DRAM pricing is killing the hobbyist SBC market
jeffgeerling.com
jeffgeerling.com
But that’s the high end of the market, which is only a few hundred million phones a year. Apple sells two or three hundred million phones annually. The bulk of the market is mid-range and low-end. It used to be that 1.4 billion smartphones were sold a year. Now we’re at about 1.1 billion. Our projections are that we might drop to 800 million this year, and down to 500 or 600 million next year.
We look at data points out of China from some of our analysts in Asia, Singapore, Hong Kong, and Taiwan. They’ve been tracking this, and they see Xiaomi and Oppo cutting low-end and mid-range smartphone volumes by half.
Yes, it’s only a $150 BOM increase on a $1,000 iPhone where Apple has some larger margin. But for smaller phones, the percentage of the BOM that goes to memory and storage is much larger. And the margins are lower, so there’s less capacity to even eat the margins. And they have also generally tended not to do long-term agreements on memory.
Why this is a big deal is that if smartphone volumes halve, that drop will happen in the low and mid-range, not the high end.
With the 4-7 year support window on Android? Maybe that's why Google is trying to kill off Graphene et al.
My Samsung Galaxy S8 died at 7 years. Some kind of thermal failure, I was able to recover my data by keeping the phone in the freezer while I copied. Known issue.
My Samsung Galaxy S21? I figure I've got another year or two in it before it, too, dies.
Having beautiful dead phones that have never had a broken screen or a hard drop is pretty depressing.
My Galaxy Tab also has dead EMMC. My HTC One M8 still works and even holds a day of charge. Too bad Android doesn't support 32bit ARM anymore.
The great filter for incompetence by the big G was real and necessary.
I'd been running CyanogenMod until they quit giving updates to the S3.
Maybe that's why the previous owner sold it.
My Galaxy Note 4 still works. Had to sideload updated web certificates.
My Galaxy S1 would still be going, but somebody got the charging port wet.
> My Samsung Galaxy S8 died at 7 years.
The bigger issue was no more OS updates since 2020, and no Play updates since 2023. The battery can be replaced but getting a fully updated OS is more involved.
The device integrity is a bigger deal, this is also a backup for some banking apps so if they don't work it kind of defeats the purpose. I removed all other apps to minimize the attacks surface.
Like everything else, phones need to be backed up.
When new cars got more expensive, used cars got more expensive, too. I expect the same to happen with the phones.
But I'm glad I don't need to upgrade for the next couple of years. I honestly want to get 4-5 years out of any phone going forward. There's basically no difference between models 12 months apart.
I can see the prices going up this year. IT's already happened to the PS5, which is bascially unheard of.
It really sucks more because the reason for it--AI--is just so godawful and pointless.
> It really sucks more because the reason for it--AI--is just so godawful and pointless.
Strong disagree.
AI is the best thing I've seen in 30 years working in software and expensive RAM for 2 years is a price I think is worth it.
I think generative AI is pretty neat, but I'm not sure it's worth the RAM increases. I use Claude like everyone else does, and it's cool, but I am a little concerned at how much absolute low-effort crap is being produced with it.
It has made YouTube considerably worse; there was already a lot of low-effort shit flooding it, but now it's almost cartoonish. A lot of the videos that I'm recommended will have thousands of views, and give kind of a facsimile of a video with "effort", only for me to realize about a quarter of the way through a bunch of AI tropes in the writing and/or the visuals. It has made the already-mediocre experience of YouTube actively bad.
I am also not convinced that the prices will go down after two years. We already have big memory vendors completely leaving the consumer market, and we have these AI companies buying literal years worth of entire production lines of RAM chips.
This is something that could be solved by competitors jumping in to fill the niche, but it takes a lot of time to build new factories for this stuff, I think more than two years.
It’s amazing how quickly people forget why we only have a handful of DRAM manufacturers in the first place.
I've offered to buy her a replacement phone but at this point I think she's kind of curious as to how much life she can get out of it.
I have an iPhone 13 Pro Max; I bought it in 2023 but it was a refurb so I don't actually know how old it actually is. Regardless, it's still going strong, and I am hoping it can last through whatever RAM crunch is going on.
Said the user who didn't learn the lesson.
Apple, you do not own anything, if Apple wanna release an update next month that makes your current phone useless, there is nothing you can do to prevent it.
Apple was caught hacking battery level, hacking users GPS signal, etc.
You don't own an Apple device, Apple owns you!!!
Turns out not much more.
I don't see myself replacing mine any time soon.
Photos taken in the dark also became much better.
The default photos app can find words inside photos and translate without going on the internet. My wife’s 13 couldn’t even though we had the same OS.
I’m sure there’s more that I don’t even know about.
Anyone can do it in the cloud, but I don’t want them looking at my photos
I don't think they do that at the low-end (nor the high-end, though that doesn't matter here - higher-end manufacturers have a small margin they can eat into). People on the low-end phones want a new phone, they just cannot afford it!
There's also the issue of phones occasionally getting broken, of course.
My desktop PC is from 2008 but I'd never consider this to represent anything like common usage. In fact it's so unusual that I get to point it out in posts like this.
https://www.apple.com/iphone/compare/?modelList=iphone-17,ip...
Is the newer model better? Sure!
But it had 4k 60fps video, optical image stabilisation, a "super retina display" etc five generations ago. The specs have kept improving, but it's not a quantum leap in performance.
You buy a new phone every two years, it comes with a camera, a cpu, a gpu, a host of sensors. Same as phones did two years ago, and ten years before that.
I don’t use my current smart phone in any ways that are different to the iMate PDA2K I had twenty years so.
Because that's what happens if you use an old cheap phone in the modern day.
I even had a phone that would occasionally just crash when on a heavy website and the onscreen keyboard popped up. That was not at all infuriating!!! Especially when it would crash when I try to refine a Google search.
> You buy a new phone every two years, it comes with a camera, a cpu, a gpu, a host of sensors. Same as phones did two years ago, and ten years before that.
Some of these serve some interesting niches that might now disappear due to this DRAM supply issue, e.g. Unihertz for extra small phones or CAT for extra durable worksite phones.
It's easier to name the countries where iPhone ISN'T a luxury, as you can count them on very few hands.
This is a great thing to happen, actually. Those phones are all essentially trash that ends up in a landfill within a year or so. They should not exist at all.
This coming from an android user, that recently bought iphones for my daughters. Paying 600$ for not even top-of-the-line phones does not scream low-end to me.
https://www.wheresyoured.at/the-ai-industry-is-lying-to-you/
The prices make no sense, but that doesn't matter, they got away with it and are fighting to hold onto high prices, even as consumers balk. Their solution? Ditch poorer consumers. New cars and (branded) junk foods are luxury items now, apparently.
It looks like the current price on Amazon for the Raspberry Pi 4 4GB is ¥18,800 (~$117 at current rates), which is indeed expensive AF. Oddly, the Raspberry Pi 5 4GB is priced about the same, at ¥18,950 (~$119).
Considering inflation and the speed increases over the 4, the Raspberry Pi 5 price doesn't seem too unreasonable to me. But having the price go up well over ¥10,000 definitely takes it out of the realm of impulse buy and more into something I would only buy if I had a specific and urgent need. So I can definitely see this killing off a good chunk of the hobbyist market.
As it stands, my two older Pis are currently sitting unused in a closet, so I would definitely try to use those before buying anything new.
My big regret at the moment is not buying a 4TB M.2 SSD last year when prices were dipping down below ¥30,000. Now they have more than doubled to ¥65,000 or more. I had one in my cart, but decided not to buy it with the rationale that "well I still don't need the space right now, and the price per TB will probably come down even further by the time I do need it". That is, after all, the way that prices on computer component have worked for most of my life.
Around Christmas I tried to order one more. They wanted to above MSRP, like $500. Given the price of everything else I decided to just bite the bullet and do it.
After about a month they canceled my order. Whether that’s because they didn’t actually have one and couldn’t get one, or because they just wanted to wait for the prices to go up further I don’t know.
I went looking again two weeks ago. The exact same drive is back in stock. MSRP is now $1000. Amazon has it “on sale” for $900. Other retailers that often have slightly higher prices are asking $1250.
That’s 3-4x price increase in 2 years.
