There's clearly easy/irrational money distorting the markets here. Normally this wouldn't be a problem: prices would go up, supply would eventually increase and everybody would be okay. But with AI being massively subsidized by nation-states and investors, there's no price that is too high for these supplies.
Eventually the music will stop when the easy money runs out and we'll see how much people are truly willing to pay for AI.
Regardless where demand comes from, it takes time to spin up a hard drive factory, and prices would have to rise enough that, as a producer, you would feel confident that a new hard drive factory will actually pay off. Conversely, if you feel that boom is irrational and temporary, as a producer you’d be quite wary of investing money in a new factory if there was a risk it would be producing into a glut in a few years.
If I remember during a previous GPU shortage (crypto?), Nvidia (and/or TSMC?) basically knew the music would stop and didn't want to be caught with its pants down after making the significant investments necessary to increase production
Not to mention that without enough competition, you can just raise prices, which, uh (gestures at Nvidia GPU price trends...)
You're talking about how higher prices can motivate higher supply. The parent commenter was talking about how higher prices shift the current point on the demand curve to the right. If hard drives sold for $1 billion per gigabyte, we wouldn't see even AI companies buying as many as they are, and current production would go idle. Even assuming supply is locally inelastic (as it is given no time or space to scale, or given a lack of confidence that scaling is wise), you should be able to find a price point that avoids supply shortages by manipulating demand.
The problem with this expectation of usual market behavior is that demand from AI will still be unsatisfied even after buying out the current providers' whole supply, so any new manufacturer entering the market will also prioritize high-paying AI companies above consumers.
No it’s not an easy fix. Manufacturers don’t have a good pulse on long term demand. The he capex to spin up a new manufacturing plant is significant. Especially with the recency of Covid where some folks did get caught with their pants down and over invested during the huge demand boom.
I don’t quite follow the narrative like yours about nation states and investors. There is certainly an industrial bubble going on and lots of startups getting massive amounts of capital but I here is a strong signal that a good part of this demand is here to stay.
This will be one of those scenarios where some companies will look brilliant and others foolish.
Smart manufacturers will sell 'hard drive futures'. Ie. "Give us $100/drive now for 100k drives for delivery in march 2028".
These contracts are then transferrable. The manufacturer can start work on a factory knowing they'll get paid to produce the drives.
If the AI boom comes to an end, the manufacturer is still going to get paid for their factory, and if the AI company wants to recoup costs they could try to sell those contracts back to the manufacturer for pennies on the dollar, who might then decide (if it is more profitable) to halt work on the factory - and either way they make money.
You wish. More likely all that data center capacity will be used to sell something as nefarious, like VDI for the masses. You won't need RAM, disk and GPUs when you can rent those from OpenVDI.
It's hard to increase long-run production capacity for what seems to be clearly a short-term spike in datacenter buildout. Even if AI itself is not much of a bubble, at some point spending on new AI facilities has to subside.
Seeing the first mover succeed, every Tom, Dick and Harry wants to emulate. It distorts the price because people would pay premium for everything. Then there is surplus supply and no takers. People are caught with their pants down and things go for cheap.
This repeats ad nauseum. Whether it was building ISPs during early 2000s or the abundance of streaming service where every media company wanted one. Just because the corporate overlord doesn't want to look foolish for not following a trend.
AI is going to be what fiber was to the dotcom bubble. Someone spend a lot of money on a lot of infrastructure, some of which is going to be incredibly useful, but sold for much less than it cost to build. Hardware just depreciates much much faster than fiber networks.
I'm not saying that data center buildouts can't overshoot demand but AI and compute is different than fiber buildout. The more compute you have, the smarter the AI. You can use the compute to let the AI think longer (maybe hours/days/weeks) on a solution. You can run multiple AI agents simultaneously and have them work together or check each other's work. You can train and inference better models with more compute.
So there is always use for more compute to solve problems.
Fiber installations can overshoot relatively easily. No matter how much fiber you have installed, that 4k movie isn't going to change. The 3 hours of watch time for consumers isn't going to change.
current shortages are exactly the result of fabs not wanting to commit extra capex due to overbuild risk and inference demand seems to be growing 10x yoy; you've famously got 8 year old TPUs at google at 100% load.
There's clearly easy/irrational money distorting the markets here.
No, I think it is real demand.
AI will cause shortages in everything from GPUs to CPUs, RAM, storage, networking, fiber, etc because of real demand. The physical world can't keep up with AI progress. Hence, shortages.
AI simply increases computer use by magnitudes. Now you can suddenly use Seedance 2.0 to make CGI that would have cost tens of millions 5 years ago for $5.[0] Everyone is going to need more disk space to store all those video files. Someone in their basement can make a full length movie limited only by imagination. The output quality keeps getting better quicker.
AI agents also drastically increase storage demands. Imagine financial companies using AI agents to search, scrape, organize data on stocks that they wouldn't have been able to do prior. Suddenly, disk storage and CPUs are in high demand for tasks like these.
I think the demand for computer hardware and networking gear is real and is only the beginning.
As someone who is into AI, hardware, and investing, I've been investing in physical businesses based on the above hypothesis. The only durable moats will be compute, energy, and data.
> The only durable moats will be compute, energy, and data
"Compute" is capital investment; normal and comprehensible, but on a huge scale.
"Data" is .. stolen? That feels like a problem which has been dodged but will not remain solved forever, as everyone goes shields-up against the scrapers.
"Energy" was a serious global problem before AI. All economic growth is traded off against future global temperature increases to some extent, but this is even more acute in this electricity-intensive industry. How many degrees of temperature increase is worth one .. whatever the unit of AI gain-of-function is?
The question isn’t if the demand is real or not (supplies are low, so demand must exist). The question is if the demand curve has permanently shifted, or is this a short-term issue. No one builds new capacity in response to short term changes, because you’ll have difficulty recouping the capital expense.
If AI will permanently cause an increase in hard drives over the current growth curve, then WD, et al will build new capacity, increasing supply (and reducing costs). But this really isn’t something that is known at this point.
I wonder if I'm alone in being optimistic about this. I believe that the gigantic inflow of money into hardware will lead to large increase in production capabilities, accelerated progress and perhaps even new, better architectures.
> AI will cause shortages in everything from GPUs to CPUs, RAM, storage, networking, fiber, etc because of real demand.
Real demand, sure, I agree, but maybe not retail or business demand; at the moment the "demand" is entirely VC demand.
It's a really distorted market which is to be expected in any bubble/hype phase. The current retail/business demand doesn't appear to exist at the price point these investments require - even at the low low cost of "free, gratis and for nothing", not enough consumers and businesses are signing up.
The ones really going all-in on AI are the slop-producers. I dunno if slop is enough to pay back the investment into AI - I mean, even the slop producers are going to realise that paying $200/m to produce something in 1/10th of the time is a race to the bottom because someone else on the same plan is going to do the same, but cheaper.
> The physical world can't keep up with AI progress. Hence, shortages.
I think the word "progress" is inaccurate there - the physical world supply product at the demand maintained by VC's money.
It's not "cannot keep up with progress", it's "cannot keep up with demand from VCs".
> The only durable moats will be compute, energy, and data.
That'll be a first :-) Physical commodities have never been moats on their own before.
"Punished" implies a moral valence to the whole thing which isn't there. It's not like the AI companies were aware of gamers and set out to do this. You simply got run over, like everyone else in front of the trillion dollar bulldozer.
GPUs before crypto had a lot less amount of VRAM. Crypto investment funded a lot of stupid experiments, of which some did stick to the wall. I don't think gamers had lives completely ruined by crypto in the end.
> Normally this wouldn't be a problem: prices would go up, supply would eventually increase and everybody would be okay.
This sounds like economic dogma based on pointing at some future equilibrium.
I like the saying that goes something like "life is what is happens when you are waiting for the future". In the same way, it seems to me that equilibrium is increasingly less common for many of us.
Markets are dynamic systems, and there are sub-fields of economics that recognize this. The message doesn't always get out unfortunately.
> But with AI being massively subsidized by nation-states and investors, there's no price that is too high for these supplies.
