I saw this video yesterday and considered posting it, but I wasn’t sure if it was appropriate for HN.
This channel has another video where it shows how the clean room lab is created starting from a basic backyard shed, and that was truly astounding. The positive pressure to keep the number of particles low in someone’s backyard is almost mystical to me.
Recently I saw a post about Bonsai trees on the front page. Making your own RAM is 100% more relevant to HN than quite a few posts I see on the main page.
You’re not sure if someone building a RAM clean room in a shed is appropriate for HackerNews, literally “news for nerds”? A dictionary purchase may be warranted
To put it that into context, some tags count as upvotes, others count as downvotes, "Troll" is a downvote. So to have your post labelled as "Troll" with a positive score, it has to have enough upvotes to compensate the penalty from the "Troll" votes, but without having another tag dominate. 5 is the maximum score.
"Score: 5, Troll" is therefore the mark of a very successful troll.
They also have "Underrated" and "Overrated" which apply points but do not act as tags. So I guess the easiest way to get +5 Troll is to have many Troll and Underrated votes, if it works the way I think it does.
I think he plans to go far beyond just making RAM in that clean room. This is pure speculation, but I suspect the goal of that channel is to just make doom from scratch.
Given that the shed in this guy’s backyard is already approaching the entire national technological output of any country in the 1970s I think he may get there.
There's no mention of AGI, climate change, AWS outages, Trump, crypto schadenfreude or my new MVP that you should totally sign up for even though I just vibe coded it 20 minutes ago and the DNS hasn't fully propagated yet, but the API is amazing plz like comment and subscribe.
Ok, maybe I'm being a bit cynical. Stories about bikeable cities are welcome too. And wasn't Soylent popular for a hot minute back in the day?
Which is a pity, because lots of videos really need to be seen to be fully appreciated. Especially the ones showing stuff being made. And the ones that tend to show up here are usually worth the time.
I'm totally with the text folks on the 5 hour Fossdem sessions, though. Give me an accurate transcript I can grep or don't even bother.
I think if it's interesting to you then it's worth posting, and letting the voting system do it's thing. I only rarely post because by the time I've seen something it's usually already been posted
why even allow the cognitive overhead to worry about such a thing? it's not for you to decide anyway - let the users decide using the voting system that's built to task
2028. How to forage for edible plants and identify safe water sources
2030. How to make horseshoes, nails, and plate armor
2032. How to send smoke signals to neighboring villages
2034. How to defend against barbarian incursions
Jokes aside, seeing as this person has created their own clean room in a shed, and is making RAM, what exactly is stopping any company from doing this themselves and breaking into the RAM business?
I'd pay less for RAM that wasn't "certified" in some official way, at least it works.
> what exactly is stopping any company from doing this themselves and breaking into the RAM business?
nothing, except the terrible yields that they would obtain, and the lack of scale making the entire enterprise not profit generating (as the amount of profit per sale is too low if it even is positive, but you can't set it higher as there's cheaper, "better" ram available from pre-established fabs that do have economies of scale).
You could play the artisanal angle, and market it as home grown, organic ram. Not sure how much real buyers of ram care, but might get a few hobbyists in the market.
The angle right now I think is pretty obvious, there is a massive shortage that might cause actual incidents.
OpenAI should do their own production, I say slightly bitter because I'm in a health care sector that might be affected because we can't scale up or repair our infrastructure due to their massive pre-orders.
It's really easy to set up a manufacturing process for basically anything if you can spend 100x per unit compared to the big optimized factories and you don't mind the product being a lot slower.
The clean room isn't the hard part about being competitive. It's using advanced lithography to cram billions of cells into a single chip. If you want make DDR2 chips on a 90 nanometer process, that is accessible to a whole lot of companies, but nobody will buy the product. And in the micrometer range you can DIY like this guy.
Correct, nobody will buy your 1GB stick of DDR2-speed RAM for the $100 it cost you to produce it.
> And what happens when smaller companies have to repair or scale their infrastructure and can't get affordable RAM?
That situation sucks but bringing up more obsolete fabrication isn't going to help. They can't compete with modern chips even when those modern chips have a 5x price penalty.
Ok sorry I misread you saying that it was easy with nm manufacturing but nobody will buy the product, you said it's easy to manufacture micrometer DDR2 speeds.
Single digit micrometer is really easy and makes toys and/or microcontrollers. 90nm is sort of easy if you have factory money, and is about what you need for DDR2. It gets a lot more difficult as you go beyond that.
In one of Dwarkesh's interviews, he mentioned that China is trying to replicate the entire stack. Ironically now that they have mastered all pieces of the stack for older nodes, they actually have an advantage in a collapse scenario. The US does not appear to have the ability to do all steps in the stack for any node. They still rely on other western countries that could go offline. China despite being behind does at least have top to bottom capability for older nodes. Combine that with their rock bottom electricity prices and they have a unique card that they can play.
Just imagine if electricity costs were trending towards 0. Instead of e-waste run all those machines till the chips burn out.
You can make eDRAM using logic processes (which are currently less bottlenecked than RAM, at least for non cutting-edge nodes) but the cost is still prohibitive compared to specialty DRAM processes (even when considering the recent increase in DRAM prices). If you were doing that, you'd want to use it for compute-in-memory instead, which basically pushes NUMA to an extreme of having lots of tiny cores each with direct access to its own local DRAM.
I'm not really sure what you mean by "certified"; I don't think JEDEC are handing out stickers. Although I am reminded of the Bunnie Huang article about SD cards, where the Shenzen vendor would give you the same SD card with the manufacturer logo + holographic authenticity .. of your choice.
The real problem with the RAM business is that it was commodified; normally manufacturers make a relatively small margin. No incentive to build a factory for that. These are not normal times because (a) someone has bought all the RAM and (b) someone has blown up a whole load of globally critical infrastructure in the Middle East.
The risk the existing RAM manufacturers are being cautious about is the risk of normal: if you start building a factory now, will you be selling into a RAM glut?
i assume the reason is that this is a very competitive market and you need hundreds of billions in investment to just start producing at a competitive quality and price, with massive uncertainty that you will be able to make that money back
i mean, you'd think if someone is willing to through 60B (or 10B for an option to buy at a 60B valuation) at Cursor, then other people would give it a shot at starting a SotA semiconductor fab, but apparently the profit expectations are not there
EUV lithography is the state of the art. It makes far denser chips and is quite out of reach for the backyard fellow. Find a documentary on how ASML machines work: they're near the pinnacle of human accomplishment!