The exact same SSD is $479 on Amazon today. It's not a fancy super fast NVMe. It's a slow SATA drive. I have no idea why anyone would even consider building a PC with prices this inflated.
I did recently, specifically targeting lower capacities for the components that have been increasing (RAM and storage).
It didn’t seem like prices would be going down for a while and I didn’t have a desktop pc otherwise, so just went for it. We’ll see how it all plays out but I don’t think it was a terrible decision, as long as prices stay high for a couple years it still makes sense to just suffer through the increases
I did think for more basic cameras I'd use a lower spec Pi like 1GB, I use the full pis now for the high resolution DSI displays
Meanwhile, a refurbished corporate laptop with 16GB RAM and a 512GB SSD can be yours for $199 [1]
I'm sure there will still be people who want the Pi 5 but at these prices, I ain't one of them.
You also better hope the aliexpress dont figure out a way to get the RAM out of those things because they will start harvesting it for sure if there is money to made.
We're talking about a pi replacement. The Pi 5 is slower than a 10yo laptop. That's gives us a very vast pool of used laptops.
> You also better hope the aliexpress dont figure out a way to get the RAM
That is a real worry and I can see used machines being gutted because selling DDR3/4/5 sticks is way easier and profitable than the whole machine. Adapters for SODIMM to regular DIMM are readily available and cheap, too.
Delaying that refresh might actually end up the more expensive option.
To bypass the check during installation:
Boot the laptop from your USB.
When you see the "This PC can't run Windows 11" screen, press Shift + F10 to open a command prompt.
Type regedit and hit Enter.
Navigate to HKEY_LOCAL_MACHINE\SYSTEM\Setup.
Right-click Setup, create a new Key named LabConfig.
Inside LabConfig, create two DWORD (32-bit) values:
BypassTPMCheck = 1
BypassSecureBootCheck = 1
Close the registry and the command prompt; the installer will now let you proceed.And worse, they're shucking surplus for RAM And SSD's now. I am seeing more and more eBay auctions for surplus PC's sans SSD and RAM. So the second hand market is going to be invaded by the reseller parasites leaving us with $50 CPU-in-a-box and $500+ RAM/SSD parts
--
[1] Korg Kronos with its crazy Intel Atom based architecture notwithstanding.
I feel like for the first time in our lives we might have seen peak technology for the next few years. Everyone is going to have to make do instead of depending on ever increasing performance.
Pascal inspired syntax
Ada inspired type system
Lisp inspired templating and Macros
Compiles to C
The language has well defined syntax, strong types, and I turned up the compiler strictness to the max, treat all warnings as errors etc. After a few hours I put the agent aside, committed to git then deleted everything and hand coded some parts from scratch.
I then compared the results. Found one or two bugs in the AI code but honestly, the rest of our differences were “maters of taste” (is a helper function actually justified here or not kind of things).
Remove the "I actually only want a slideshow" instruction from your prompt :-)
Or is this just a temporary thing based on where processing is located?
Most helium from most wells is simply vented because it is expensive to separate even with its relatively high concentration, and I imagine even the best case scenario for capturing it from a fab has abysmal concentration of helium. But because most of it is vented it also means if the capital is put down to build more helium separators on gas wells it wouldn't take long to increase supply. Short term for a year or two it can be a problem, but beyond that it is simply a cost versus demand issue. There is neither a technological nor source limitation, it is a pure capital investment limitation.
> Helium is almost all captured from gas wells by cryogenically liquefying the nitrogen out of it.
It's more of a byproduct in that sense.
I remember a similar situation with neon early in the Ukraine invasion a few years ago. What I expect to happen is some other source coming online that currently doesn't try to capture it for economic reasons.
Helium recovery in scientific settings for cost saving reasons is already done, so it's not like there isn't expertise in using it.
Were you born after COVID and the 24 months of dire component shortages that followed?
If there is an escalation over Taiwan, then that will cause the loss of most of the world's high grade chip manufacturing capacity. TSMC is busy doing technology transfers into the US, but it is going to take time, those fabs won't have capacity for the whole world, and they still heavily depend on Taiwan based engineers if something goes wrong etc.
Just like with COVID you don't know how long this shortage will last.
Almost everything you own that runs on electricity has some parts from Taiwan in it. TSMC alone makes MEMS components, CMOS image sensors, NVRAM, and mixed-signal/RF/analog parts to name a few.
Also, people seem to assume that TSMC is an autonomous entity that receives sand at one loading dock and ships wafers out at another. That's not how fabs work. Their processes depend on a continuous supply of exotic materials and proprietary maintenance support from other countries, many of them US-aligned. There is no need to booby-trap any equipment at TSMC; it will grind to an unrecoverable halt soon after the first Chinese soldier fires a rifle or launches a missile.
Hopefully Xi understands that. But some say it's a personal beef/legacy thing with him, and that he doesn't even care about TSMC.
I wouldn't be surprised if there was enough damage that building a new fab from scratch is easier.
Which will put the whole world back for a decade while we rebuild the factories from scratch.
I believe helium, although important constitutes a small percent of the cost of semiconductors, so its effect on price will be less severe. It will be more noticeable in other uses of helium though - party balloons could get very expensive etc.
If there is a reasonable downside, probably no? You have a right to try to keep yourself alive, in nearly all contexts.
The problem is not that it's finite, the problem is that by the time prices rise enough to discourage people from using it frivolously, you might already be dangerously low on it.
Helium could be made with nuclear fusion, but a 1 Gigawatt nuclear fusion plant would only produce 200kg of helium per year, so it's still not a viable path to make the quantities of helium we currently use. Current usage is almost 30 Million kg per year.
You should think about this some more.
A hospital isn't going to shut down because their MRI's new helium load is getting more expensive - they'll pay a fortune for it. For a lot of other applications there are no suitable alternatives either.
The real question then becomes: what's going to happen when there's a 1000x price increase?
He had two floating balloons, one about twice as big as the other. Pointed a blowtorch at the smaller one and it (of course) popped.
"That one was filled with helium. Now, there's only one gas less dense than helium..." and right as I thought to myself "he's not gonna do what I think he's gonna do", he pointed the blowtorch at the other balloon which exploded into a much larger (and much louder) fireball.
Attention captured, for sure.
This happened for a while with CPUs in 2004 or 2005, IIRC. At the end of the Pentium 4 era clock speeds and TDPs were so high that we hit a wall. Nobody was pushing past 4 GHz even with watercooling (I tried).
Dual-core processors were neither widely available nor mainstream yet, and those that were available had much lower clock speeds. It definitely felt like we hit a lull, or a stagnation, in those years. It picked back up with a fury when Intel released the Core 2 Duo in 2006, though.
This is obviously logical. If I know how to program in Python or JS but not C and am familiar with SSH, I can do something with a SBC in a few minutes.
I get paid $200/hr. If I spent even one hour to learn what I need to deal with a microcontroller, the time cost is four times the cost of materials if I stick with what I know.
How many small projects do I need to do in my free time before it's financially smart to learn a whole new technology?
I find it remarkable that they haven't tried to make all of that easier. Any board with arduino support is easy to start using, with pared down c++, boards similar to the micro:bit support micropython and javascript as well as a few others, and a ton of modern development boards have UF2 support.
UF2 is a step change in how easy it is to flash a binary onto a microcontroller. You hold down a button before connecting it to a USB port, and then it appears as a USB drive for you to drop a file onto, once it's done "copying" the board is flashed and will run your code as soon as it resets.
If you want to gain familiarity with a board, you can drop a .uf2 file with a REPL on it and run code on the board a line at a time.
(Anyway, I still remember the thrill of writing assembly for a 68HC11 and getting a pair of hobby servos to respond.)
void setup() {
pinMode(LED_BUILTIN, OUTPUT);
}
void loop() {
digitalWrite(LED_BUILTIN, HIGH);
delay(1000);
digitalWrite(LED_BUILTIN, LOW);
delay(1000);
}(Of course, there are those to whom an Arduino is an overpriced piece of junk and they don't understand how I can't solder a three cent chip myself.)
But let's be realistic - all of these things are like my Steam library - purchases made but never used (I have a drawer full of Pis and other SBCs, and Arduino dev kits, etc. Someday I'll have time time time!).
int main() {
setup();
for(;;) loop();
}
I'd assume the average user on HN should be able to figure it out pretty easily.