This feels like more dogma: find a convenient scape-goat: governments.
Time to wake up to what history has shown us! Markets naturally reflect boom and bust cycles, irrationality of people, and various other market failures. None of these are news to competent economists, by the way. Be careful from whence you get your economic "analysis".
Yes, this is why the prices of housing has dropped dramatically. The market stepped up and filled the demand needed and now everyone can afford a place to live
Higher price encourages more supply. Typically when you see a acute shortage, its quickly followed by a glut as supply starts coming online in an over correction.
I like to imagine the reference in the movie margin call is that of a merry go round or a game of Musical chair. Like we are all on a ride, none of us are the operator, and all we can do is guess when the music will stop (and the ride ends).
The problem with this AI stuff is we don't know how much we will be willing to pay for it, as individuals, as businesses, as nations. I guess we just don't know how far this stuff will be useful. The reasons for the high valuation is, in my guess, that there is more value here than what we have tapped so far, right?
The revenues that nVidia has reported is based on what we hope we will achieve in the future so I guess the whole thing is speculation?
Better stock up with used laptops. I'm going to buy another one this year. Those used ones usually don't last very long.
What if in the near future it is simply too expensive to own "personal" computers? What if you can no longer buy used computers from official channels but have to find local shops or sharpen up on soldering skills and find parts from dumps? The big techs will conveniently "rent out" cloud computer for us to use, in exchange of all of your data.
Bingo. A number of corporate interests don't want to let you own your personal computers for different reasons. Google/Apple wants you to get locked down devices, and cloud/AI providers want you to use their services from a weak client.
Time for folks to familiarize themselves with Linux distros designed to run on older hardware. My 2009 laptop runs great, with the exception of the browser. Oh and the fact that 32-bit software is harder and harder to find.
Yes, I'm grateful I run Linux. You can get quite a bit done with 4GB RAM and a 6th generation (or even earlier) CPU. All 64-bit. I don't think such ancient hardware will be affected by AI demand to the same degree (though I think we'll still see some prices rise if people stop buying new stuff).
The worry is that at some point the older hardware will stop working.
I think about this too. There are several headwinds. Rent-seeking and collapse of economies of scale in the consumer sector for sure, but also I feel like we've basically peaked in hardware's ability to meet routine needs.
Once the phone makers realize that they can sell phones and docking stations to businesses because 90% of knowledge work seems to happen in a web browser through one SaaS or other I think personal computers will be cooked.
I don't think phones are really any cheaper than the mini PCs that businesses can already buy. Which makes sense, because a phone has to include a battery, touch screen, and is under tighter space constraints.
I have heard that you can get used laptops. But they do not come with memory or SSD anymore... As even used components are now valuable enough to be removed and sold.
Many laptops from last few years have soldered memory. Your previous laptop's SSD can also be reused, since those don't die that quickly compared to the laptop.
In a lot of cases, owners remove the storage not because it has any value but rather they don't want to risk making a mistake letting a device go that still has data on it.
Also pulling and shredding hard drives is cheaper than paying someone to run DBAN or equivalent (which can take many hours to complete on a large drive), and there's no easy way to securely erase an SSD if it wasn't encrypted from the beginning.
Damn really? One of my go to moves when helping small political campaigns is to buy like a 2015 MBP and turn it into a locally hosted server to run all their stuff on the cheap.
I no longer feel obligated to apologize for holding on to older devices for a long time. I have several desktops and laptops that are all still usable.
It's still absolutely fascinating to me that basically the whole modern tech industry and the economic growth from it rests on the shoulders of a single company that has all of their important factories on a single island that's under constant threat of invasion. On top of that they themselves are reliant on a single company that's able to produce the machine required to print the wafers.
I don't know if TSMC has anything to do with hard drive production, but the reliance on very few players is also a problem in that industry.
Indeed, investors left to their own devices act in this way. Underlying such a single point-of-failure is an implied but immense hope and thus pressure for stability. I wonder what the prediction markets saying about current levels of geopolitical stability in Taiwan?
> rests on the shoulders of a single company that has all of their important factories on a single island
Isn't this just taking the oft-proposed explore vs exploit dichotomy to the logical conclusion of the "exploit" side?
Every single arbitrarity-finely-divided thing "should" be handled by the single (group|process) that has the greatest relative advantage at that one thing.
And you end up with the total variety/detailedness of everything matching what the substrate of the economy (ie, people with specialized training or education) has capacity to support. So at the limit there is at most one person who knows how to do any one specific thing.
(And the global economic system becomes infinitely fragile, but eh who's counting.)
It's only this way because the American ruling class would rather ship jobs overseas to increase their wealth than competently establish an industrial sector that would pay good wages to average people.
Turns out letting a bunch of MBAs plan your economy is extremely foolish.
> It's still absolutely fascinating to me that basically the whole modern tech industry and the economic growth from it rests on the shoulders of a single company
Stop getting your news from news.
> that has all of their important factories on a single island that's under constant threat of invasion.
Threat of invasion? Who would dare invade taiwan when it's protected by china?
> I don't know if TSMC has anything to do with hard drive production
Then why bother commenting here?
> but the reliance on very few players is also a problem in that industry.
If AI continues at this trajectory, sure, likely to the picks and shovels.
If AI has a bubble burst, you could see a lot of used hardware flood the market and then companies like WD could have a hard time selling against their previous inventory.
The problem is more likely that companies like WD doesn't know if this will be a bubble or not. Currently they can milk the market by raising their prices and just rely on their current production facilities, maybe expand a little. If there's going to be crash, then it's better to have raised the price, even if just temporarily, rather than being left standing with excessive production capacity.
If it's long term, it would be better to be the front runner on additional capacity, but that's assuming continuous growth. If it all comes down, or even just plateaus, it's better to simply raise prices.
Given how hard AI is on I/O, while restarting if hardware might go second hand. I dont see hard drives go second hand. Most hardware that we get might be used beyond redeeming even at free price.
What are companies needing all of these hard drives for? I understand their need for memory, and boot. But storing text training data and text conversations isn't that space intensive. There's a few companies doing video models, so I can see how that takes a tremendous amount of space. Is it just that?
Hearing about their scrapping practises it might be that they are storing same data over and over and over again. And then yes, audio and video is likely something they are planning for or already gathering.
And if they produce lot of video, they might keep copies around.
All the latest general purpose models are multimodal (except DeepSeek I think). Transfer learning allows to improve results even after they exhausted all the text in the internet.
Speaking from personal experience.. we treat cloud storage like an infinitely deep bucket. At rest data efficiency is not really a consideration because compute costs are so absurd. Why worry about a $2M year storage bill when your compute bill is $500M? It’s not worth the engineering time to optimize
I think the somewhat hallucinatory canned response is that they distribute data across drives for a massive throughput. Though idk if that even technically makes sense...
Everyone moved to SDDs or NVMe. If you're right, that includes manufacturers. HDDs still have advantages over SSDs for specific needs, like more reliable long-term unelectrified storage. It's also possible that the high price of SSDs made HDDs an option again.
I was recently involved in a large server purchase for work, where we wanted 72 hard drives of 24TB each for a server. They were available last year, but last month the largest we could get were 20TB drives.
And what's wrong with not wanting to write functions yourself? It is a perfectly reasonable thing, and in some cases (ex: crypto), rolling your own is strongly discouraged. That's the reason why libraries exist, you don't want to implement your own associative array every time your work needs it do you?
As for plagiarism, it is not something to even consider when writing code, unless your code is an art project. If someone else's code does the job better then yours, that's the code you should use, you are not trying to be original, you are trying to make a working product. There is the problem of intellectual property laws, but it is narrower than plagiarism. For instance, writing an open source drop-in replacement of some proprietary software is common practice, it is legal and often celebrated as long as it doesn't contain the original software code, in art, it would be plagiarism.
Copyright laundering is a problem though, and AI is very resource intensive for a result of dubious quality sometimes. But that just shows that it is not a good enough "plagiarism machine", not that using a "plagiarism machine" is wrong.