He's producing semiconductors with a 1000nm (one micron) feature size. This kind of tech was cutting edge in the mid 80s. You might be able to produce a 32KB memory chip with it.
It would be difficult to break into the RAM business with that sort of product as most of the demand these days is for higher capacities.
So, I get that we charge the capacitor up, and that it leaks so we must recharge it periodically. I don't get two things:
1. How is the value read? Is it reading the leak?
2. How is recharging done? I guess the leak itself (assuming my guess in 1. holds) could provide charge for some logic that checks "if has charge then recharge else nop".
I still don't really get transistors :P, but this was cool.
It turns out they intentionally drain a bit of the storage capacitor, and amplify that weak signal. Some of that amplified charge is then fed back to storage.
A transistor effectively is an AND gate. If there is a charge on both the source and the gate, then charge can move to drain. So if you charge up the capacitor and you connect it to the source of another transistor then you can check whether it is still charged by putting a charge to the gate and see if there is charge on the drain.
And you are right, that charge on the drain can then be used both to drive some logic and to activate the recharging of the capacitor that was just discharged.
By the way I am being handwavy about "charge" moving about, if you really want to learn the electronics it is more correct to call it a voltage relative to some ground that the charge always moves towards.
You measure the charge (if it is there) before it is completely gone (if it were ever there). Capacitors always leak, these capacitors leak fast.
Measuring the charge also removes some of the charge -- fast, compared to the leak spead.
DRAM chips have a circuit that writes the value back -- charges the capacitor up if there's supposed to be a charge, drains it if there's not supposed to be a charge.
Refreshes and normal reads are the same, except that normal reads sends the value(s) to the output pin(s) on the chip.
He has "only" shown the basic grid of capacitors and transistors. The chip he shown has no circuit to measure charges or to write them back afterwards. This makes it easier to test the basic grip and the basic capcacitors.
Pretty sure the proper read out and write back circuit comes in the next video.
In the "real" DRAM chip, there is a large array of very tiny capacitors, with the switches which allow to connect one row of the array at a time to the readout column wires.
The capacitance of the wires themselves is typically an order of magnitude greater than that of the storage capacitors. So when the memory is read, the wires are first precharged to some standard voltage. Then the desired row of storage capacitors is connected to the wires, and the charges from the storage capacitors spread onto the wires, changing their voltages very slightly. These voltage deviations from the standard value are amplified by the "sense amplifiers". The amplifiers are sort of like flip-flops. Once they start in a state which "tilts" slightly to "zero" or slightly to "one", they go all the way to the full magnitude zeros and ones. This not only amplifies the signal, but also automatically brings the voltages on the wires and the still connected to them capacitors to the full magnitude, thus "refreshing" the data. The row is disconnected, and the next read cycle can start for some other row.
In the video, an array of 4x5 capacitors and the associated with them switches was fabricated. The capacitors in the video are several hundred times larger (12400 fF) than typical capacitors in a 64 Kbit DRAM (about 50 fF). I assume this is done so that in the later episodes the author could implement the readout electronics outside of the chip.
- any two conductive plates close together with an insulator between make a capacitor (1)
- when a capacitor is charged, the energy goes into the electric FIELD
- that's the FIELD of the field effect transistor
- if the field is strong enough, that causes conduction between source and drain (along the "channel")
- the insulator is nanometers thick, so current leaks across it; at that scale, you can detect individual electrons quantum teleporting straight through it.
(1) technically like gravity there is a capacitance between any two objects in the universe, but it's only significant when you have relatively wide and close conductors
This is the most pathological technology I have heard of in a long time and I am not even going to apologize for upvoting and telling other people about the evil genius I found on the internet today.
I've thought about doing stuff like that (or mounting S3), but I've never done it because I figured that the latency would be enormous and it would be too slow to actually be useful.
That said, I could actually kind of see it being useful for stuff that doesn't need to be super fast but does allocate a lot of RAM (though I'm drawing a blank on what that would be at the moment).
Enormous latency is all relative. A network drive on a local network holding a swap file would still outperform some number of computers I've owned that put their swap on much, much slower hard drives of their time. Of course, nobody was trying to swap two gigabytes to these drives as that would have been 10 times their capacity....
If they put a pricing page, I think there would be someone who would buy it, especially nowadays when with embedded llms there is a huge hunger for RAM (as well as CPU). :))
I was thinking... will x402 protocol make it super easy for scammers to commit such frauds in future? By tricking online searches done by LLM's to trick them into spending money?
Is it? Looks like this video is "locked" on Patreon requiring "this post and more exclusive work. Join for $10/month" yet got, as of now, 329 611 views ... so are they just making $3M/month or is it not really working?
What are you talking about? The video is free on YouTube. Patreon is just another way of supporting the channel beyond the usual ad-sense, sponsors, etc.
No one is expecting them to make a fortune over night, but it at least shows there is space for new people. This video was actually push into my feed by youtube which is rare for new creators. So they clearly are doing pretty well in the grand scheme of things.
What do you mean with "What are you talking about?" while my comment literally post to the Patreon account and I copy/pasted their message.
The comment of the video also says "Thanks for the support!" whereas I (maybe you are a subscriber to their Patreon but I know I'm not) have not contributed anything.
Does this look normal to have locked videos behind a paywall then have it public and free elsewhere?
PS: I obviously didn't expect them to make this amount (even though it would be nice) only trying to highlight that there is a mismatch and thus trying to understand what the problem is, namely is that video actually meant to be public.
Well that's the first time I encounter this. I have been using Patreon for years and typically a creator who locks content even often mention "Please do not share outside". It's also often hosted on YouTube but usually unlisted.
Can you please give me an example of locked content that is openly shared outside? I guess I don't understand the monetization in that situation so I'd appreciate if you could help me get what they are doing here.
It appears that the creator built a setup inspired by HackerFab, which is a collection of open source tools and resources for chip manufacturing. It's a fantastic project, and I recommend people check it out:
Technologies that were considered "high-tech" 20-30 years ago are now accessible to regular people. Making DRAM. I remember a video recently of a guy making his own floppy disks.
Next I expect people to manufacture integrated circuits that have been discontinued. Like the Z80
It is not really feasible to fabricate usable integrated circuits at home. There is a huge difference between a one off demonstration of a principle that "sort of works", and perfecting the process to the point that it produces finished parts that can be relied on.