It is c++ though. Just limited in similar ways to the US air force's requirements for using the language.
In fact it's a lot more straight forward to not have to deal with Network Manager config files or systemd unit files or read only rootfs headaches of Linux world.
You can usually find libraries for your language of choice to speak GPIO and expose the pin's state as a variable in your code.
These days, with ESP32, Pi Pico W etc... things have changed a lot.
But before they got popular, Why deal with MCU + wiring some weird peripheral for wifi / ethernet when you get a Pi Zero W / Clone with built in wifi for the same price?
Now what I would have loved to have done is come up with some crazy analog circuit to implement an arbitrary transfer function from potentiometer input to LED voltage, but I didn't know how to do this at the time and the dev cycle would be a lot more painful than with software.
I would not have fallen in love with microcontrollers without Raspberry Pi and PocketCHIP as stepping stones.
The messaging of "it's a tiny computer, make whatever you want with it" is so much more approachable than anything I've found on the microcontroller side. Even Arduino. I dismissed it for a long time because I misunderstood it. I thought I had to buy Arduino devices, then Arduino shields, then program them in the Arduino language using the Arduino IDE.
The Raspberry Pi 3B 1GB is $35 on CanaKit, Adafruit. The 3B+ is on PiShop for $40.
The Raspberry Pi 3A+ 512MB is $25 on CanaKit, Adafruit, PiShop, SparkFun.
The Raspberry Pi Zero 2 W is $16.35 on CanaKit, $17.25 on PiShop.
We are unavoidably headed at financial collapse, authoritarianism, and potentially the collapse of Western civilization. And ya'll are worried that you won't have 8GB of RAM in an embedded GPIO computer for your hobby? Maybe it's time to make the ultimate sacrifice: use less RAM.
The same accounts that defended and promoted LLM use just a few weeks ago are now telling RPi users to use less RAM.
I remember 15-20 years ago when hard drive prices went up through the roof because there was a flood in Thailand and it too years for prices to come down.
There is going to be supply chain issues due to the current Geopolitical situation (Helium comes out of the Gulf and that is need in chip manufacture) is also going to affect the price of components.
Eventually in a few years (as the article states) the situation will change. It just sucks at the moment.
TBH I am more worried about my ability to fill up the tank on my car as both Petrol and Diesel is unavailable locally. I can make do with whatever computer equipment I have.
inb4 AI has the same supply chain effects as a worldwide pandemic. I guess those AI doomers that talked about it being the end of the world had it right!
There is a saying that is often trotted out my economists "That the cure for high prices, is high prices".
There is a consumer market and business need for DRAM outside of AI. Someone will fulfil the need as there is a high incentive to. It just going to take a bit of time for this to happen. My equipment is going to be fine for another few years. So I am going to just hang tight and make do with what I got for now.
And those uses which fall short of the new threshold, e.g. hobbyist SBCs, slowly fall away.
Raspberry PI is the defacto standard for SBCs. Almost all the other SBCs had significant problems usually around software support and also third party support e.g. Hats, cases etc.
Oh look, there is a player coming into the market it seems:
https://economy.ac/news/2026/02/202602288291#:~:text=If%20eq...
EDIT: In fact many other chinese companies are now expanding into DRAM because of the high prices. Which confirms exactly what I said.
> chinese companies are now expanding into DRAM because of the high prices
In the US. The rest might do the other way. The US of course will try to do some arm twisting. Hopefully the world can learn to fight back.
Unless the capital cost to compete is too high and the risk of the existing manufacturers undercutting you is very real. Plus it can take 5-10 years or more to build a new fab, debug/iterate your process, then start shipping product.
Markets are prone to natural distortions. This is one form of that. It can be perfectly natural for all potential competitors to choose not to compete no matter how much demand exists.
Frankly I'd expect nationalization of some of the DRAM makers before we see the rise of useful competitors. The more likely scenario is government pressure, up to and including arresting executives, to rattle the cages of the existing players who are way better placed to expand production quickly for relatively low capex. Not that I think any action is likely in the short term. My guess is the existing players are betting on an AI bubble pop so they don't see the use in really expanding capacity only to be left with idle fabs later. None of us really knows.
The price for a couple of 32GB sticks is now over $1200 after being stable at about $200 for several years until last September. That's not a blip; that's 6-fold hike and there is no sign it is slowing down any time soon.
> That's not a blip; that's 6-fold hike and there is no sign it is slowing down any time soon.
How does that invalidate anything I said? As states in the article this will change, it will take years but it isn't forever.
I find it hard to believe that people here cannot make do with whatever hardware they already have.
I also don't believe those small SBCs would have survived long term anyway. Most people just use a Raspberry PI. It is either a MiniPC or a Raspberry PI.
Discord groups that had real-time line counts and pictures of the line at most best buys across the country (US).
The only way I got one was overpaying and a lottery system that bundled it with other hardware because they knew everyone would still buy it. It was impossible to buy online normally as you needed some kind of automated way to buy it before stock zeroed the minute it was posted.
You could pay a scalper for a gfx card, but stores had none. Now, stores have RAM at least.
You're comparing to memory sticks that went up 6x. If you were offering anywhere near 6x MSRP and you couldn't get a video card... I don't believe you.
https://www.pcmag.com/news/scalpers-have-sold-50000-nvidia-r...
https://www.pcmag.com/news/read-it-and-weep-heres-how-bad-nv...
These show GPUs available for 1.5-2.5x price, which fits what I remember.
> I couldn't even get a Raspberry PI (any model) for about a year.
https://picockpit.com/raspberry-pi/why-are-raspberry-pi-pric...
I didn't look into Pi prices a whole lot, but this suggests they were continuously available for 2-3x price.
> If you were offering 5x MSRP and you couldn't get a video card... I don't believe you.
My 1080Ti had died. I had to use a 8800GTS from the late 2000s for about a year. As that was the only GPU I had. I have no iGPU on my CPU.
There was at one time, no stock available. Not on Amazon, Not on Overclockers, Not on Scan. They had some weird lotto system taking place on most sites.
Scalpers claimed to have cards. But I wouldn't risk sending a lot of money to some random seller on ebay.
> Unless this article is massively misleading, sure it was out of stock at 1x price but it wasn't out of stock at 2-3x price.
Again I am in the UK. You could not buy any PI other than 1GB model and maybe the zero. Both of which were useless to me.
Edit since you added: Scalpers claimed to have cards. But I wouldn't risk sending a lot of money to some random seller on ebay.
Even with ebay's buyer protection?
Well not to be mean but I think "I refused to use ebay" invalidates your claim that you couldn't buy a card.
I've had problems with it before (I can't remember specifics as it was a while ago). I'd rather not going through the hassle and/or risk in the first place.
There are still plenty of scams on ebay. During this era there were people scamming. e.g the box for a GPU. Listing the entire specs and then putting right at the bottom of the listing it was only the box and not the card.
> Well not to be mean but I think "I refused to use ebay" invalidates your claim that you couldn't buy a card.
What you are doing is being hyper-pedantic. It is fucking tiresome when people do this online.
If you are going to be a smart arse, I will modify my statement to say "I could not get a card from a reputable online store as they were all out of stock and did not wish to risk buying from a less reputable one".
I would be foolish to trust some overpriced (or underpriced) listing on ebay. I've had an ebay/paypal account now for 25+ years, I've learned to never do this because I got screwed every time I did.
I'd rate my pedantry level as quite low. From my point of view this is not a nitpick.
Especially because you emphasized "at any price". It's the scalpers and the used market that were selling at any price. Sticking to reputable stores means sticking close to MSRP.
I would expect people to understand that unreasonable options should be omitted from conversation.
There was no stock at any of the online outlets that are commonly used in the UK when it came to GPUs for what seemed like a long time.
> I'd rate my pedantry level as quite low. From my point of view this is not a nitpick.
"I have investigated myself and found that I did nothing wrong".
Ebay is not all scalpers either. You could have gotten another 1080Ti from a legitimate previous owner.
Paying a scalper on ebay isn't. Which is what I said. Misstating what I said is disingenuous.
> You could have gotten another 1080Ti from a legitimate previous owner.
They were being scalped as well. Also people were holding onto their 10 series cards because the other cards were too expensive. So I would have had to buy an older card (which I had already had one fail) at an inflated price.