If I use a package for crypto stuff, it will generally be listed as part of the project, in an include or similar, so you can see who actually wrote the code. If you get an LLM to create it, it will write some "new original code" for you, with no ability to tell you any of the names of people who's code went into that, and who did not give their consent for it to be mangled into the algorithm.
If I copy work from someone else, whether that be a paragraph of writing, a code block or art, and do not credit them, passing it off as my own creation, that's plagiarism. If the plagiarism machine can give proper attribution and context, it's not a plagiarism machine anymore, but given the incredibly lossy nature of LLMS, I don't foresee that happening. A search engine is different, as it provides attribution for the content it's giving you (ignoring the "ai summary" that is often included now). If you go to my website and copy code from me, you know where the code came from, because you got it from my website
The ones who spend billions on integrating public cloud LLM services are not the ones writing that function. They are managers who based on data pulled out of thin air say "your goal for this year is to increase productivity by X%. With AI, while staffing is going slightly down".
I have to watch AI generated avatars on the most boring topics imaginable, because the only "documentation" and link to actual answer is in a form of fake person talking. And this is encouraged!
Then the only measure of success is either AI services adoption (team count), or sales data.
That is the real tragedy and the real scale - big companies pushing (external!) AI services without even proof that it justifies the cost alone. Smooth talking around any other metric (or the lack of it).
In my experience LLMs mimic human thought, so they don't "copy" but they do write from "experience" -- and they know more than any single developer can.
So I'm getting tired of the argument that LLMs are "plagiarism machines" -- yes, they can be coaxed into repeating training material verbatim, but no, they don't do that unless you try.
Opus 4.6's C compiler? I've not looked at it, but I would bet it does not resemble GCC -- maybe some corners, but overall it must be new, and if the prompting was specific enough as to architecture and design then it might not resemble GCC or any other C compiler much at all.
Not only do LLMs mimic human thinking, but also they mimic human faults. Obviously one way in which they mimic human faults is that there are mistakes in the LLMs' training materials, so they will evince some imperfections, and even contradictions (since there will be contradictions in their training materials). Another way is that their context windows are limited, just like ours. I liken their hallucinations to crappy code written by a tired human at 3AM after a 20 hour day.
If they are so human-like, we really cannot ascribe their output to plagiarism except when prompted so as to plagiarize.
LLMs just predict the next token. They mimic humans because they were trained on terabytes of human-created data (with no credit given to the authors of the training data). They don't mimic human thinking. If they did, you would be able to train them by themselves, but if you do that you get Model Collapse
6-7 years ago when GPU prices went up, I hoped nothing would break. Last year when RAM prices went up I did the same. Now with drive prices going up, it's the same thing.
It's interesting because I've always built mid-tier machines over the years and it was in the neighborhood of ~$700 at the time. Now the same thing is almost double that but the performance is no where near twice as good for general computer usage.
This is all basically a textbook example of irrational market decisions. There’s clearly a bubble and not enough money coming in to pay for the AI bonanza.
It’s building materials being in short supply when there’s obviously more houses than buyers. That’s just masked at the moment because of all the capital being pumped in to cover for the lack of actual revenue to pay for everything. The structural mismatch at the moment is gigantic, and the markets are getting increasingly impatient waiting for the revenue to materialize.
Mark this post… in a few years folks will be coming up with creative ideas for cheap storage and GPUs flooding the market after folks pick up the pieces of imploded AI companies.
(For the record, I’m a huge fan of AI, but that doesn’t mean I don’t also think a giant business and financial bubble is about to implode).
> in a few years folks will be coming up with creative ideas for cheap storage and GPUs flooding the market
COVID was six years ago. In that time, GPU prices haven't gone down (and really have only increased). Count me skeptical that there will be a flood of cheap components.
> It’s building materials being in short supply when there’s obviously more houses than buyers.
That I think is a hard one to prove and is where folks are figuring it out. There is obvious continued demand and certainly a portion of it is from other startups spending money. I don’t think it’s obvious though where we are at.
These really are some of the toughest years for people trying to buy a computer. We only recently emerged from the cryptocurrency-driven crunch, and now AI has arrived—and this wave is far more severe than crypto.
In that light, Apple’s Macs are actually quite good value for money.
I console myself with knowledge of the economics maxim that every supply shortage is usually, eventually, followed by a supply glut.
One can only hope that that's the principle at work here, anyway. It could also be a critically damped system for all I know. Unfortunately I studied control systems too...
If storage and memory manufacturer don't respond with increasing supply. There might not be glut. Just postponed demand that will slowly get fulfilled over longer period. That is if we were in steady state.
On other hand if there is bigger economic turmoil that might mean that the postponed demand does not realise as there is no purchasing power...
People with a control theory background are welcome in economics; the field is more diverse than some would recognize. Certain professions and subfields are more open than others. There are plenty of economists who care about things like resilience and dampening shocks.
I would love if more non-traditional economists got involved in the public sphere by which I mean: writing about economic trends, public policy, regulation, rate-adjustment, etc.
As an engineer with a passing control theory background and a breadth of general knowledge, I'd love to explore this space more and find a way to apply my knowledge and share the results. Are there any particular problems you think well-suited to this treatment?
I am interested in hearing how to get hard drives to last longer. Should you keep them locked away in the closet? Spin them up occasionally but not too much? Keep them always-on? I understand the less reading and writing, the better.
How does external compare to internal, if at all? Is 3.5" going to last longer than something smaller?
Spinning HDDs will eventually be at risk of failing for purely mechanical reasons, so beyond handling them with care you can't really do all that much. Keeping them always-on may be a viable strategy for drives that are already mostly on, otherwise, just spin them up once in a while, but don't expect this to lower risk significantly. An old drive should simply have its contents transferred to new media, and then be treated as something ephemeral that may fail at any time without warning.
I'm not against subsidies but the concentration is a problem. This money could have spurned grass roots participation in these emerging industries, but instead they chose to go with the most heinous of monocultures, leaving billions of people out of the loop.
The only way to combat this is the same way to combat toilet paper shortages.
Market says the more you buy the better pricing you get. Once you start capturing large market share of the product, the price should go up and not down; exponentially.
Example, a person that owns 10 houses means that they are restricting the ability of others to own a single home. By increasing the cost of excessive product ownership ... it will reduce the amount of product that people will hoard and allow others to gain access to it.
If you try to use government to force reality to conform to your idea of how things should work you're just going to get 1,000 companies buying 10,000 hard disks each rather than 10 companies buying 1,000,000 each. And if you try to outlaw that somehow then the market will just route around your new scheme in another way, creating even more unintended consequences in the process.
If you must meddle, you're much better off working with market forces rather than trying to fight against them.
>The only way to combat this is the same way to combat toilet paper shortages.
I buy one hard disk, AI company busy 40% of global supply. Me not buying that one hard disk is not going to change anything.
>According to Western Digital, thanks to a surge in demand from its enterprise customers, the consumer market now accounts for just 5 percent of the company's revenue.
It's interesting to see here that spending is irrational, but actually even if AI improvements slow down it's more rational for companies to spend more and underutilize the machines than to underspend and get disrtupted.
On the otherhand lots of people here are even more uncomfortable of the other option, which is quite possible: AI software algorithms may scale better than the capacity of companies that make the hardware. Personally I think hardware is the harder to scale from the two and this is just the beginning.
Great. I’ve just returned a WD drive to Amazon after it arrived crushed in a torn-open paper bag.
The replacement arrived also in a paper bag and went straight back, this time for a refund.
I guess I should have kept that one and hoped for the best.
Good alternatives? I’ve only recently been enlightened on how profoundly sh__ty SSD is for long-term storage and I have a whole lot of images my parents took traveling the last few years of their lives.
This is getting ridiculous. Never before has an unwanted product been thrust so forcefully and artificially into the market that it disrupts the supply line for real products with actual demand.
HDDs, RAM, and chips have so many health metrics and methods that you really shouldn't be afraid to buy them used. The only special requirement is a RasPi test rig. That and a 30 day return window.