This guy is not exactly a regular person. He is a pretty unique case of a talented semiconductor engineer who has a home lab for side hassles. It is not a low effort thing. He runs the equipment 24/7, scrubs all the surfaces in the lab daily to keep it clean.
Still, with the lab and all of the equipment already at hand, it cost him several weeks of work to produce this demonstration of transistors and capacitors, which kind of work, but are still long ways from a "completely complete" 20 bit DRAM chip.
Unfortunately it is simply too much work for one person to maintain a viable semiconductor fabrication process, even when it is done semi-professionally.
Backyard semiconductor production is pretty similar to backyard barbecue. Lots of heating, smoking (diffusion), injecting (ion implant), and layering..
There's another video on YouTube by the same guy detailing how he built his backyard clean room shed. I was kind of surprised at how easy it was - it's definitely a construction project that requires some specialized knowledge, but the fact that it's tractable at all for one person with a shed is pretty amazing to me.
Of course it not easy. He spent $20k on materials and a good amount of elbow grease just to construct this one room. In order to maintain the clean environment the air circulation runs 24/7 and he washes all the surfaces daily using cleanroom wipes. Deionized water system provides water for washing the cleanroom gowns. It is a lot of work just to have a small facility at home for odd jobs.
The furnace is a low cost off-the-shelf Chinese product from Anhui Beq Equipment Technology Co.
Much more impressive are the modifications to the microscope, transforming it into an improvised lithography machine, and the home made plasma etching machine, cobbled together from surplus components.
Of course, the whole thing, starting from the clean room, is extremely impressive -- Intel started their business in a much simpler facility.
I think I’ve seen a similar video where the person was explaining how to increase your RAM on your android device. I don’t know if it’s possible. I personally never tried it. has anyone tried it before?
Incredible video! I've been in tech for 20 years but never really bothered to dive in to the process of making RAM. This is eye-opening and makes me appreciate the sophistication of the tech involved in making those ram-sticks.
Good video, informative, easy to digest. You should start your own company, perhaps as in the form of an employee-owned entity or using Kickstarter to create an all-in-one RAM Fab kit that can be sold online. I'm sure the maker community may consider buying it.
The graphs towards the end were discharge curves for a single transistor/capacitor cell out of only 16 present, if I understood correctly? So "enough cells to count as memory" and "addressing logic" are definitely future work (it looked like he wanted to characterize what the refresh cycle would have to look like before actually building more.) I was kind of surprised that the "use a microscope as a photolithography projector" approach worked at all, it will be interesting to see how that scales up...
Considering how AI companies and older hardware manufacturing
companies are driving up the RAM prices - thus, milking all of
us ultimately - I think 3D printing needs to become the new
default. And affordable too, eventually (it already is, for
cheap things, e. g. plastics or PLA based printing, this is
often much cheaper than ordering this from a company, but I
mean with regards to computer systems too. Naturally right now
we are far away from the nanoscale here, but I see this simply
as a situation that will change eventually, given enough time).
Those AI companies and hardware manufacturers lost all right to
further dictate and increase prices. Capitalism does not work
as de-facto blackmail monopoly - or should not. If a state fails
to protect the people, such as in the USA right now under the
orange king, then the people need to insist on change. ALL
steps against this tyranny from a few superrich needs to end.
Right now the legislation is going in the way how lobbyists
want this, e. g. trying to make 3D printing illegal, but I
think technology will obsolete such illegal laws eventually.
Tyranny will eventually fail.
I do but my days in the fab taught me that you do NOT want people to do this, considering the extremely dangerous chemicals involved. People have died changing EMPTY tanks of phosphine gas used for doping… and HF acid used for etch is another nightmare entirely.
No acids at all? That would be stupendously difficult for no real benefit. So many things are acids, so many useful reactions involve acids, and there's not a significant correlation between "is an acid" and "danger".
No. Silicon oxide (glass) is extremely tough from a chemical perspective. That's why it's used in chemistry for everything. Barely anything touches it. Also this is the main reason I think that the meme of "silicon based life" is completely absurd and comes from people who only took high school chemistry and built their worldview on that.
Is it conceivable that some organic solvent could be synthesized that is, simultaneously harmless to water-based biological life, and capable of etching Silicon oxides?
Not really. Organics don't really have any affinity for this type of compound. You could, of course, create some kind of organic fluorinating compound, but it would basically just put you back at square one for safety.
It is only half as bad as working in the places that make tbose chemicals for use in clean rooms. Swaping out "empty" phosphine tanks is bad, but filling and shipping hundreds of full tanks is worse.
I used to graduate at an institute having physicists as well as chemists, I gues it was no coincidence that only physicists operated with HF, one chemist told me that no chemist in their right mind would touch it
That's not quite accurate (but close enough). We had HF in the chem lab. It lived in a dedicated metal box with a massive neon warning label and a padlock.
It's notable in comparison that all the deadly organics lived together in an unlocked cupboard (vented OFC). I think the only thing I ever saw treated as more of a pariah than HF were radioactive isotopes. Those generally get an entire dedicated room with restricted entry and a tedious mandatory cleaning procedure.
Makes sense. HF deserves the same awe as radioactive material. I've always found both fascinating. Like some kind of dark magic that curses you if you contact it.
HF is routinely used in analytical labs; it's standard to microwave HF solutions for ICP digestions. It's not even the most hazardous reagent in my lab right now.
Now, perhaps this chemist meant that no chemist in their right mind would physically touch HF--in that case, I agree completely!
This is the issue I have with people saying that solar power is "clean and eco friendly".
It sure is, if you ignore the fact that you have to have a factory to make it where one of the *nicest* things around is the fucking hydrofluoric acid, and most of the rest will kill you instantly in trace amounts.
The technology is, the production is not but you can contain that, at least in theory. Compare that with gasoline that everytime you obtain energy from it you burn it out of existence and create a mess of the environment.
This is why we should have converted all the cars to run on propane, instead of scrapping them in favour of "cleaner greener diesels" 20 years ago when they started all the "scrappage scheme" bollocks.
The propane is going to get burnt anyway. May as well extract some useful work from it, and when you run a car off it they become ultra low emission.