I could have bought a GT 710 or a GT1030, but that wouldn't have been any better than my 8800GTS really.
I could have flown to Taiwan and bought a card. I could have stolen one. I am sure you will invent another fantasy scenario where I could have gotten a graphics card that I didn't think about at the time.
The fact is that I could not buy a new card from an online retailer in the UK as they were out of stock. Even when they did come into stock there was a lotto system. So you couldn't really buy one then. That is a fact.
That's not pedantry. There's a huge difference between "they were unavailable and I couldn't get one at any price" and "I could have bought one from a scalper but I didn't trust them". Even if it's reasonable not to trust them (it is!), the first statement is sensational, and untrue, especially considering you emphasized "at any price" in your comment upthread.
> If you are going to be a smart arse, I will modify my statement to say "I could not get a card from a reputable online store as they were all out of stock and did not wish to risk buying from a less reputable one".
That's what you should have said in the first place; that would have been honest and correct.
And please, there's no need to call the other poster names. That's uncalled-for and childish. You seem to be new here (9-day-old account), so please read the site guidelines and turn it down a notch or three.
It is for any normal person in relatively normal setting.
Only amongst technical people is this sort of discourse tolerated where someone pretends that an unreasonable option (the scalper in this case as you admitted yourself) should be included in a statement when it is perfectly obvious it should not be included because it is not in any way reasonable.
I could have flown to the US and bought a card or China. Is that reasonable? For most people it isn't reasonable. It wasn't for me. Buying from an untrustworthy seller, is unreasonable.
> the first statement is sensational, and untrue, especially considering you emphasized "at any price" in your original comment.
They were out of stock on every reputable site. Therefore I could not buy a card at any price from them because they didn't exist.
> That's what you should have said in the first place; that would have been honest and correct.
I was honest and correct to begin with. The poster was using prices and availability in the US and not the UK.
> And please, there's no need to call the other poster names.
I never called them names. I expressed my annoyance at their behaviour.
A normal person understands scalping and that if they want it badly enough they can go on ebay.
They're not going to say it's "unavailable at any price" when it's right there for double the price.
If you're willing to pay the scalped price, the risk of using ebay is not in fact unreasonable.
So I suggest in future you should learn that using this line of logic (where you expect me to do something unreasonable to a huge number of people) is not something that people are going to put up with. It is really annoying to have to converse in this manner and in fact I believe that often that is wholly disingenuous and I no longer wish to speak to you.
But I see things a different way. The logic I'm actually using is not pedantic.
You calling me disingenuous over this is painful to look at. Get out of your own head for a second. We're using different premises, and we're reaching different conclusions because of that. My logic is fine, and your logic is fine.
I am not categorising any situation. The vast majority of people would omit unreasonable options.
I could buy a racing bike that is £5000 new, for £200 when I live in London (back in 2000s). The bike would most likely would have been stolen. So technically I can buy a £5000 bike for £200. But most people wouldn't want to buy from a thief and consider it unethical.
People feel similarly about scalpers and other untrustworthy sellers.
> You calling me disingenuous over this is painful to look at. Get out of your own head for a second.
You started the conversation claiming I was outright lying. Then when I clarified to you what I meant you continued claiming I was lying/misstating. That is really annoying.
If you could have just said "okay that is fair, while you might have been doing X and Y, I can understand why you didn't want to do that". That would have been fine. But that didn't happen.
You like the other people are was arguing with are pretending that the options were reasonable. They weren't at the time. Many other people I know thought the same.
There was no stock for any GPU except for absolute crap on any of the retail sites in the UK. There are not many options in the UK generally. It is not like the US.
As far as I am concerned what you are engaging is effectively gas-lighting.
> The line from “I just didn’t feel like doing something once” through to “My predictions for the future about a different problem are obviously true” is clear as day. Can’t see why anyone would disagree
If you deliberately want to misunderstand what is said you could draw that conclusion. Which is blatantly what you are doing.
The only thing I claimed about the current high price DRAM situation is:
1) It is likely to get worse before it gets better (due to supply chain issues due to current wars). 2) It resolve itself over time and you should be patient and just make your existing stuff last as long as possible.
That is how any crisis often plays out and I was actually telling people in my original statement not to be all doom and gloom and just be patient. It will sort itself out. It won't be this year for sure.
>I've had problems with it before (I can't remember specifics as it was a while ago). I'd rather not going through the hassle and/or risk in the first place.
As your evidence that
> Doomers IMO are just click baiting.
Like you admitted that you _do not remember_ why it was entirely unreasonable or impossible and are arguing against people that do possess memory of it being possible and reasonable enough for them at the time. Amazing stuff.
You are misunderstanding what is being said. I suspect it is deliberate.
It is often said that "Prevention is often better than the cure". Similarly it is often better not to risk spending your money unwisely than to have to go through processes to recover your money. It matters not what the specifics of the situation was (it happened a decade or more ago)
I communicated that quite clearly. So you either didn't understand or you are deliberately misunderstanding what I said.
> Like you admitted that you _do not remember_ why it was entirely unreasonable or impossible and are arguing against people that do possess memory of it being possible and reasonable enough for them at the time. Amazing stuff.
I bet you felt really clever constructing that. However as explained the specifics weren't the point. Avoiding the process entirely for funds recovery is the point.
I said "If you were offering anywhere near 6x MSRP" I didn't believe you, and it turns out you weren't offering 6x MSRP. So I wasn't calling you a liar.
> If you could have just said "okay that is fair, while you might have been doing X and Y, I can understand why you didn't want to do that". That would have been fine. But that didn't happen.
So if I had explicitly said "I think it's fine you didn't use ebay" that would have fixed everything? Because I never argued about your personal choice, I argued about you calling ebay "unreasonable".
Well for the record, I was going to say something like that in response to "If you are going to be a smart arse, I will modify my statement to say "I could not get a card from a reputable online store as they were all out of stock and did not wish to risk buying from a less reputable one"."
But then I saw you had called me "hyper-pedantic" and I focused on rebuffing that attack instead.
Edit: And it doesn't help that you never actually did that modification, and instead keep insisting that what you originally said means the same thing.
Ebay in itself isn't unreasonable.
Ebay is unreasonable when the only sellers are untrustworthy sellers, when there was a bunch of scams at the time. Which there were.
I've clarified this many times now. I don't care what interpretation is now of what I said.
> Well for the record, I was going to say something like that in response to "If you are going to be a smart arse, I will modify my statement to say "I could not get a card from a reputable online store as they were all out of stock and did not wish to risk buying from a less reputable one"."
I don't believe you. I've had plenty of stupid conversations like this, with plenty of tech nerds. Rarely happens with non-tech people. I spend some time in non-tech hobby spaces that are technical (Classic Car / Bike repairs) and this convo style never happens.
People like yourself think you are being clever buy poking holes in everything that said. I am quite happy to be quite obnoxious in pointing this out. I am tired of it. I am this cantankerous IRL about this btw.
The fact is that you could not buy a new graphics card in the UK for some time during COVID via almost every online retailers. I had conversations with other people in the UK that wanted to buy PC hardware and they were in the same situation. The same was true for the Pi 4 at the time. Making stupid semantic arguments doesn't change that fact.
> Edit: And it doesn't help that you never actually did that modification, and instead keep insisting that what you originally said means the same thing.
For all intents and purposes it is the same thing if you aren't engaging in pedantry and semantics. I try not to engage in it anymore (unless it is tit for tat), because I understand it pisses people off. You obviously don't care.
I disagree. But clearly I'm not going to convince you (and vice versa), so let's just call it a day.
> I never called them names. I expressed my annoyance at their behaviour.
"Smart arse" is name-calling.
Why don't you step back from the keyboard for a bit and cool down. Might do you some good.
Try it in a IRL conversation and see how quickly someone gets annoyed with you. It won't be very long.
> "Smart arse" is name-calling.
I said "If you are going to be a smart arse". Which means "If you are going to engage in this behaviour then ...".
I never called anyone names.
> Why don't you step back from the keyboard for a bit and cool down. Might do you some good.
I am perfectly fine. I can be mildly annoyed by someone and still be quite rational.