I feel like the goal is to avoid buying something broken in the first place, not just to be able to tell if you've bought something that turns out to be broken
I listed some hard drives on Friday on eBay.. most of them refurbished... within 5 minutes got a message from a person who wanted them all... shipped them an hour later
I built a new server this time last year. My board does 6 channel RAM so I bought 6x32GB ECC DDR5. $160 a stick at the time. Just for grins I looked up the same product number at the same supplier I originally bought from. $1300 apiece. One of the VMs running on that server is TrueNAS, with 4 20TB WD Red Pros. God help me if I have to replace a drive.
I bought 6x refurbished ultrastars for ~$100/ea Black Friday 2024. They were over $200/ea 2025. Samsung T7 (and shield) SSD’s have 2x-3x. Can’t get 1TB for less than like $180 right now. It’s ridiculous
I pray someone steps into the market and takes all their consumer/SME business from them. I know thats not that simple or at all probable, but it would be great time to take market share and great for us little guys.
The reason this is a problem is because whatever value AI may have (personally I’m as long as one can get), companies believe that right now it’s a government sponsored financial bubble.
So they’re unwilling to spend on increasing capacity because they don’t expect this demand to last.
Any point in setting my laptop to minimal power usage by default and drives to aggressively sleep, in order to try to extend life? Or is sleep and extra power cycle and better to not sleep my M2 drives so they aren't being powered up/down?
The main thing that shortens the lifetime of solid state drives is sustained writes over time. You should disable all system logging options and all "spotlight databases for fast search", avoid swap use and not let the drives "sleep" too aggressively since this will force the system to persist some volume of writes from RAM to disk that might have turned out to be unnecessary down the road.
I'm running multiple services such as Forgejo, Audiobookshelf, Castopod and they all need no more than roughly 100 MB RAM.
There is one exception though. Open WebUI with a whopping 960 MB. It's literally a ChatGPT interface. I'm only using external API providers. No local models running.
Meanwhile my website that runs via my own Wordpress-like software written in Rust [1] requires only a few MB of RAM, so it's possible.
Is this an inevitable future? The amount of people ready to cede their computational resources, thinking, digital sovereignty, to centralised platforms, all in the name of progress, is truly shocking to me, especially in the current political moment.
The main reason I do not prioritise AI usage in my own life is to retain my skills and mental acuity. All of the forms of computing and opportunities that I value do not require AI to achieve. I can understand why people feel differently from me, though, because AI and AI-adjacent things are where all of the money is right now.
You know what is the sad part. I don't think software developers or LLMs know how or want anymore to make low resource consumption software that runs on a thin client. It will be some browser based thing capping to whatever memory is available on the system.
Even if the AI bubble bursts, having successfully cornered the compute market they can just go rent seeking instead by renting out cloud workstations, given that they've made the hardware to build a workstation yourself unaffordable.
Do they really think they will get some money from the AI ponzi scheme ?
Well, at least they might still have a product to sell once the AI bubble pops, unlike with NVIDIA which does seem to kinda forgot to design new consumer GPUs after getting high on AI money.
They haven't forgotten, they've expressly decided to soft-pivot away from consumer GPUs. RTX 60x0 series is apparently coming in 2018… (oops, 2028. No time travel involved. Probably). If the bubble has burst by then.
the current spikes tempt me to sell off my home lab. a mac mini to sell to the open claw bros, 5tb HDD, Intel NUC, some SSDs, and a 5 year old dell laptop. can always buy back after the crash.
It includes all forms of storage except for USB devices, GPUs and high end CPUs. The latter you can still get but you're going to have some severe sticker shock.
I read it as both, but UK suppliers have stock of various SATA HDDs available in large and small sizes. It's hard to say if prices will rocket or availability decline, or both. I don't normally advocate panic-buying, but if it's needed now is the time. I have one NAS spare on hand, I don't want or need a drawer full of them, but it'll be a royal pain if I do and can't get parts.
Lower performance/capacity consumer drives might be comparatively safe because there's Chinese end-to-end production capacity for those. Of course the price can still increase, but probably not that much.
I might have been more concerned if it was a different drive manufacturer.. some users won't forgive them for the WD Red debacle where they lied about what was in the drive.
There's clearly easy/irrational money distorting the markets here. Normally this wouldn't be a problem: prices would go up, supply would eventually increase and everybody would be okay. But with AI being massively subsidized by nation-states and investors, there's no price that is too high for these supplies.
Eventually the music will stop when the easy money runs out and we'll see how much people are truly willing to pay for AI.
Regardless where demand comes from, it takes time to spin up a hard drive factory, and prices would have to rise enough that, as a producer, you would feel confident that a new hard drive factory will actually pay off. Conversely, if you feel that boom is irrational and temporary, as a producer you’d be quite wary of investing money in a new factory if there was a risk it would be producing into a glut in a few years.
I'll add that the GPU, CPU, storage, and RAM industries crashed in 2022 after a Covid-induced boom.[0]
Everything was cheap. Samsung sold SSDs at a loss that year.
TSMC and other suppliers did not invest as much in cap ex in 2022 and 2023 because of the crash.
Parts of the shortage today can be blamed by those years. Of course ChatGPT also launched in late 2022 and the rest is history.
[0]www.trendforce.com/presscenter/news/20221123-11467.html
15 replies →
If I remember during a previous GPU shortage (crypto?), Nvidia (and/or TSMC?) basically knew the music would stop and didn't want to be caught with its pants down after making the significant investments necessary to increase production
Not to mention that without enough competition, you can just raise prices, which, uh (gestures at Nvidia GPU price trends...)
19 replies →
You're talking about how higher prices can motivate higher supply. The parent commenter was talking about how higher prices shift the current point on the demand curve to the right. If hard drives sold for $1 billion per gigabyte, we wouldn't see even AI companies buying as many as they are, and current production would go idle. Even assuming supply is locally inelastic (as it is given no time or space to scale, or given a lack of confidence that scaling is wise), you should be able to find a price point that avoids supply shortages by manipulating demand.
Thus far, we've not found that point.
> it takes time to spin up a hard drive factory
Very good.
The problem with this expectation of usual market behavior is that demand from AI will still be unsatisfied even after buying out the current providers' whole supply, so any new manufacturer entering the market will also prioritize high-paying AI companies above consumers.
Are these factories already running 24/7 that labor can't be added to make more without adding capital infra?
And if they were running 24/7, maybe setting up another factory or line will avoid some of the 24/7 scheduling.
1 reply →
No it’s not an easy fix. Manufacturers don’t have a good pulse on long term demand. The he capex to spin up a new manufacturing plant is significant. Especially with the recency of Covid where some folks did get caught with their pants down and over invested during the huge demand boom.
I don’t quite follow the narrative like yours about nation states and investors. There is certainly an industrial bubble going on and lots of startups getting massive amounts of capital but I here is a strong signal that a good part of this demand is here to stay.
This will be one of those scenarios where some companies will look brilliant and others foolish.
Smart manufacturers will sell 'hard drive futures'. Ie. "Give us $100/drive now for 100k drives for delivery in march 2028".
These contracts are then transferrable. The manufacturer can start work on a factory knowing they'll get paid to produce the drives.
If the AI boom comes to an end, the manufacturer is still going to get paid for their factory, and if the AI company wants to recoup costs they could try to sell those contracts back to the manufacturer for pennies on the dollar, who might then decide (if it is more profitable) to halt work on the factory - and either way they make money.
7 replies →
> Eventually the music will stop when the easy money runs out and we'll see how much people are truly willing to pay for AI.
Cheap hard drives and ram, yay! Perhaps GPUs too!
You wish. More likely all that data center capacity will be used to sell something as nefarious, like VDI for the masses. You won't need RAM, disk and GPUs when you can rent those from OpenVDI.
It's hard to increase long-run production capacity for what seems to be clearly a short-term spike in datacenter buildout. Even if AI itself is not much of a bubble, at some point spending on new AI facilities has to subside.
This is what a business cycle looks like.
Seeing the first mover succeed, every Tom, Dick and Harry wants to emulate. It distorts the price because people would pay premium for everything. Then there is surplus supply and no takers. People are caught with their pants down and things go for cheap.
This repeats ad nauseum. Whether it was building ISPs during early 2000s or the abundance of streaming service where every media company wanted one. Just because the corporate overlord doesn't want to look foolish for not following a trend.