I visited a pcb making factory once. Left with an appreciation for the amount of work needed for 80-layer pcbs, and knowing I would not want to deal with making them myself.
my goal was simple 2 sided pcbs, machined traces because i wanted to avoid chemical etching, but when it comes to via's chemicals are really the only way. The chemcials needed for plating via's are very toxic. my current thinking to avoid the really nasty ones is to try conductive ink (probably pretty bad too) but it maybe would work to coat the fr4 material and then allow a copper plating to take... really it's a fun process machining and laser the soldier mask.
I keep thinking that for home tinkering this is really the wrong approach. Surely there are other more DIY-friendly ways to make switches besides with semiconductors? Sure, they wouldn't achieve anywhere near the same density as SOTA semiconductors, but that's not really possible at home anyway.
I guess you could always go back to electromechanical relays. I wonder if it might be possible to come up with a method of 3D printing those?
But in practical terms the Z80 was a 4 um process node so unless you're willing to go back to the proverbial stone age it seems like you need semiconductors and lithography.
Magnetic core memory final form were single large perforated plates with many conductors plated on the ferrite surface through-holes, and only the vertical stack of bus wires were threaded through the plates. This meant weaving was less of an issue, and higher >1kiB modules were feasible in a smaller area. The main draw back is it sill had destructive read-once access, so always had higher latency in addition to being slow.
The DDR market will adapt, as China grey market state fab smells the opportunity. They have been counterfeiting cmos chips for decades already, and dram is not as complex as people like to assume.
Neuromorphic computing will likely kick over the LLM sand pile at some point, and all that discounted hardware will need re-homed. We can wait for the bubble to run its course, and actual investors realize they were conned. =3
> Neuromorphic computing will likely kick over the LLM sand pile at some point, and all that discounted hardware will need re-homed.
I don't see a lot of work going on in neuromorphic - there was some work at Intel, IIRC. Not saying you're wrong, but just wondering where you think it's going to come from?
> just wondering where you think it's going to come from?
Will likely evolve like any regular biological system, and consume translated LLM weight sets into its initial training condition 3D propagation structure.
The speed at which this occurs will likely initially be measured in weeks due to slower growth state writes, but once bootstrapped the GC is self-propagating.
I normally don't like to speculate, but the barrier to entry would actually be much lower than traditional silicon fabrication processes. It was an old idea from science fiction, that until recently was highly impractical. Asimov was likely wrong about the physical process, but not about how it is made.
Due to theoretical constant morphological changes under GC, one must acknowledge the inherent lack of safety such systems would pose. Have a nice day. =3
I expect you'd carry out most such work in the form of simulations, only moving to hardware once you'd demonstrated an efficient algorithm. If I'm right about that then it would be easy for any corporate R&D on the topic to fly under the radar indefinitely.
On the academic side of things there's a steady drip of papers on things like spiking neutral networks so I'd say the general theme is being explored.
I figure if a breakthrough happens it will be overnight just like what happened with transformers.
The next big shift will be HBF. All that DRAM holding essentially static weights that are read in nice, long linear reads in inference machines is wasted; if you had a proper interface to it you could replace it all with flash for a tenth of the cost.
I saw this video yesterday and considered posting it, but I wasn’t sure if it was appropriate for HN.
This channel has another video where it shows how the clean room lab is created starting from a basic backyard shed, and that was truly astounding. The positive pressure to keep the number of particles low in someone’s backyard is almost mystical to me.
Recently I saw a post about Bonsai trees on the front page. Making your own RAM is 100% more relevant to HN than quite a few posts I see on the main page.
The HN crowd decides what is relevant
It's about intellectual curiosity, so it's both.
You’re not sure if someone building a RAM clean room in a shed is appropriate for HackerNews, literally “news for nerds”? A dictionary purchase may be warranted
Agree with the sentiment, but “news for nerds” is Slashdot.
Slashdot still exists?
Well, “exists” is a pretty broad spectrum.
I miss the comment tagging system: insightful, informative, interesting, funny. It would make sense for hn.
You forgot Troll, you insensitive clod!
"Score: 5, Troll" is the ultimate achievement.
To put it that into context, some tags count as upvotes, others count as downvotes, "Troll" is a downvote. So to have your post labelled as "Troll" with a positive score, it has to have enough upvotes to compensate the penalty from the "Troll" votes, but without having another tag dominate. 5 is the maximum score.
"Score: 5, Troll" is therefore the mark of a very successful troll.
They also have "Underrated" and "Overrated" which apply points but do not act as tags. So I guess the easiest way to get +5 Troll is to have many Troll and Underrated votes, if it works the way I think it does.
you must be new here
Their standard is higher than that, "Stuff that matters."
I think he plans to go far beyond just making RAM in that clean room. This is pure speculation, but I suspect the goal of that channel is to just make doom from scratch.
Given that the shed in this guy’s backyard is already approaching the entire national technological output of any country in the 1970s I think he may get there.
In a comment he says he is doing it for some research into a related thing (somethiing to do with GaN sheets?)
Yeah but it's a YouTube video. Those tend not to do super well on the front page.
Just gimme the transcripts for a speed read
It’s perfectly ok to submit links that don’t reach the front page!
There's no mention of AGI, climate change, AWS outages, Trump, crypto schadenfreude or my new MVP that you should totally sign up for even though I just vibe coded it 20 minutes ago and the DNS hasn't fully propagated yet, but the API is amazing plz like comment and subscribe.
Ok, maybe I'm being a bit cynical. Stories about bikeable cities are welcome too. And wasn't Soylent popular for a hot minute back in the day?
If you haven't seen this one, I highly recommend it:
Indistinguishable From Magic: Manufacturing Modern Computer Chips
https://www.youtube.com/watch?v=NGFhc8R_uO4&t=2070s
It's quite old but I think there is no modern version of it.
I've tried posting to HN a few times but it hasn't gained traction for some reason, but I find it absolutely mind blowing.
Videos in general don’t get much traction here. Most of the time I don’t want to watch them in this context either, when other sites I do.
Maybe it’s just I come here for the old web feel when video was costly, rare and short.
Which is a pity, because lots of videos really need to be seen to be fully appreciated. Especially the ones showing stuff being made. And the ones that tend to show up here are usually worth the time.
I'm totally with the text folks on the 5 hour Fossdem sessions, though. Give me an accurate transcript I can grep or don't even bother.
Yeah, I understand and partially agree.
However I've discovered wonderful gems like this RAM video.
There's something to it. I personally am happy to have one of these few precious places left where I can find content to read rather than watch.