Also this sort of statement is close to concern trolling.
there might be a cultural difference between the old world and new world for what "at any price" means, but id take it to mean that to be at least spending $1M for it
Ah, so you could have bought one, but you judged the available suppliers to be too risky.
Completely fair, but then it's not true that you couldn't buy one "at any price". It was just not a price+risk that you were willing to take.
Also, re: Raspberry Pis, you couldn't always get the exact RAM configuration you wanted, but they were pretty continuously available during COVID on Aliexpress. You did have to pay 3-5x normal price, but you could do it. I really needed one after one at home died, and paid the 3x markup, and it was annoying but fine. Not sure if Aliexpress is equally as available in the UK as it is here in the US, though.
You are being pedantic. I find this type of discussion very tiresome. I've explained why in other forks of this thread. Quite honestly it pisses me off.
> Also, re: Raspberry Pis, you couldn't always get the exact RAM configuration you wanted, but they were pretty continuously available during COVID on Aliexpress. You did have to pay 3-5x normal price, but you could do it. I really needed one after one at home died, and paid the 3x markup, and it was annoying but fine. Not sure if Aliexpress is equally as available in the UK as it is here in the US, though.
Not in the UK. Someone was running a site with all the places that you could buy from. I was checking most days. Stock was extremely limited other than a few models.
This was my experience, too. Pis would disappear from online retailers before you noticed the stock alert email.
I only got hold of a Pi 4 by chance when Raspberry Pi did an official pop-up store in Southampton for one day only. The queue to get in was about 45 mins long.
The closest option on the pcpartpicker chart was about $75 as a stable price. So that one's only a 4x increase.
Versus DDR5 where... it looks like a 5x increase to me? I'm seeing a jump from 200USD up to 1000USD. Edit: Oh there's an extra jump in the last month on the CAD version but not the USD version.
If you are a gamer, chances are you want one of the AMD X3D CPUs. Whilst AMD did produce 5600X3D, 5700X3D and the highly sought after 5800X3D, these are effectively unobtainable now (outside of the Used Market, which is already about 2X MSRP).
You are effectively forced into AM5 (or whatever Intel is doing) and they require DDR5. You don't have the "choice" to use DDR4 anymore in most circumstances.
If your question is more of a hypothetical (assuming we could use newer CPUs with DDR4 or even DDR3) the answer is a bit more blurred, but at least in a lot of gaming workloads, you aren't memory speed bound. There is some performance regressions, sometimes up to 15%, but a lot of this is negated with the X3D chips anyways (:
CPUs don't support previous generation, so even if you don't need the RAM, you don't have a choice (assuming you need the latest CPU).
A more apt analogy is "do you really need a 28 SEER rated air conditioning system when there's a much cheaper 21 SEER available?"
(For read only workloads (no writes or only very rare writes) any ordinary SSD would suffice; the point of resorting to Optane is its unique wearout resistance.)
I think the doomers are probably anticipating another round of that and they're probably right.
We just had a vendor uplift our quote 50% per unit for some machines because of a mix of memory + supply chain issues.
"Memory card prices have TRIPLED in the last few months: when will this madness stop?!" https://www.digitalcameraworld.com/cameras/memory-cards/memo...
After all, a truck can carry a 10kg sack of rice, or a 10kg nvidia gpu. If shipping costs for 10kg rise by $15 the sack of rice has doubled in price, but the GPU is only 0.5% more expensive.
The GPUs can though.
But still 30% of the worlds helium production is apparently shut down and ships can't get to where they need to be as efficiently as they have been so there is going to be knock on effects from this.
GPUs, ram, ssds, hdds, hell even CPUs are starting to climb in price. It's an everything shortage and it's only getting worse.
A workstation that two years ago cost $3,000 was $10,000 last month and $10,500 this month. There are parts which aren't available at any price.
Between this revelation and that post recently on HN about the scanned receipts and egg prices, I find myself wondering if we're worrying about the wrong things.
We're seeing massive inflation in computing, but because the dollar is holding its value we call it increased prices. But the buying by the big buyers is the thing driving the inflation, its mechanism is scarcity.
But it's also localized. Only we experience this as a problem because compared to the hyperscalers we're poor.
The same idea applies to the price of groceries. As the prices increase, base increase being inflation, but logistic efficiency also plays a big role.
The effect is the same. The ones with more spendable income don't experience an issue yet in the projects nobody is eating fresh veggies.
The part that scares me is the creep, as I call it. Throughout the years I've always been able to carry price shocks and such but this time I'm out of the game. No more DRAM for me.
I then wonder if one day, without losing my job, I won't be able to pay for veggies.
My hard drive tree will take years to develop before it bears fruit!
Having issues with both price and availability on NVMe, SATA flash, starting to see some CPUs, and for a personal project high density spinning rust (24TB+).
We decided to just keep our current hardware for now and extend a support contract for ~ normal price.
Try 200% (tho tbf our boss sit on that quote for like a year and a half because he thought it was too pricey. Bet he regretted it now).
And all the quotes are now only valid for a week due to insane price fluctuation.
Good thing they didn't increase it.
Sounds unlikely.
Unless RAM use would impact financial metrics, it would be hard to justify the work on memory footprint if customer doesn’t complain a lot.
or to teach that again
Most programmers are JS web devs writing client side code or server side CRUD.
I would guess < 10% of programmers writing code today get perf / valgrind out on the regular. I know I dont.
My point, which I should have been clearer with, is that we aren't at a state where you can just one shot a rewrite of a complex application into another language and expect some sort of free savings. Once we are at that state, and it's good enough to pull it off, why wouldn't the AI be able to pull it off in C as well?
let foo = [1, 2, 3];
unsafe {
*foo.get_unchecked_mut(4) = 5;
}
> impossible to express in Rust
I’m not going to argue with Rust folks who misrepresent the language.
A lot of people are very excited by the idea that now language capabilities (and almost every other technical nuance) somehow don't matter but much like gravity they will continue to assert themselves whether you believe in them or not.
So far humans have proven unable to write large apps in C without those issues, given their work is the training basis for LLMs this creates two problems, one being that they don't 'know' what a safe app looks like either and any humans reviewing the outputted code will be unable to validate that either.
Now that the work is delegated to an LLM, the test and documentation quality ultimately decides the quality of the product.
Since you as the programmer no longer have to deal with the language's annoyances directly and force the LLM to perform the drudgery for you, you can build a language that makes a trade off between drudgery and quality and receive a software quality upgrade essentially for free.
LLMs are really good at producing tokens faster than developers, so make those tokens count.
Nobody who works with LLM generated code believes that LLMs produce fault-free code.
Its languages like C that you have to watch out for, because the LLM will gladly say "this is safe!" when its not.
In like 4 hours. (and most of that was me copy pasting things around to feed it reasonable chunks of information, feature by feature)
It also wrote a real-time passive DTLS-SRTP decryptor in C in like 1 hour total based on just the DTLS-SRTP RFC and a sample code of how I write suckless things in C.
I mean people can believe whatever they want. But I believe LLMs can write a reasonably fine C.
I believe that coding LLMs are particularly nice for people who are into C and suckless.
"if"
If it could you wouldn't need to use Rust. It can't, qed.
Rust gives you a lot of pain (= useful signals), before damage occurs.
Now imagine you build a reinforcement learning harness around Rust and C. Which is better for reinforcement learning? Impossible to detect failures in the final product or loud and annoying compiler errors that force you to address them?
I would not describe median quality C code as free from these issues
1. Check that you really need a SaaS SPA to solve the communication issues between your team members.
2. HTML and css should be enough for 99% of corporate websites.
3. Resize the images on your websites, they're too big.
4. Use teams in the browser, not as stand-alone app.
The old graybeards who know how to optimize efficiency may not work for them anymore, though.
If you go look, you often discover that 90% of the requests are useless, or at least could be combined. That 60% of bandwidth is used up by 3 high res images which get displayed at 30x30 pixels. That CPU performance is dominated by some rubbish code that populates an array of a million items every call, then looks up 1 element then throws the whole thing away, only to regenerate the exact same list again a few microseconds later.
We have plenty of RAM. In absolute terms, 8gb of ram in the macbook neo is 8 billion bytes. 64 billion ones and zeros. You don't need rocket science to make a CRUD app that runs well with that much ram.
Computers don't get slower over time. If we were merely as lazy with computing resources as programmers 10 years ago, most programs would scream on modern hardware.