AI is going to be what fiber was to the dotcom bubble. Someone spend a lot of money on a lot of infrastructure, some of which is going to be incredibly useful, but sold for much less than it cost to build. Hardware just depreciates much much faster than fiber networks.
I'm not saying that data center buildouts can't overshoot demand but AI and compute is different than fiber buildout. The more compute you have, the smarter the AI. You can use the compute to let the AI think longer (maybe hours/days/weeks) on a solution. You can run multiple AI agents simultaneously and have them work together or check each other's work. You can train and inference better models with more compute.
So there is always use for more compute to solve problems.
Fiber installations can overshoot relatively easily. No matter how much fiber you have installed, that 4k movie isn't going to change. The 3 hours of watch time for consumers isn't going to change.
9 replies →
current shortages are exactly the result of fabs not wanting to commit extra capex due to overbuild risk and inference demand seems to be growing 10x yoy; you've famously got 8 year old TPUs at google at 100% load.
The manfucaturing capacity expanded to meet the demand for new hardware doesn't (as much)
4 replies →
This goes beyond profits. It will be important for national security.
No, I think it is real demand.
AI will cause shortages in everything from GPUs to CPUs, RAM, storage, networking, fiber, etc because of real demand. The physical world can't keep up with AI progress. Hence, shortages.
AI simply increases computer use by magnitudes. Now you can suddenly use Seedance 2.0 to make CGI that would have cost tens of millions 5 years ago for $5.[0] Everyone is going to need more disk space to store all those video files. Someone in their basement can make a full length movie limited only by imagination. The output quality keeps getting better quicker.
AI agents also drastically increase storage demands. Imagine financial companies using AI agents to search, scrape, organize data on stocks that they wouldn't have been able to do prior. Suddenly, disk storage and CPUs are in high demand for tasks like these.
I think the demand for computer hardware and networking gear is real and is only the beginning.
As someone who is into AI, hardware, and investing, I've been investing in physical businesses based on the above hypothesis. The only durable moats will be compute, energy, and data.
[0]https://seed.bytedance.com/en/seedance2_0
> The only durable moats will be compute, energy, and data
"Compute" is capital investment; normal and comprehensible, but on a huge scale.
"Data" is .. stolen? That feels like a problem which has been dodged but will not remain solved forever, as everyone goes shields-up against the scrapers.
"Energy" was a serious global problem before AI. All economic growth is traded off against future global temperature increases to some extent, but this is even more acute in this electricity-intensive industry. How many degrees of temperature increase is worth one .. whatever the unit of AI gain-of-function is?
3 replies →
The question isn’t if the demand is real or not (supplies are low, so demand must exist). The question is if the demand curve has permanently shifted, or is this a short-term issue. No one builds new capacity in response to short term changes, because you’ll have difficulty recouping the capital expense.
If AI will permanently cause an increase in hard drives over the current growth curve, then WD, et al will build new capacity, increasing supply (and reducing costs). But this really isn’t something that is known at this point.
7 replies →
I wonder if I'm alone in being optimistic about this. I believe that the gigantic inflow of money into hardware will lead to large increase in production capabilities, accelerated progress and perhaps even new, better architectures.
1 reply →
> AI will cause shortages in everything from GPUs to CPUs, RAM, storage, networking, fiber, etc because of real demand.
Real demand, sure, I agree, but maybe not retail or business demand; at the moment the "demand" is entirely VC demand.
It's a really distorted market which is to be expected in any bubble/hype phase. The current retail/business demand doesn't appear to exist at the price point these investments require - even at the low low cost of "free, gratis and for nothing", not enough consumers and businesses are signing up.
The ones really going all-in on AI are the slop-producers. I dunno if slop is enough to pay back the investment into AI - I mean, even the slop producers are going to realise that paying $200/m to produce something in 1/10th of the time is a race to the bottom because someone else on the same plan is going to do the same, but cheaper.
> The physical world can't keep up with AI progress. Hence, shortages.
I think the word "progress" is inaccurate there - the physical world supply product at the demand maintained by VC's money.
It's not "cannot keep up with progress", it's "cannot keep up with demand from VCs".
> The only durable moats will be compute, energy, and data.
That'll be a first :-) Physical commodities have never been moats on their own before.
This fact never ceases to amaze me. It's so cool how relentlessly AI is pushing the horizons of our current hardware!
Maybe now we will start to see the "optical" CPUs start to be a thing. Or the 3D disk storage,;or other ground breaking technology.
1 reply →
AI's output is not reproducible. It's a disaster.
5 replies →
> Normally this wouldn't be a problem: prices would go up, supply would eventually increase
Sounds right
> there's no price that is too high for these supplies.
Are you saying even higher prices won't increase supply? I don't understand.
A comment above yours clearly explains why:
https://news.ycombinator.com/item?id=47034480
Earlier gamers got punished by crypto and now they are being punished by AI.
"Punished" implies a moral valence to the whole thing which isn't there. It's not like the AI companies were aware of gamers and set out to do this. You simply got run over, like everyone else in front of the trillion dollar bulldozer.
1 reply →
So what?
Why gamers must be the most important group?
11 replies →
GPUs before crypto had a lot less amount of VRAM. Crypto investment funded a lot of stupid experiments, of which some did stick to the wall. I don't think gamers had lives completely ruined by crypto in the end.
1 reply →
> Normally this wouldn't be a problem: prices would go up, supply would eventually increase and everybody would be okay.
This sounds like economic dogma based on pointing at some future equilibrium.
I like the saying that goes something like "life is what is happens when you are waiting for the future". In the same way, it seems to me that equilibrium is increasingly less common for many of us.
Markets are dynamic systems, and there are sub-fields of economics that recognize this. The message doesn't always get out unfortunately.
> But with AI being massively subsidized by nation-states and investors, there's no price that is too high for these supplies.
This feels like more dogma: find a convenient scape-goat: governments.
Time to wake up to what history has shown us! Markets naturally reflect boom and bust cycles, irrationality of people, and various other market failures. None of these are news to competent economists, by the way. Be careful from whence you get your economic "analysis".
Yes, this is why the prices of housing has dropped dramatically. The market stepped up and filled the demand needed and now everyone can afford a place to live
.....
5 replies →
Higher price encourages more supply. Typically when you see a acute shortage, its quickly followed by a glut as supply starts coming online in an over correction.
These factories take years to make and massive amounts of money. That and there are so few manufacturers now they are far more likely to collude
Loved the reference. Probably from Margin Call[0]
0. https://youtu.be/fij_ixfjiZE
I like to imagine the reference in the movie margin call is that of a merry go round or a game of Musical chair. Like we are all on a ride, none of us are the operator, and all we can do is guess when the music will stop (and the ride ends).
The problem with this AI stuff is we don't know how much we will be willing to pay for it, as individuals, as businesses, as nations. I guess we just don't know how far this stuff will be useful. The reasons for the high valuation is, in my guess, that there is more value here than what we have tapped so far, right?
The revenues that nVidia has reported is based on what we hope we will achieve in the future so I guess the whole thing is speculation?
2 replies →
I think AI companies are involving these other industries so when the money runs out they will claim the whole thing is too big too fail.
By buying flash and thus shifting demand to HDD? How does that work?
1 reply →
The quote was from the screenwriter he never said it.
Better stock up with used laptops. I'm going to buy another one this year. Those used ones usually don't last very long.
What if in the near future it is simply too expensive to own "personal" computers? What if you can no longer buy used computers from official channels but have to find local shops or sharpen up on soldering skills and find parts from dumps? The big techs will conveniently "rent out" cloud computer for us to use, in exchange of all of your data.
"Don't you all have cellphones?"
Bingo. A number of corporate interests don't want to let you own your personal computers for different reasons. Google/Apple wants you to get locked down devices, and cloud/AI providers want you to use their services from a weak client.
Time for folks to familiarize themselves with Linux distros designed to run on older hardware. My 2009 laptop runs great, with the exception of the browser. Oh and the fact that 32-bit software is harder and harder to find.