Video links are naturally gonna get less clicks from people scrolling HN at work :)
I rewatch that every year as a reminder of what we're capable of. Great video.
Yep, me too. Still feels magical after all these years.
The whole process was deep magic to me before I watched that video. It didn't seem less magical after I've watched it. More so, if anything.
Asianometry[0] has a number of videos on EUV lithography that cover some of the mind-blowing advances in the years since.
Veritasium[1] recently also made a video on the subject.
[0] https://www.youtube.com/playlist?list=PLKtxx9TnH76RYHY7L1YzE...
[1] https://www.youtube.com/watch?v=MiUHjLxm3V0
This was the one that did it for me: "38C3 - From Silicon to Sovereignty: How Advanced Chips Are Redefining Global Dominance" https://www.youtube.com/watch?v=NdppYYfQJgg
Absolutely insane stuff.
This one I did not know! Thanks for sharing
I think if it's interesting to you then it's worth posting, and letting the voting system do it's thing. I only rarely post because by the time I've seen something it's usually already been posted
Tbh this is exactly the sort of thing I'd come here to see
Yeah we need 20 more LLM submissions instead, that's the hard hitting content. /s
> but I wasn’t sure if it was appropriate for HN
why even allow the cognitive overhead to worry about such a thing? it's not for you to decide anyway - let the users decide using the voting system that's built to task
2027. Upgrade your LLM with home made RAM so you can afford to have it design your flying car
2027. Just-in-time built software and hardware.
2030. All flying cars are actually armed drones - homefab is outlawed
Turns out Cessnas are actually low performance missiles with an organic computer for the guidance system.
How to make an axe from a broken RTX 4060
I'm not sure this is what they meant when they said they wanted to bring manufacturing back to the USA lol
Chairman Trump will have a semiconductor furnace in every backyard.
You have to surrender your silicon(e) to the state!
Jokes aside, seeing as this person has created their own clean room in a shed, and is making RAM, what exactly is stopping any company from doing this themselves and breaking into the RAM business?
I'd pay less for RAM that wasn't "certified" in some official way, at least it works.
So, CXMT?
> what exactly is stopping any company from doing this themselves and breaking into the RAM business?
nothing, except the terrible yields that they would obtain, and the lack of scale making the entire enterprise not profit generating (as the amount of profit per sale is too low if it even is positive, but you can't set it higher as there's cheaper, "better" ram available from pre-established fabs that do have economies of scale).
You could play the artisanal angle, and market it as home grown, organic ram. Not sure how much real buyers of ram care, but might get a few hobbyists in the market.
The angle right now I think is pretty obvious, there is a massive shortage that might cause actual incidents.
OpenAI should do their own production, I say slightly bitter because I'm in a health care sector that might be affected because we can't scale up or repair our infrastructure due to their massive pre-orders.
Local, small-batch RAM! Top fermenting Get it at your local farmer's market!
It's really easy to set up a manufacturing process for basically anything if you can spend 100x per unit compared to the big optimized factories and you don't mind the product being a lot slower.
The clean room isn't the hard part about being competitive. It's using advanced lithography to cram billions of cells into a single chip. If you want make DDR2 chips on a 90 nanometer process, that is accessible to a whole lot of companies, but nobody will buy the product. And in the micrometer range you can DIY like this guy.
Nobody will buy the product?
I'm an infrastructure architect and work with health care, local tax agency, they're all getting 25-30% inflated bills now for new hardware.
And what happens when smaller companies have to repair or scale their infrastructure and can't get affordable RAM?
I'd say if people aren't desperate already, they're about to be.
> Nobody will buy the product?
Correct, nobody will buy your 1GB stick of DDR2-speed RAM for the $100 it cost you to produce it.
> And what happens when smaller companies have to repair or scale their infrastructure and can't get affordable RAM?
That situation sucks but bringing up more obsolete fabrication isn't going to help. They can't compete with modern chips even when those modern chips have a 5x price penalty.
Ok sorry I misread you saying that it was easy with nm manufacturing but nobody will buy the product, you said it's easy to manufacture micrometer DDR2 speeds.
Single digit micrometer is really easy and makes toys and/or microcontrollers. 90nm is sort of easy if you have factory money, and is about what you need for DDR2. It gets a lot more difficult as you go beyond that.
In one of Dwarkesh's interviews, he mentioned that China is trying to replicate the entire stack. Ironically now that they have mastered all pieces of the stack for older nodes, they actually have an advantage in a collapse scenario. The US does not appear to have the ability to do all steps in the stack for any node. They still rely on other western countries that could go offline. China despite being behind does at least have top to bottom capability for older nodes. Combine that with their rock bottom electricity prices and they have a unique card that they can play.
Just imagine if electricity costs were trending towards 0. Instead of e-waste run all those machines till the chips burn out.
Lucky for America, in the case of civilizational collapse there will be a lot of spare semiconductors thanks to almost everyone being dead!
You can make eDRAM using logic processes (which are currently less bottlenecked than RAM, at least for non cutting-edge nodes) but the cost is still prohibitive compared to specialty DRAM processes (even when considering the recent increase in DRAM prices). If you were doing that, you'd want to use it for compute-in-memory instead, which basically pushes NUMA to an extreme of having lots of tiny cores each with direct access to its own local DRAM.
I'm not really sure what you mean by "certified"; I don't think JEDEC are handing out stickers. Although I am reminded of the Bunnie Huang article about SD cards, where the Shenzen vendor would give you the same SD card with the manufacturer logo + holographic authenticity .. of your choice.
The real problem with the RAM business is that it was commodified; normally manufacturers make a relatively small margin. No incentive to build a factory for that. These are not normal times because (a) someone has bought all the RAM and (b) someone has blown up a whole load of globally critical infrastructure in the Middle East.
The risk the existing RAM manufacturers are being cautious about is the risk of normal: if you start building a factory now, will you be selling into a RAM glut?
i assume the reason is that this is a very competitive market and you need hundreds of billions in investment to just start producing at a competitive quality and price, with massive uncertainty that you will be able to make that money back
i mean, you'd think if someone is willing to through 60B (or 10B for an option to buy at a 60B valuation) at Cursor, then other people would give it a shot at starting a SotA semiconductor fab, but apparently the profit expectations are not there
EUV lithography is the state of the art. It makes far denser chips and is quite out of reach for the backyard fellow. Find a documentary on how ASML machines work: they're near the pinnacle of human accomplishment!
https://en.wikipedia.org/wiki/Extreme_ultraviolet_lithograph...