I am a web developer of over 20 years. I can create insanely optimised pages using nothing other than vanilla CSS and JS.
I have been paid exactly once to do this. There is a site I built in 2023 that has a JS and CSS footprint of less than 100KB after GZip (large site). We even had the Go templates compiled when the web app initialised so the server responded as fast as possible.
Guess what happened when it went live? The content team use 8mb images for everything and every single optimisation I did at CSS/JS was totally useless.
Devs don't care because the people above them don't care and therefore there is zero incentive to even bother.
I hear you, and this is a real problem. But it's kind of depressing to need incentives to care about the quality of your work.
You get beaten down eventually. Late last year. I spent like an hour going through why a PR (and this developer's work) in general wasn't acceptable to my superior. He said to me that he was perfectly fine with someone not understanding basic language features (after 6 months using the langauge). He then merged it.
It didn't work (as I had warned) and created a situation where I had to turn off tests in some projects as it totally broke them. I've spent months fixing his crap and still haven't recovered from one bad PR. Now add two other employees that are like this and my manager does nothing about it. I bought a AI package from Jetbrains and now have it do almost all the work. I normally spend some time cleaning it up. Management have made it clear to me that they don't care about quality, they won't hold anyone accountable and won't even fire people that clearly cannot program.
I am 43 years old this year. I just can't be bothered trying to be a hero anymore.
Similarly, my father who retired last week was a joiner/carpenter and would be considered a master boat builder. When my sister was little my dad made her new bed with hearts and flowers carved in the headboard.
He described how adversarial he was too his employer before he retired. He was engaging in Malicious compliance (he is a layman and didn't know it was called that) because management was making his life miserable by employing the same sort of the stand-up meeting ceremony nonsense in carpentry.
They managed to make someone with that level of skill hate their job because of process.
Some companies. A lot of companies, maybe. But far from all of them.
I've done a lot of consulting work, which means I've done short stints at a lot of different places over the years. Some were absolute stinkers - like you describe. But I've also worked with some wonderful people and on some great, high performance teams. I understand that its not so easy when you're 43 (and maybe, with kids). But you don't need to stay in a job like this. Its not worth getting ground down like this. Its bad for your health. And its horrible for your career in the long run.
Move to a smaller company. Or sniff around and find a better team within your existing org. In the words of my favorite poet: The world is made to be free in. Give up all the other worlds except the one to which you belong.
I honestly think it is most of them.
> Some were absolute stinkers - like you describe. But I've also worked with some wonderful people and on some great, high performance teams.
I've totally given up on it. People don't value your work. I did a piece for a particular company. It worked perfectly. It was thrown away after a year and half because management decided everything should be rewritten in <new framework> ignoring the fact that what I had written was well documented and worked absolutely fine.
Now I shouldn't really care right? I was paid and all. But it pissed me off. What the point in doing a good job if people just throw your work in the bin?
I am looking at what my options are going forward. I am honestly considering being a car mechanic (I fix my own vehicles) or work outside for the canal trust. Realistically I suspect I might pivot to QA or doing something security related.
> I understand that its not so easy when you're 43 (and maybe, with kids). But you don't need to stay in a job like this. Its not worth getting ground down like this. Its bad for your health. And its horrible for your career in the long run.
I've been looking for over 2 years. I want to move to be closer to my family which are 300 miles away (the other side of the UK). So remote is a must. A large number of positions are hybrid, so not an option.
Outside of that many of the position in the UK are working Defence, Intelligence or Law Enforcement. All of those I have ethical reasons why I won't work for them. Outside of that there is Gambling, Pay day loans, and spooky stuff like tracking people via facial recognition.
> In the words of my favorite poet: The world is made to be free in. Give up all the other worlds except the one to which you belong.
I find this condescending.
I would not advise. If you're feeling what you are feeling now, the entire job in QA within the software field is to be essentially "controlled opposition". Unless you bust into Food Safety, Aeronautics, manufacturing, or medical devices/boomed, you will not be anything but someone the CTO is gonna go "fuck it, override" anyway. Over 14+ years experience.
The problem is management. The only solution is to make them responsible for cleaning up the consequences of their own actions, and not allowing them to delegate the task It is the only way. We have made it far to easy for assholes not responsible for doing the work to get positions managing it. We have to fix that if we're ever going to get back on the right track.
>Outside of that many of the position in the UK are working Defence, Intelligence or Law Enforcement. All of those I have ethical reasons why I won't work for them. Outside of that there is Gambling, Pay day loans, and spooky stuff like tracking people via facial recognition.
Ah, someone else feeling the sting of actually having Ethics. I feel your pain. I truly do. Our societies are dying because we forgot capitalism wasn't it's own end, but was supposed to be a means to manage solving societal ills. Once you go finance for finance sake, you end up creating more ills than you solve just because there is profit to be made.
All I can say is keep pushing. The system is on the verge of breakdown, and if the only way to get people paying attention to it is to let it happen... Then that's just what we're going to have to do I guess.
the engineering and leadership failure was at requirements time. why on earth would somebody pay for all that optimization without knowing about whats gonna be on the page first?
?OUT OF MEMORY ERROR IN 0
READY.
>Because we all know that DRAM prices have spiked since production is going to those infernal chatbot training data centers. Same as a lot of the electricity in some parts of the world, BTW.
> We all know that DRAM prices have spiked since production is going to those infernal chatbot training data centers
I know it's very fashionable here to talk about capitalism as some hand-washes-hand big corp organized scam, but if you put that ideology aside for a moment, you contradicted yourself here, I think.
I personally don't like conspiracy-theory-thinking. If I was a DRAM manufacturer and had to choose between servicing a single customer, who orders hundreds of millions worth of my product, or service a very large number of customers who order tiny amounts of the product a piece, then of course I would focus on the large client, because they are easier to service for the expected profit margin. I wouldn't even need to think about advertisement, sales, all that jazz. Looking at it from that perspective, it seems pretty logical to me that a spike in demand from datacenter operators would rise prices dramatically. I struggle to see room for collusion / conspiracy here.
If one customer buys a majority of your product, your entire business is at their mercy. They can dictate terms, or quit buying from you which can end your business.
So even with RAM - if a company goes all in on RAM for an AI company, what happens when the AI bubble bursts, or the AI company spins up/buys their own RAM factory and quits buying? Did you make enough money to tide you over until you can regain your old customers that have gotten used to not being your customer?
See coverage here:
https://bizety.com/2025/12/28/the-dirty-dram-deal-how-openai...
https://www.tomshardware.com/pc-components/dram/openais-star...
Unfortunately I don't see this reversing until HBM demand plateaus or new fabs come online, which is 2-3 years out at minimum.
I've been having fun getting Linux 7.0 running on my Milk-V Duo S, it's still available super cheap (though tariffs make buying single quantities expensive) so i stocked up on Duo boards. I guess I'm hoping for an upside where there's more interest in cheaper overstock boards from 2022+
I use the 16GB SKU to host a bunch of containers and some light debugging tools, and the power usage that sips at idle will probably pay for the whole board over my previous home server, within about 5 years.
If stuff is written in .NET, Java or JavaScript. Hosting a non-trivial web app can use several hundred megabytes of memory.
If you run normal web applications they often take many hundreds of megabytes if they are built with some popular languages that I happened to list off the top of my head. That is a fact.
Comparing that to cut down frameworks with many limitations meant for embedded devices isn't a valid comparison.
I run (regular) .NET (8) in <50mb, Javascript in <50mb, PHP in <50mb. C, Perl, Go in <20mb.
Unless you're talking about disk space.. runtimes take space.
Couldn't have seen many then! Maybe you should look elsewhere.
> Yes if you're running a repeated full OS underneath (hello VMs) then it'll waste more.
Docker is not VMs. Other people have stated this.
> I run (regular) .NET (8) in <50mb, Javascript in <50mb, PHP in <50mb. C, Perl, Go in <20mb.
Good for you. I run web services that are heavier. The container has nothing to do with it.
And most of this 50-100 MiB will be shared if you run multiple node services on the same machine the old way. So you can run 6 node app servers this way, and they'll consume eg. 150MiB of RAM total.