A full-sized laptop from 2009 will probably boot a 64-bit distro, even Pentium M was no longer in use by then.
Yes, I'm grateful I run Linux. You can get quite a bit done with 4GB RAM and a 6th generation (or even earlier) CPU. All 64-bit. I don't think such ancient hardware will be affected by AI demand to the same degree (though I think we'll still see some prices rise if people stop buying new stuff).
The worry is that at some point the older hardware will stop working.
I think about this too. There are several headwinds. Rent-seeking and collapse of economies of scale in the consumer sector for sure, but also I feel like we've basically peaked in hardware's ability to meet routine needs.
Once the phone makers realize that they can sell phones and docking stations to businesses because 90% of knowledge work seems to happen in a web browser through one SaaS or other I think personal computers will be cooked.
I don't think phones are really any cheaper than the mini PCs that businesses can already buy. Which makes sense, because a phone has to include a battery, touch screen, and is under tighter space constraints.
1 reply →
I have heard that you can get used laptops. But they do not come with memory or SSD anymore... As even used components are now valuable enough to be removed and sold.
Many laptops from last few years have soldered memory. Your previous laptop's SSD can also be reused, since those don't die that quickly compared to the laptop.
In a lot of cases, owners remove the storage not because it has any value but rather they don't want to risk making a mistake letting a device go that still has data on it.
Also pulling and shredding hard drives is cheaper than paying someone to run DBAN or equivalent (which can take many hours to complete on a large drive), and there's no easy way to securely erase an SSD if it wasn't encrypted from the beginning.
Or worse, they have memory and SSD soldered on board, and are broken, so you have to learn soldering skills too.
Damn really? One of my go to moves when helping small political campaigns is to buy like a 2015 MBP and turn it into a locally hosted server to run all their stuff on the cheap.
1 reply →
I have 3 old employer laptops and my personal gaming laptop, which I use for work now. I'm happy about this now ;)
I probably will only need to return newest laptop if I leave the company.
Non state of the art lithography is pretty much commoditized (DDR3 & DDR4) so we will always have compute, although slower.
I no longer feel obligated to apologize for holding on to older devices for a long time. I have several desktops and laptops that are all still usable.
Why laptops specifically (and not also desktops)?
Eh, if this demand is really sustainable they will eventually start producing in adequate volume
It's still absolutely fascinating to me that basically the whole modern tech industry and the economic growth from it rests on the shoulders of a single company that has all of their important factories on a single island that's under constant threat of invasion. On top of that they themselves are reliant on a single company that's able to produce the machine required to print the wafers.
I don't know if TSMC has anything to do with hard drive production, but the reliance on very few players is also a problem in that industry.
Investors love a monopoly, and establishing this required more than a trillion dollars of investment sustained over a couple decades.
> Investors love a monopoly...
Indeed, investors left to their own devices act in this way. Underlying such a single point-of-failure is an implied but immense hope and thus pressure for stability. I wonder what the prediction markets saying about current levels of geopolitical stability in Taiwan?
3 replies →
> rests on the shoulders of a single company that has all of their important factories on a single island
Isn't this just taking the oft-proposed explore vs exploit dichotomy to the logical conclusion of the "exploit" side?
Every single arbitrarity-finely-divided thing "should" be handled by the single (group|process) that has the greatest relative advantage at that one thing.
And you end up with the total variety/detailedness of everything matching what the substrate of the economy (ie, people with specialized training or education) has capacity to support. So at the limit there is at most one person who knows how to do any one specific thing.
(And the global economic system becomes infinitely fragile, but eh who's counting.)
There are three pillars for the bleeding edge, aren't there? TSMC, ASML, and the Spruce Pine quartz mine.
It's only this way because the American ruling class would rather ship jobs overseas to increase their wealth than competently establish an industrial sector that would pay good wages to average people.
Turns out letting a bunch of MBAs plan your economy is extremely foolish.
Hey now they went to school for at least 1.5 years and not all of that was at a 9th grade reading level! Some of it was 10th grade
> under constant threat of invasion
And isn't it also in a seismically active region, also prone to eathquake and/or tsunami?
Prone to earthquake yes. Tsunami no
> It's still absolutely fascinating to me that basically the whole modern tech industry and the economic growth from it rests on the shoulders of a single company
Stop getting your news from news.
> that has all of their important factories on a single island that's under constant threat of invasion.
Threat of invasion? Who would dare invade taiwan when it's protected by china?
> I don't know if TSMC has anything to do with hard drive production
Then why bother commenting here?
> but the reliance on very few players is also a problem in that industry.
Ah, you have a political agenda.
Are these the picks and shovels?
Is the profitability of these electronics manufacturers more likely than the companies that are buying up all their future inventory?
If AI continues at this trajectory, sure, likely to the picks and shovels.
If AI has a bubble burst, you could see a lot of used hardware flood the market and then companies like WD could have a hard time selling against their previous inventory.
The problem is more likely that companies like WD doesn't know if this will be a bubble or not. Currently they can milk the market by raising their prices and just rely on their current production facilities, maybe expand a little. If there's going to be crash, then it's better to have raised the price, even if just temporarily, rather than being left standing with excessive production capacity.
If it's long term, it would be better to be the front runner on additional capacity, but that's assuming continuous growth. If it all comes down, or even just plateaus, it's better to simply raise prices.
Given how hard AI is on I/O, while restarting if hardware might go second hand. I dont see hard drives go second hand. Most hardware that we get might be used beyond redeeming even at free price.
8 replies →
What are companies needing all of these hard drives for? I understand their need for memory, and boot. But storing text training data and text conversations isn't that space intensive. There's a few companies doing video models, so I can see how that takes a tremendous amount of space. Is it just that?
Hearing about their scrapping practises it might be that they are storing same data over and over and over again. And then yes, audio and video is likely something they are planning for or already gathering.
And if they produce lot of video, they might keep copies around.
All the latest general purpose models are multimodal (except DeepSeek I think). Transfer learning allows to improve results even after they exhausted all the text in the internet.
Speaking from personal experience.. we treat cloud storage like an infinitely deep bucket. At rest data efficiency is not really a consideration because compute costs are so absurd. Why worry about a $2M year storage bill when your compute bill is $500M? It’s not worth the engineering time to optimize
Storing training data: for example, Anthropic bought millions of second hand books and scanned them:
https://www.washingtonpost.com/technology/2026/01/27/anthrop...
All of Annas archive can be put on 40 drives
2 replies →
I think the somewhat hallucinatory canned response is that they distribute data across drives for a massive throughput. Though idk if that even technically makes sense...
I am surprised by that too. I thought everyone moved to SDDs or NVMe ?
I was toying with getting a 2T HDD for a BSD system I have, I guess not now :)
Everyone moved to SDDs or NVMe. If you're right, that includes manufacturers. HDDs still have advantages over SSDs for specific needs, like more reliable long-term unelectrified storage. It's also possible that the high price of SSDs made HDDs an option again.
1 reply →
I was recently involved in a large server purchase for work, where we wanted 72 hard drives of 24TB each for a server. They were available last year, but last month the largest we could get were 20TB drives.
This is the consequence of "I don't want to write this function myself, I'll get the plagiarism machine to do it for me"
And what's wrong with not wanting to write functions yourself? It is a perfectly reasonable thing, and in some cases (ex: crypto), rolling your own is strongly discouraged. That's the reason why libraries exist, you don't want to implement your own associative array every time your work needs it do you?
As for plagiarism, it is not something to even consider when writing code, unless your code is an art project. If someone else's code does the job better then yours, that's the code you should use, you are not trying to be original, you are trying to make a working product. There is the problem of intellectual property laws, but it is narrower than plagiarism. For instance, writing an open source drop-in replacement of some proprietary software is common practice, it is legal and often celebrated as long as it doesn't contain the original software code, in art, it would be plagiarism.
Copyright laundering is a problem though, and AI is very resource intensive for a result of dubious quality sometimes. But that just shows that it is not a good enough "plagiarism machine", not that using a "plagiarism machine" is wrong.