He's producing semiconductors with a 1000nm (one micron) feature size. This kind of tech was cutting edge in the mid 80s. You might be able to produce a 32KB memory chip with it.
It would be difficult to break into the RAM business with that sort of product as most of the demand these days is for higher capacities.
The US version of The Great Leap Forward
So, I get that we charge the capacitor up, and that it leaks so we must recharge it periodically. I don't get two things:
1. How is the value read? Is it reading the leak?
2. How is recharging done? I guess the leak itself (assuming my guess in 1. holds) could provide charge for some logic that checks "if has charge then recharge else nop".
I still don't really get transistors :P, but this was cool.
https://en.wikipedia.org/wiki/Dynamic_random-access_memory#P...
It turns out they intentionally drain a bit of the storage capacitor, and amplify that weak signal. Some of that amplified charge is then fed back to storage.
A transistor effectively is an AND gate. If there is a charge on both the source and the gate, then charge can move to drain. So if you charge up the capacitor and you connect it to the source of another transistor then you can check whether it is still charged by putting a charge to the gate and see if there is charge on the drain.
And you are right, that charge on the drain can then be used both to drive some logic and to activate the recharging of the capacitor that was just discharged.
By the way I am being handwavy about "charge" moving about, if you really want to learn the electronics it is more correct to call it a voltage relative to some ground that the charge always moves towards.
You measure the charge (if it is there) before it is completely gone (if it were ever there). Capacitors always leak, these capacitors leak fast.
Measuring the charge also removes some of the charge -- fast, compared to the leak spead.
DRAM chips have a circuit that writes the value back -- charges the capacitor up if there's supposed to be a charge, drains it if there's not supposed to be a charge.
Refreshes and normal reads are the same, except that normal reads sends the value(s) to the output pin(s) on the chip.
He has "only" shown the basic grid of capacitors and transistors. The chip he shown has no circuit to measure charges or to write them back afterwards. This makes it easier to test the basic grip and the basic capcacitors.
Pretty sure the proper read out and write back circuit comes in the next video.
I know how to spell "speed" and "grid". Honest!
In the "real" DRAM chip, there is a large array of very tiny capacitors, with the switches which allow to connect one row of the array at a time to the readout column wires.
The capacitance of the wires themselves is typically an order of magnitude greater than that of the storage capacitors. So when the memory is read, the wires are first precharged to some standard voltage. Then the desired row of storage capacitors is connected to the wires, and the charges from the storage capacitors spread onto the wires, changing their voltages very slightly. These voltage deviations from the standard value are amplified by the "sense amplifiers". The amplifiers are sort of like flip-flops. Once they start in a state which "tilts" slightly to "zero" or slightly to "one", they go all the way to the full magnitude zeros and ones. This not only amplifies the signal, but also automatically brings the voltages on the wires and the still connected to them capacitors to the full magnitude, thus "refreshing" the data. The row is disconnected, and the next read cycle can start for some other row.
In the video, an array of 4x5 capacitors and the associated with them switches was fabricated. The capacitors in the video are several hundred times larger (12400 fF) than typical capacitors in a 64 Kbit DRAM (about 50 fF). I assume this is done so that in the later episodes the author could implement the readout electronics outside of the chip.
> I still don't really get transistors :P,
The important bits here are:
- any two conductive plates close together with an insulator between make a capacitor (1)
- when a capacitor is charged, the energy goes into the electric FIELD
- that's the FIELD of the field effect transistor
- if the field is strong enough, that causes conduction between source and drain (along the "channel")
- the insulator is nanometers thick, so current leaks across it; at that scale, you can detect individual electrons quantum teleporting straight through it.
(1) technically like gravity there is a capacitance between any two objects in the universe, but it's only significant when you have relatively wide and close conductors
"There is no DownloadMoreRAM, it's just some guy in a backyard shed."
https://downloadmoreram.com/
With memory prices what they are maybe there is a business opportunity for a return of SoftRAM 95
Actually possible if you mount google drive and move your swap file there.
This is the most pathological technology I have heard of in a long time and I am not even going to apologize for upvoting and telling other people about the evil genius I found on the internet today.
icymi https://scp-iota.github.io/software/2025/06/16/download-ram-...
tldr it's not theorical, some madman actually did that!
How about swap on pingfs so you aren't relying on some cloud provider?
https://github.com/yarrick/pingfs
I've thought about doing stuff like that (or mounting S3), but I've never done it because I figured that the latency would be enormous and it would be too slow to actually be useful.
That said, I could actually kind of see it being useful for stuff that doesn't need to be super fast but does allocate a lot of RAM (though I'm drawing a blank on what that would be at the moment).
Enormous latency is all relative. A network drive on a local network holding a swap file would still outperform some number of computers I've owned that put their swap on much, much slower hard drives of their time. Of course, nobody was trying to swap two gigabytes to these drives as that would have been 10 times their capacity....
If they put a pricing page, I think there would be someone who would buy it, especially nowadays when with embedded llms there is a huge hunger for RAM (as well as CPU). :))
I was thinking... will x402 protocol make it super easy for scammers to commit such frauds in future? By tricking online searches done by LLM's to trick them into spending money?
This guy is proof that newcomers to YouTube can still succeed, if they find the right niche.
And are doing exceptional things (like building their own cleanroom in a shed)
Content creation was always about content. If you can make something special and engaging, audience will come.
> If you can make something special and engaging
... and sufficiently interesting to more than a very small group of persons.
Is it? Looks like this video is "locked" on Patreon requiring "this post and more exclusive work. Join for $10/month" yet got, as of now, 329 611 views ... so are they just making $3M/month or is it not really working?
What are you talking about? The video is free on YouTube. Patreon is just another way of supporting the channel beyond the usual ad-sense, sponsors, etc.
No one is expecting them to make a fortune over night, but it at least shows there is space for new people. This video was actually push into my feed by youtube which is rare for new creators. So they clearly are doing pretty well in the grand scheme of things.
What do you mean with "What are you talking about?" while my comment literally post to the Patreon account and I copy/pasted their message.
The comment of the video also says "Thanks for the support!" whereas I (maybe you are a subscriber to their Patreon but I know I'm not) have not contributed anything.
Does this look normal to have locked videos behind a paywall then have it public and free elsewhere?