With docker, it's anyone guess how much running 6 node backend apps will consume, because it depends on how many things can be shared in RAM, and usually it will be nothing.
this is naive
"just as well"? lmao sure i guess i could just manually set up the environment and have differences from what im hoping to use in productio
> 1GiB machine can run a lot of server software,
this is naive
it really depends if you're crapping out some basic web app versus doing something that's actually complicated and has a need for higher performance than synchronous web calls :)
in addition, my mq pays attention to memory pressure and tunes its flow control based on that. so i have a test harness that tests both conditions to ensure that some of my backoff logic works
> if RAM is not wasted on having duplicate OSes on one machine.
this is naive
that's not how docker works...
I don't know why people instantly forget the context of the discussion, when their favorite way of doing things gets threatened. :)
Context is hobbyists and SBC market (mostly various ARM boards). Maybe I'm weird, but I really don't care about minor differences between my arch linux workstation, and my arch linux arm SBCs, because 1) they're completely different architectures, so I can't avoid the differences anyway 2) it's a hobby, I have one instance at most of any service. 3) most hobbyist run services will not work with a shitton of data or have to handle 1000s of parallel clients
What you described is exactly the opposite of how it works. There is no reasonable scenario in how that is how it works. In fact, what you're saying is opposite of the whole point of containers versus using a VM.
> when their favorite way of doing things gets threatened
No, it's when someone (like you) thinks they have an absolute answer without knowing the context.
And by the way, in my scenario, container overhead is in the range of under a hundred MiB total . The thing I'm working on HAPPENS to require a fair amount of RAM.
But you confidently asserted that "1GiB machine can run a lot of server software". And that's true for many people (like you), but not true for a lot of other people (like me).
> most hobbyist run services will not work with a shitton of data or have to handle 1000s of parallel clients
neither of these are true for me but you need to take a step back and maybe stop making absolute statements about what people are doing or working on :)
you dont get to define "where it matters" for a hobbyist
> which is where you are installing random third-party apps/containers that you want to run on your SBC locally
this is such a consoomer take. for those of us who actually build software, we have actual valid reasons for using it during development
> they're completely different architectures, so I can't avoid the differences anyway
ironically this is a side benefit modern containers are useful
i think you have a fundamental misunderstanding of how containers work and why theyre useful for software development. based on your other posts in this thread only makes me more sure of that. im not saying containers/etc are a perfect solution or always the right solution, but your misconceptions are separate from that
I've been working with "containers" since before docker existed, and I also wrote several applications that use basic technologies so called "docker containers" are based on in Linux. You can use these technologies (various namespaces, etc.) in a way that does not waste RAM. That will not happen for common docker use, where you don't control the apps and base OS completely. You can if you try hard make it efficient, but you have to have a lot of control. The moment you start pulling random dockerfiles from random sources, you'll be wasting colossal amounts of resources compared to just installing packages on your host OS, to share maximum amount of resources.
And for all these "let's have just a big static binary and put it into a container" containers, that don't really have/or need a real full OS userspace under them, there's barely any difference deployment wise from just running them without docker. In fact docker is just a very complicated additional duplicated layer in this case for what systemd does, that most people already have on their OS. So that's another RAM waste and additional overhead from what is now reduced to a service manager in this use case scenario.
i said modern containers. and you do have a FUNDAMENTAL MISUNDERSTANDING. you are repeating falsehoods throughout this entire thread.
> That will not happen for common docker use
again you are asserting a "common" use of software, when the people youre replying to are clearly using it for development
> where you don't control the apps and base OS completely
stopping saying "you" to me. id tell you to speak for yourself but you seem incapable of doing that
> And for all these "let's have just a big static binary and put it into a container" containers, that don't really have/or need a real full OS userspace under them, there's barely any difference deployment wise from just running them without docker.
ironically enough it does have differences, glaring big differences. like ironically the deployment differences are about the only reason to use docker in this situation
another stark example of you popping off with incorrect assertions. and yes there are reasons not to do use docker for this as well but it depends on multiple factors
> In fact docker is just a very complicated additional duplicated layer in this case for what systemd does, that most people already have on their OS. So that's another RAM waste and additional overhead from what is now reduced to a service manager in this use case scenario.
there are so many misconceptions in there asserted as if theyre the entire truth. yes people can use docker containers poorly but its not everyone.
> The moment you start pulling random dockerfiles from random sources, you'll be wasting colossal amounts of resources compared to just installing packages on your host OS, to share maximum amount of resources.
its a good thing that I'm not doing that! ive already stated that im using them to build software, not just "pulling random dockerfiles from random sources"
you are digging your heels in and you are now trying to assert a set of conditions and situation in which youre correct, even though youre dead wrong for the use cases that the people youre replying to are describing
you have repeated falsehoods as fact repeatedly and seem unable to adjust to people telling you "im not doing that thing youre complaining about"
frankly, i think youre out of your depth on this subject and youre trying to do anything you can justify your original claim that 1GiB is enough, or whatever
TLDR
feel free to have the last word, im sure youll have lots of them. maybe youll get lucky and a few will be correct. im exiting this conversation
and you don't provide any other technical arguments
what remains is you strongly telling me something I already acknowledged in the previous post (that you can perhaps make efficient use of docker, but it's hard to make it not waste resources in general use case)
Which bigcorp does use cubicles?
I was going to buy a small nuc and load it up on memory but I've acquired an old Mac Mini with 16GB of ram, which will do.
Maybe the company will extend the availability of the Model 2b for a few more years and release versions with less RAM (<= 500 MB)?
I remember my company buying RAM expansion boards for our PCs back in 1989 so we could run OS/2. The 4MB boards (MB! Not GB.) cost around $2000 at the time.
Like everyone, I love getting tons of RAM or SSD storage on the cheap; but we have a ways to go before we reach the 'unaffordable' level.
1. Shifting Hobbyist Focus: Because hobbyists typically prefer parts under the $100 mark (so they don't "fret over breaking them"), the community is shifting away from modern, high-powered SBCs. Instead, people are moving toward:
2. Older SBC models (like the Pi 3 or 4 with lower RAM).
3. Microcontrollers (like the RP2040) which remain cheap. So Used hardware and "repurposing" old tech is retro trending again.
IMO, perhaps there will be push to make software/firmware more RAM efficient with AI assisted coding?
My preference is responsible AI development which prevents it from turning into an arms race but that’s clearly not on the cards, especially with current leadership.
They're already regretting spending so much now that prices have started to tick downward.
I keep telling everyone: If you don't have a pressing need to buy right now, please wait 6 months and check again.
After adding to it DRAM and SSDs, the cost of the barebone remained of only 40% of the total, so the price of the memories was 50% higher than the barebone computer.
At that time, the memories were still cheaper than today, so now the price ratio would be even worse. (The barebone NUC had an Intel Arrow Lake H CPU and it cost $500, while 32 GB DDR5 + 3 TB SSDs cost $750.)
Def don't regret doing that, though I regret not springing for the extra RAM.
Where?
I think it is interesting that, at least thus far, Apple has chosen not to raise the price of their comps despite presumably the price of RAM going up multiples.
Tipping point for me: It will be a pretty kickass media server for at least a decade.
With prices steadily going up, for me it's starting to feel more sensible to repurpose the RAM sticks I've collected from old PC builds / laptops and just throw together small amd64 boxes instead of buying more RPis.
The idea of putting sixteen gigs of RAM in a raspberry pi is nuts. The legit thing you want to use a raspberry pi (or a competitor) for as an embedded headless thing with no KB/mouse/display attached should run fine in 2GB of RAM or less, assuming an ordinary debian-based OS environment.
I would much rather have a used, ex-corporate/ex-lease, small form factor or ultra small form factor x86-64 desktop PC (Dell, HP, Lenovo, whatever) with 16GB of RAM in it and an SSD on a SATA3 or NVME interface. Whatever is the "best" SFF that you can buy via huge eBay used equipment dealers on any given month.
Despite being many years old, whatever you can buy on ebay for 200 bucks (at least before the recent RAM fiasco) with some recent-ish quad core core i5/i7 or Ryzen in it will run circles around a raspberry pi 5.