If I use a package for crypto stuff, it will generally be listed as part of the project, in an include or similar, so you can see who actually wrote the code. If you get an LLM to create it, it will write some "new original code" for you, with no ability to tell you any of the names of people who's code went into that, and who did not give their consent for it to be mangled into the algorithm.
If I copy work from someone else, whether that be a paragraph of writing, a code block or art, and do not credit them, passing it off as my own creation, that's plagiarism. If the plagiarism machine can give proper attribution and context, it's not a plagiarism machine anymore, but given the incredibly lossy nature of LLMS, I don't foresee that happening. A search engine is different, as it provides attribution for the content it's giving you (ignoring the "ai summary" that is often included now). If you go to my website and copy code from me, you know where the code came from, because you got it from my website
4 replies →
I honestly think it's not that simple.
The ones who spend billions on integrating public cloud LLM services are not the ones writing that function. They are managers who based on data pulled out of thin air say "your goal for this year is to increase productivity by X%. With AI, while staffing is going slightly down".
I have to watch AI generated avatars on the most boring topics imaginable, because the only "documentation" and link to actual answer is in a form of fake person talking. And this is encouraged!
Then the only measure of success is either AI services adoption (team count), or sales data.
That is the real tragedy and the real scale - big companies pushing (external!) AI services without even proof that it justifies the cost alone. Smooth talking around any other metric (or the lack of it).
In my experience LLMs mimic human thought, so they don't "copy" but they do write from "experience" -- and they know more than any single developer can.
So I'm getting tired of the argument that LLMs are "plagiarism machines" -- yes, they can be coaxed into repeating training material verbatim, but no, they don't do that unless you try.
Opus 4.6's C compiler? I've not looked at it, but I would bet it does not resemble GCC -- maybe some corners, but overall it must be new, and if the prompting was specific enough as to architecture and design then it might not resemble GCC or any other C compiler much at all.
Not only do LLMs mimic human thinking, but also they mimic human faults. Obviously one way in which they mimic human faults is that there are mistakes in the LLMs' training materials, so they will evince some imperfections, and even contradictions (since there will be contradictions in their training materials). Another way is that their context windows are limited, just like ours. I liken their hallucinations to crappy code written by a tired human at 3AM after a 20 hour day.
If they are so human-like, we really cannot ascribe their output to plagiarism except when prompted so as to plagiarize.
LLMs just predict the next token. They mimic humans because they were trained on terabytes of human-created data (with no credit given to the authors of the training data). They don't mimic human thinking. If they did, you would be able to train them by themselves, but if you do that you get Model Collapse
[dead]
My machine was built up from parts in 2014.
6-7 years ago when GPU prices went up, I hoped nothing would break. Last year when RAM prices went up I did the same. Now with drive prices going up, it's the same thing.
It's interesting because I've always built mid-tier machines over the years and it was in the neighborhood of ~$700 at the time. Now the same thing is almost double that but the performance is no where near twice as good for general computer usage.
Yeah, this is slowing down growth and profits. The AI hype is sucking everything dry, from HVAC services to hardware.
This is all basically a textbook example of irrational market decisions. There’s clearly a bubble and not enough money coming in to pay for the AI bonanza.
It’s building materials being in short supply when there’s obviously more houses than buyers. That’s just masked at the moment because of all the capital being pumped in to cover for the lack of actual revenue to pay for everything. The structural mismatch at the moment is gigantic, and the markets are getting increasingly impatient waiting for the revenue to materialize.
Mark this post… in a few years folks will be coming up with creative ideas for cheap storage and GPUs flooding the market after folks pick up the pieces of imploded AI companies.
(For the record, I’m a huge fan of AI, but that doesn’t mean I don’t also think a giant business and financial bubble is about to implode).
> in a few years folks will be coming up with creative ideas for cheap storage and GPUs flooding the market
COVID was six years ago. In that time, GPU prices haven't gone down (and really have only increased). Count me skeptical that there will be a flood of cheap components.
I feel like the most recent time you could reasonably get an nvidia *80 GPU at the store for a normal amount of money was almost a decade ago.
2 replies →
That was largely due to the crypto mining bubble popping just as the AI bubble was getting started
Is there an industrial bubble? Probably.
> It’s building materials being in short supply when there’s obviously more houses than buyers.
That I think is a hard one to prove and is where folks are figuring it out. There is obvious continued demand and certainly a portion of it is from other startups spending money. I don’t think it’s obvious though where we are at.
More reckless and irresponsible than irrational.
These really are some of the toughest years for people trying to buy a computer. We only recently emerged from the cryptocurrency-driven crunch, and now AI has arrived—and this wave is far more severe than crypto. In that light, Apple’s Macs are actually quite good value for money.
I console myself with knowledge of the economics maxim that every supply shortage is usually, eventually, followed by a supply glut.
One can only hope that that's the principle at work here, anyway. It could also be a critically damped system for all I know. Unfortunately I studied control systems too...
If storage and memory manufacturer don't respond with increasing supply. There might not be glut. Just postponed demand that will slowly get fulfilled over longer period. That is if we were in steady state.
On other hand if there is bigger economic turmoil that might mean that the postponed demand does not realise as there is no purchasing power...
I was thinking than until my NAS gave me a error on one of my harddrives, now I'm in the market for a replacement while I still have redundancy
People with a control theory background are welcome in economics; the field is more diverse than some would recognize. Certain professions and subfields are more open than others. There are plenty of economists who care about things like resilience and dampening shocks.
I would love if more non-traditional economists got involved in the public sphere by which I mean: writing about economic trends, public policy, regulation, rate-adjustment, etc.
As an engineer with a passing control theory background and a breadth of general knowledge, I'd love to explore this space more and find a way to apply my knowledge and share the results. Are there any particular problems you think well-suited to this treatment?
2 replies →
I am interested in hearing how to get hard drives to last longer. Should you keep them locked away in the closet? Spin them up occasionally but not too much? Keep them always-on? I understand the less reading and writing, the better.
How does external compare to internal, if at all? Is 3.5" going to last longer than something smaller?
Spinning HDDs will eventually be at risk of failing for purely mechanical reasons, so beyond handling them with care you can't really do all that much. Keeping them always-on may be a viable strategy for drives that are already mostly on, otherwise, just spin them up once in a while, but don't expect this to lower risk significantly. An old drive should simply have its contents transferred to new media, and then be treated as something ephemeral that may fail at any time without warning.
I'm not against subsidies but the concentration is a problem. This money could have spurned grass roots participation in these emerging industries, but instead they chose to go with the most heinous of monocultures, leaving billions of people out of the loop.
There's always πfs. https://github.com/philipl/pifs
The only way to combat this is the same way to combat toilet paper shortages.
Market says the more you buy the better pricing you get. Once you start capturing large market share of the product, the price should go up and not down; exponentially.
Example, a person that owns 10 houses means that they are restricting the ability of others to own a single home. By increasing the cost of excessive product ownership ... it will reduce the amount of product that people will hoard and allow others to gain access to it.
The laws of supply and demand are not optional.
If you try to use government to force reality to conform to your idea of how things should work you're just going to get 1,000 companies buying 10,000 hard disks each rather than 10 companies buying 1,000,000 each. And if you try to outlaw that somehow then the market will just route around your new scheme in another way, creating even more unintended consequences in the process.
If you must meddle, you're much better off working with market forces rather than trying to fight against them.
This only works if you exclude datacenters Georg who lives in a server rack and buys 10,000 hard drives a day.
Unfortunately, there are several such outlier entities which collectively control enough resources to price literally everyone else out.
>The only way to combat this is the same way to combat toilet paper shortages.
I buy one hard disk, AI company busy 40% of global supply. Me not buying that one hard disk is not going to change anything.
>According to Western Digital, thanks to a surge in demand from its enterprise customers, the consumer market now accounts for just 5 percent of the company's revenue.
[dead]
It's interesting to see here that spending is irrational, but actually even if AI improvements slow down it's more rational for companies to spend more and underutilize the machines than to underspend and get disrtupted.
On the otherhand lots of people here are even more uncomfortable of the other option, which is quite possible: AI software algorithms may scale better than the capacity of companies that make the hardware. Personally I think hardware is the harder to scale from the two and this is just the beginning.