PS: I obviously didn't expect them to make this amount (even though it would be nice) only trying to highlight that there is a mismatch and thus trying to understand what the problem is, namely is that video actually meant to be public.
> Does this look normal to have locked videos behind a paywall then have it public and free elsewhere?
Yes, this is very normal on youtube with both patreon and nebula.
Well that's the first time I encounter this. I have been using Patreon for years and typically a creator who locks content even often mention "Please do not share outside". It's also often hosted on YouTube but usually unlisted.
Can you please give me an example of locked content that is openly shared outside? I guess I don't understand the monetization in that situation so I'd appreciate if you could help me get what they are doing here.
The pattern I often see is a Patreon release a week before the public release.
You can go to their Patreon and see that they are making CAD545/month [1]
[1] https://www.patreon.com/cw/DrSemiconductor
Thanks, I do know, that what my point, to highlight the discrepancy between the two.
It appears that the creator built a setup inspired by HackerFab, which is a collection of open source tools and resources for chip manufacturing. It's a fantastic project, and I recommend people check it out:
https://docs.hackerfab.org/home
Technologies that were considered "high-tech" 20-30 years ago are now accessible to regular people. Making DRAM. I remember a video recently of a guy making his own floppy disks. Next I expect people to manufacture integrated circuits that have been discontinued. Like the Z80
It is not really feasible to fabricate usable integrated circuits at home. There is a huge difference between a one off demonstration of a principle that "sort of works", and perfecting the process to the point that it produces finished parts that can be relied on.
This guy is not exactly a regular person. He is a pretty unique case of a talented semiconductor engineer who has a home lab for side hassles. It is not a low effort thing. He runs the equipment 24/7, scrubs all the surfaces in the lab daily to keep it clean.
Still, with the lab and all of the equipment already at hand, it cost him several weeks of work to produce this demonstration of transistors and capacitors, which kind of work, but are still long ways from a "completely complete" 20 bit DRAM chip.
Unfortunately it is simply too much work for one person to maintain a viable semiconductor fabrication process, even when it is done semi-professionally.
"Side hassles"... hahaha. Love it.
Makes me wonder if in the far future it will be ever possible to do something crazy like 7nm processes at home
The Titanic is pretty old. Could you build that in your backyard? The scale of the equipment required has drastically increased over time.
That is really not the takeaway from this video, at all.
Backyard semiconductor production is pretty similar to backyard barbecue. Lots of heating, smoking (diffusion), injecting (ion implant), and layering..
With enough sauce, the end result won’t even taste like chewing on a rock. :-)
And lighting your charcoal with piranha solution...
Nobody tell openai about this, they’ll buy up all his stock
They will preemptively rent all the sheds in the usa.
Great time to get into the shed business!
I wasn't expecting what the inside of the shed would be like!
There's another video on YouTube by the same guy detailing how he built his backyard clean room shed. I was kind of surprised at how easy it was - it's definitely a construction project that requires some specialized knowledge, but the fact that it's tractable at all for one person with a shed is pretty amazing to me.
Of course it not easy. He spent $20k on materials and a good amount of elbow grease just to construct this one room. In order to maintain the clean environment the air circulation runs 24/7 and he washes all the surfaces daily using cleanroom wipes. Deionized water system provides water for washing the cleanroom gowns. It is a lot of work just to have a small facility at home for odd jobs.
This is incredible! 1100 degrees in your backyard shed! And the video explains it well too
The furnace is a low cost off-the-shelf Chinese product from Anhui Beq Equipment Technology Co.
Much more impressive are the modifications to the microscope, transforming it into an improvised lithography machine, and the home made plasma etching machine, cobbled together from surplus components.
Of course, the whole thing, starting from the clean room, is extremely impressive -- Intel started their business in a much simpler facility.
I think I’ve seen a similar video where the person was explaining how to increase your RAM on your android device. I don’t know if it’s possible. I personally never tried it. has anyone tried it before?
Incredible video! I've been in tech for 20 years but never really bothered to dive in to the process of making RAM. This is eye-opening and makes me appreciate the sophistication of the tech involved in making those ram-sticks.
This is fascinating, but it really highlights how insane modern hardware manufacturing has become.
Would be interesting to know where the biggest bottleneck shows up first: materials, lithography, or error rates?
Mom: We have RAM at home!
RAM at home:
A was waiting for the part where he slides his little wafer chip into an ASML High NA EUV...
Great video, very much has NileRed vibes.
You're telling me we can free download more RAM now?
Subscribed. Genuinely looking forward to what this gent gets up to.
We need to be able to make semiconductors at home for computing freedom to be preserved the way the world is going.
Oh man my shed could never be this clean
I'm still to own a shed. The best I have is an apartment with no garden.
That's not a lab assistant, looks more like a poodle
"Mom can we get some RAM?" "No dear, we have RAM at home." RAM at home...
I always dreamed about this when i was a child
literally tackling the RAM problem bit by bit
Good video, informative, easy to digest. You should start your own company, perhaps as in the form of an employee-owned entity or using Kickstarter to create an all-in-one RAM Fab kit that can be sold online. I'm sure the maker community may consider buying it.
Spoiler: we never actually get to see the RAM tested
The graphs towards the end were discharge curves for a single transistor/capacitor cell out of only 16 present, if I understood correctly? So "enough cells to count as memory" and "addressing logic" are definitely future work (it looked like he wanted to characterize what the refresh cycle would have to look like before actually building more.) I was kind of surprised that the "use a microscope as a photolithography projector" approach worked at all, it will be interesting to see how that scales up...
2 bytes of memory ought to be enough for anyone!
The Atari 2600 only had 128 bytes of ram. It’s not that far off…
nowadays, we need cheap ram solutions :')
This is cool.
such good idea ( so called home made )
Considering how AI companies and older hardware manufacturing companies are driving up the RAM prices - thus, milking all of us ultimately - I think 3D printing needs to become the new default. And affordable too, eventually (it already is, for cheap things, e. g. plastics or PLA based printing, this is often much cheaper than ordering this from a company, but I mean with regards to computer systems too. Naturally right now we are far away from the nanoscale here, but I see this simply as a situation that will change eventually, given enough time).
Those AI companies and hardware manufacturers lost all right to further dictate and increase prices. Capitalism does not work as de-facto blackmail monopoly - or should not. If a state fails to protect the people, such as in the USA right now under the orange king, then the people need to insist on change. ALL steps against this tyranny from a few superrich needs to end.