So timed that all pretty great. What worries me is my desktop is up for a full new buy somewhere around early '28. That could be a train wreck depending on how taiwan situation goes
That's a very specific date / timeline. How do you decide to do a full new buy? I ask because I own a desktop that I built 15 years ago which I was flirting with replacing completely last year, but unfortunately I didn't pull the trigger ... oops :(
My old rig is still going strong. The motherboard can only take up to 32GB DDR3 though. CPU is an Intel i7-4790k which is still very fair today if you are not running a resource hog OS (looking at you Windows). Overall it is completely serviceable for my needs. Being honest with myself the only reason I wanted to upgrade was for nerd cred but I don't game much anymore and don't do any ML tasks that require lots of local compute.
Unfortunately I do also have server gear now as well. I'm going to have to really think about what I actually need now...
It's aimed at roughly hitting Zen 6 and switching to high refresh rate 4K gaming.
Was really hoping I could hop to ddr6 and pcie 6 too, but that seems less plausible
>How do you decide to do a full new buy?
Gut feel, it's a 2019 build that had a mid life refresh so 2028 is near a decade, which for a gaming rig is good going. Gaming also means a big VRAM GPU makes sense for toying with LLMs because I get dual use out of it. Plus I don't think the 3090 much as I like it will drive 4K high refresh.
...If it wasn't for gaming I could definite keep this rig till 2030+. It's comically overspec'd for browsing and casual dev stuff. Hell it's even got an optane boot drive so that'll last till the end of time
They're alive and well in ebay land. The SSD NAS I mentioned was a x570 ebay build because they can do full ECC and ebay was full of old AM4 gaming builds aging out.
>more PCIe,
Yeah that's the sticking point, though a good x570 can do 7x nvme and 8x sata...so plenty for a NAS if you're up for colouring outside the lines a bit.
nowadays the price of these 2nd hand mini PCs has shot up, and even if you do get a chance to get one, upgrading it with more RAM is gonna be painful
I will not be buying any more SBC's at this price point. I wonder if Raspberry PI will survive.
A recent nationwide poll[1] shows AI has a poorer approval rating than ICE — ICE! — probably due to their overlords being "those" SV types. Everyday AI features are being shoved down our throats. I can't even choose to not install Gemini related apps on my Android when I select "which apps to install" when booting a new phone.
But people are a weird bunch. They largely don't buy products aligning with their values. No one is jumping up and down for Graphene phones even if they had amazing privacy first software. People buy 6mi/gal hummers and iPhones for fashion, brand, money, convenience/function. The pain threshold of all bad effects still is not high enough to quit their products in a meaningful way. Values and privacy are way down in their list. I wish people would not buy/install AI related features by big tech and be more discerning, but that is likely a pipe dream.
[1] https://pos.org/wp-content/uploads/2026/03/260072-NBC-March-...
SBC (single board computer) is a fairly widely used term.
There was only one problem. The paper jumped straight from "this paper will show how our new treatment could cures cancer forever" to "as you can see, these results clearly show that our treatment cures cancer" - with neither any actual results nor any specifics on the treatment. And I don't just mean that the paper didn't go into details; writing the paper was the full extent of their "research".
Where has quality of life dropped so much that people want to die slowly at a young age of a horrible and painful disease instead of getting old?
If everything on the board but the ram costs $30, and ram is going from $10/gb to $20/gb, then they have to change the price $50 -> $70 to break even on the 2gb board, and $190 -> $350 for the 16gb board.
In other words, the raspi is now priced like a stick of ram with a bonus computer attached because ram is massively more expensive than the rest of the computer.
In the U.S., some of it was tariffs, though.
Arduino has one of the cheapest 4GB boards now, but I wonder if it's just because they made a ton and the demand for their strange board has been low?
Yes, a $250 mini PC I bought last year is now $350.
Is this pricing bad? Yeah, compared to what it was.
Is this the end of the world? Not really, and we’ve seen price spikes for all kinds of PC components in the past. It’s rarely permanent.
You can still jump on eBay and buy all kinds of dirt cheap used pieces of hardware.
My buddy just bought a used ThinkPad T14 with 32GB of RAM and 1TB of storage for about $500. You can get by with a whole lot less.
In this context, I will also present the idea that Rasperry Pi has represented quite poor cost value for many years now.
ThinkPad T14 which generation?
Placed an order with a Polish seller on eBay, received a message that Fedex wouldn't take the package due to size, replied that they could send with any shipping company and that I'm not concerned with shipping speed, after which they cancelled the order on me.
(I'm assembling a home server/workstation on a W880 mobo for ECC support - am giving the CPU all the cooling it can reasonably take, but no GPU or non-M.2 drives, so don't need much room in the case for anything other than the mobo and CPU heatsink.)
I ended up ordering one with a window since the solid side version is effectively impossible to acquire in North America.
I already moaned about this recently, but to briefly reiterate: the only hardware that's becoming available for most people in my region are Frankenstein desktops built from heavily used 10+ year old Xeons running on suspicious motherboards made by obscure Chinese manufacturers you've never heard of. This is pushing ever more people towards smartphones and away from actual computers.
But at least we got the bullshit machine in return, that's something, I guess.
It really shocks me how bad shipping has gotten. It's nearly unaffordable to buy things on eBay from the US as a Canadian due to shipping costs, so I can only imagine just how bad it is for people from other countries.
I've heard reports that these are actually surprisingly good. I wouldn't want to use one in a production environment, but for homelab stuff they're an incredible deal.
It had caused me to look around though. I have found the Pi Zero 2W to be surprisingly capable for Pi sized jobs.
Oh yeah, the market is also getting intense scrutiny from powerful geopolitical entities that are quite explicit that they don't believe in fair play or consistent, stable rules.
Would you place that bet?
Otherwise their memory manufacturing companies would be happy to exploit this opportunity.
Actually some Chinese companies already sell cheap DDR5 memory modules, but their production capacity is severely limited by the US blockade, so the cheap memories are available in few places, mainly in Asia.
So the high memory prices are caused by USA both by the AI companies that have bought most of the existing production and by the US government, who has sabotaged the Chinese memory vendors since a couple of years ago, in order to protect the market share of Micron (the US sanctions coincided with the moment when several companies, including Apple, intended to use the cheaper Chinese memories, so preventing this to happen seems a much more likely reason for the "sanctions" than the BS excuse that consumer DDR DIMMs and SSDs are dual-use products that may benefit the military. Even if that were true, the US sanctions did not prevent at all the Chinese from producing anything that would be needed in a small quantity, like for a military application. The sanctions have prevented only the mass production of devices using state-of-the-art lithography, which would have impacted the prices in consumer markets).
how long does it take to increase manufacturing capacity? how long will the decision be postponed to increase manufacturing capacity? if AI skeptics are right and the bubble bursts, increasing capacity inordinately will prove a big mistake. if AI skeptics are wrong delaying increase of capacity indordinately will prove a big mistake.
In a sense we are forcing DRAM manufacturers to play the judge, jury and executioner:
If they don't increase capacity corresponding with AI boom, the DRAM prices may ultimately cause an AI winter.
If they do increase capacity (lowering per unit costs), the lower DRAM prices may enable AI summer to continue.
This looks like self-fulfilling prophecy scenario.
I was expecting the Milk V Titan to avoid this memory nonsense since it has two unpopulated DDR4 slots, but it has fallen off the radar like several other SBCs.
I love capitalism; its ability to allocate resources on a macroeconomic scale to pick winners and more-importantly losers doesn't have a rival system. As a younger, more naive startup employee, I'm on the record making a total fool of myself responding to our CEO talking about struggling to find PMF by saying "then maybe our company doesn't deserve to exist" (yeah...) But the "capitalists" who run the world aren't actually interested in capitalism, and thus definitionally can't be capitalists. At least once upon a time we had filthy rich titans you could look up to, like Buffet and Gates (Epstein stuff aside); but at this point most of them aren't even enviable people. Despite being richer than God, people like Huang, Musk, Ellison, and Zuckerberg feel more like vampires; they want to spend their whole lives doing the exact same thing, getting richer and richer, refusing to put a ladder down for anyone else to take a shot at improving on what they've built. I actually have a modicum of respect for Bezos and, to whatever vanishingly small degree, Trump; at least they're trying something different.
If you use a PC or mini-PC that you already have, that is much cheaper than using a Raspberry Pi or similar.
2. If you are as others said, you can get very cheap GPIO addons.
For a couple years, a Pi was a decent value as a cheaper small desktop replacement.