begin with.

Right now the saving grace is a server board doesn’t look a whole lot different than a consumer board. But that may change significantly faster than anyone predicts should current trends with SoCs and the like continue. You may be in a place soon where you get to choose your CPU/RAM/storage at purchase time like you do with a MacBook today.

I’m not gonna build new until I can get a GPU and RAM as well lol. That’s gonna be a few years for sure.