Great. I’ve just returned a WD drive to Amazon after it arrived crushed in a torn-open paper bag.
The replacement arrived also in a paper bag and went straight back, this time for a refund.
I guess I should have kept that one and hoped for the best.
Good alternatives? I’ve only recently been enlightened on how profoundly sh__ty SSD is for long-term storage and I have a whole lot of images my parents took traveling the last few years of their lives.
I'm sure Amazon isn't the only shop that delivers to your area
The premise of this news is that prices are going to climb and availability is going to drop.
And I’m not keen on having anyone ship me one of these anymore.
Walmart sells what appears to be an older version of the drive and I might have to cross my fingers and just get one of those.
2 replies →
Bestbuy with local pickup. They price match Amazon. B&H Photo is another option.
It started with RAM; now with hard drives and SSDs. This is not looking good. But at least you can buy used ones for a pretty good price, for now.
This is getting ridiculous. Never before has an unwanted product been thrust so forcefully and artificially into the market that it disrupts the supply line for real products with actual demand.
Presumably they're also looking to increase production capacity as fast as possible - within the year?
I'd have thought HDDs aren't at the top of the list for AI requirements, are other component manufacturers struggling even more to meet demand?
Why would they?
If we weren’t talking about AI, was there another high demand sector / customer for spinning platters?
And their margins get fat now that supply is relatively constant but AI demand has saturated their current production numbers.
HDDs, RAM, and chips have so many health metrics and methods that you really shouldn't be afraid to buy them used. The only special requirement is a RasPi test rig. That and a 30 day return window.
I feel like the goal is to avoid buying something broken in the first place, not just to be able to tell if you've bought something that turns out to be broken
I listed some hard drives on Friday on eBay.. most of them refurbished... within 5 minutes got a message from a person who wanted them all... shipped them an hour later
I built a new server this time last year. My board does 6 channel RAM so I bought 6x32GB ECC DDR5. $160 a stick at the time. Just for grins I looked up the same product number at the same supplier I originally bought from. $1300 apiece. One of the VMs running on that server is TrueNAS, with 4 20TB WD Red Pros. God help me if I have to replace a drive.
Best Buy is actively selling 2x8gb sticks of DDR4 3200 for $80 a stick. I was floored. Ten bucks a gig, $160 for the pack.
We're fucking doomed.
Ten bucks a gig is lower than what some DDR5 memory is selling at.
Perhaps there is an incentive to go back to OS that can operate with 640KB RAM ... /s
Is this for NVME only or spinning drives too? I use both, but I actually have use cases for HDDs and hope those are less affected.
It’s affecting both. HDD maybe slightly less/slower, but you’re paying significantly more than six months ago in any case.
All I know is I saw most of my go-to refurbished enterprise HDD’s 2-3x during Black Friday a few months ago compared to a year prior.
This particular news is for spinning drives, the other types we already had news about upcoming shortages earlier on.
I bought 6x refurbished ultrastars for ~$100/ea Black Friday 2024. They were over $200/ea 2025. Samsung T7 (and shield) SSD’s have 2x-3x. Can’t get 1TB for less than like $180 right now. It’s ridiculous
Bought a few 2TB T7 shield disks last year before the boom. Thank fuck I did it then.
Take good care of them!
i am happy i bought 5x10TB drives two months ago, anticipating this exact scenario.
Not only storage, cheapest 32 GB RAM that I can find is around 200 euros.
That's actually a bargain, average market price (though highly volatile) is more than double that.
They're probably looking at DDR4.
sighs at her local backup drive that just gave up the ghost
thanks, AI-boosting motherfuckers, thanks a lot
Rotating or SSD?
I pray someone steps into the market and takes all their consumer/SME business from them. I know thats not that simple or at all probable, but it would be great time to take market share and great for us little guys.
The reason this is a problem is because whatever value AI may have (personally I’m as long as one can get), companies believe that right now it’s a government sponsored financial bubble.
So they’re unwilling to spend on increasing capacity because they don’t expect this demand to last.
Any point in setting my laptop to minimal power usage by default and drives to aggressively sleep, in order to try to extend life? Or is sleep and extra power cycle and better to not sleep my M2 drives so they aren't being powered up/down?
The main thing that shortens the lifetime of solid state drives is sustained writes over time. You should disable all system logging options and all "spotlight databases for fast search", avoid swap use and not let the drives "sleep" too aggressively since this will force the system to persist some volume of writes from RAM to disk that might have turned out to be unnecessary down the road.
Good luck to everyone. Home you made some reserve.
Yes, AI is nice, but I also like to be able to buy some RAM and drives…
The future is thin clients for everyone, requiring a minimal amount of RAM and storage because all they are is a glorified ChatGPT interface.
I'm running multiple services such as Forgejo, Audiobookshelf, Castopod and they all need no more than roughly 100 MB RAM.
There is one exception though. Open WebUI with a whopping 960 MB. It's literally a ChatGPT interface. I'm only using external API providers. No local models running.
Meanwhile my website that runs via my own Wordpress-like software written in Rust [1] requires only a few MB of RAM, so it's possible.
[1]: https://github.com/rikhuijzer/fx
Is this an inevitable future? The amount of people ready to cede their computational resources, thinking, digital sovereignty, to centralised platforms, all in the name of progress, is truly shocking to me, especially in the current political moment.
The main reason I do not prioritise AI usage in my own life is to retain my skills and mental acuity. All of the forms of computing and opportunities that I value do not require AI to achieve. I can understand why people feel differently from me, though, because AI and AI-adjacent things are where all of the money is right now.
3 replies →
You know what is the sad part. I don't think software developers or LLMs know how or want anymore to make low resource consumption software that runs on a thin client. It will be some browser based thing capping to whatever memory is available on the system.
Even if the AI bubble bursts, having successfully cornered the compute market they can just go rent seeking instead by renting out cloud workstations, given that they've made the hardware to build a workstation yourself unaffordable.
It won't last. If the demand is sustained then new factories will open up and drive the price down.
More likely a couple of big financing wobbles lead to a fire sale.
It isn't practical for HDD supply to be wedged because in 5 years the disks start failing.
Do they really think they will get some money from the AI ponzi scheme ?
Well, at least they might still have a product to sell once the AI bubble pops, unlike with NVIDIA which does seem to kinda forgot to design new consumer GPUs after getting high on AI money.
They haven't forgotten, they've expressly decided to soft-pivot away from consumer GPUs. RTX 60x0 series is apparently coming in 2018… (oops, 2028. No time travel involved. Probably). If the bubble has burst by then.
> RTX 60x0 series is apparently coming in 2018
That's either a typo, or NVidia has achieved some previously unheard of levels of innovation.
1 reply →
> "apparently coming in 2018… maybe. If the bubble has burst by then."
Spoiler from the future: it hasn't. Get your investments in while you have time.
Take a look at prices of SSDs and RAM too.
the current spikes tempt me to sell off my home lab. a mac mini to sell to the open claw bros, 5tb HDD, Intel NUC, some SSDs, and a 5 year old dell laptop. can always buy back after the crash.
You could buy a house and a private jet with these if you held just one more year.
does that only include SSDs, or does it include HDDs as well?
It includes all forms of storage except for USB devices, GPUs and high end CPUs. The latter you can still get but you're going to have some severe sticker shock.
Maybe shucking USB HDDs is the short-term answer.
11 replies →
I read it as both, but UK suppliers have stock of various SATA HDDs available in large and small sizes. It's hard to say if prices will rocket or availability decline, or both. I don't normally advocate panic-buying, but if it's needed now is the time. I have one NAS spare on hand, I don't want or need a drawer full of them, but it'll be a royal pain if I do and can't get parts.
Lower performance/capacity consumer drives might be comparatively safe because there's Chinese end-to-end production capacity for those. Of course the price can still increase, but probably not that much.
1 reply →
I might have been more concerned if it was a different drive manufacturer.. some users won't forgive them for the WD Red debacle where they lied about what was in the drive.
Not only does AI want to take my job, it also wants to make my hobbies unaffordable.
I love modern world so much /s