Right now the legislation is going in the way how lobbyists want this, e. g. trying to make 3D printing illegal, but I think technology will obsolete such illegal laws eventually. Tyranny will eventually fail.
Photolithography is way more practical, 3d printing principles are already used here in stepping of the mask.
I only buy free-range artisanal DRAM at the DRAM farmer's market.
I only have raw RAM, pastured RAM is wrong.
I get my DRAM needs at the RAM ranch.
>I get my DRAM needs at the RAM ranch.
18GB at a time
You mean at the RAMch? I'll see myself out now
Admit it, deep down, our inner engeering child also wants to build a semiconductor clean room ;)
I do but my days in the fab taught me that you do NOT want people to do this, considering the extremely dangerous chemicals involved. People have died changing EMPTY tanks of phosphine gas used for doping… and HF acid used for etch is another nightmare entirely.
Knowing that really helps you understand just how valuable semiconductors are as a product.
Honest question, is there a way to run the entire process acid-free?
No acids at all? That would be stupendously difficult for no real benefit. So many things are acids, so many useful reactions involve acids, and there's not a significant correlation between "is an acid" and "danger".
yes acids capable of etching = danger :) aktschually
No. Silicon oxide (glass) is extremely tough from a chemical perspective. That's why it's used in chemistry for everything. Barely anything touches it. Also this is the main reason I think that the meme of "silicon based life" is completely absurd and comes from people who only took high school chemistry and built their worldview on that.
Is it conceivable that some organic solvent could be synthesized that is, simultaneously harmless to water-based biological life, and capable of etching Silicon oxides?
Not really. Organics don't really have any affinity for this type of compound. You could, of course, create some kind of organic fluorinating compound, but it would basically just put you back at square one for safety.
It is only half as bad as working in the places that make tbose chemicals for use in clean rooms. Swaping out "empty" phosphine tanks is bad, but filling and shipping hundreds of full tanks is worse.
I used to graduate at an institute having physicists as well as chemists, I gues it was no coincidence that only physicists operated with HF, one chemist told me that no chemist in their right mind would touch it
That's not quite accurate (but close enough). We had HF in the chem lab. It lived in a dedicated metal box with a massive neon warning label and a padlock.
It's notable in comparison that all the deadly organics lived together in an unlocked cupboard (vented OFC). I think the only thing I ever saw treated as more of a pariah than HF were radioactive isotopes. Those generally get an entire dedicated room with restricted entry and a tedious mandatory cleaning procedure.
Makes sense. HF deserves the same awe as radioactive material. I've always found both fascinating. Like some kind of dark magic that curses you if you contact it.
As usual for particularly nasty crap, Derek Lowe has written amusingly about why he won't work with it[1].
[1] https://www.science.org/content/blog-post/things-i-won-t-tou...
HF is routinely used in analytical labs; it's standard to microwave HF solutions for ICP digestions. It's not even the most hazardous reagent in my lab right now.
Now, perhaps this chemist meant that no chemist in their right mind would physically touch HF--in that case, I agree completely!
AFAIK they're talking about anhydrous HF gas, not solutions of it. That sounds substantially less fun to work with. It's used to clean CVD chambers.
Oh fair enough, that is a nightmare.
Oh, my dream clean room is of course fully robot automated and I can watch through a big (safety) window.
This is the issue I have with people saying that solar power is "clean and eco friendly".
It sure is, if you ignore the fact that you have to have a factory to make it where one of the *nicest* things around is the fucking hydrofluoric acid, and most of the rest will kill you instantly in trace amounts.
The technology is, the production is not but you can contain that, at least in theory. Compare that with gasoline that everytime you obtain energy from it you burn it out of existence and create a mess of the environment.
This is why we should have converted all the cars to run on propane, instead of scrapping them in favour of "cleaner greener diesels" 20 years ago when they started all the "scrappage scheme" bollocks.
The propane is going to get burnt anyway. May as well extract some useful work from it, and when you run a car off it they become ultra low emission.
In my journey to make pcb’s at home I decided to stop once I almost gassed myself and shifted instead to buying gpus
I visited a pcb making factory once. Left with an appreciation for the amount of work needed for 80-layer pcbs, and knowing I would not want to deal with making them myself.
my goal was simple 2 sided pcbs, machined traces because i wanted to avoid chemical etching, but when it comes to via's chemicals are really the only way. The chemcials needed for plating via's are very toxic. my current thinking to avoid the really nasty ones is to try conductive ink (probably pretty bad too) but it maybe would work to coat the fr4 material and then allow a copper plating to take... really it's a fun process machining and laser the soldier mask.
I'm all for safer and less error prone processes, if there is a conductive ink safe for home use, just sell me a 3D printer that does it.
I keep thinking that for home tinkering this is really the wrong approach. Surely there are other more DIY-friendly ways to make switches besides with semiconductors? Sure, they wouldn't achieve anywhere near the same density as SOTA semiconductors, but that's not really possible at home anyway.
I guess you could always go back to electromechanical relays. I wonder if it might be possible to come up with a method of 3D printing those?
But in practical terms the Z80 was a 4 um process node so unless you're willing to go back to the proverbial stone age it seems like you need semiconductors and lithography.
While that would be cool, something like the 7400 series, is already pretty close to scratching that itch. And a lot less dangerous.
What a time to be alive.
I put myself in the print-it-yourself-at-home group.
Did anyone forget that core memory is woven ? With knitting, beads and everything ?
https://dl.acm.org/doi/10.1145/3173574.3174105
Did anyone ever use different coloured ferrites to make cool patterns? I'd have thought that'd be a no-brainer, Navajo blanket core memory!
They usually color ferrites to indicated different magnetic properties. I believe they used different color of wires though.
Magnetic core memory final form were single large perforated plates with many conductors plated on the ferrite surface through-holes, and only the vertical stack of bus wires were threaded through the plates. This meant weaving was less of an issue, and higher >1kiB modules were feasible in a smaller area. The main draw back is it sill had destructive read-once access, so always had higher latency in addition to being slow.
The DDR market will adapt, as China grey market state fab smells the opportunity. They have been counterfeiting cmos chips for decades already, and dram is not as complex as people like to assume.
Neuromorphic computing will likely kick over the LLM sand pile at some point, and all that discounted hardware will need re-homed. We can wait for the bubble to run its course, and actual investors realize they were conned. =3
> Neuromorphic computing will likely kick over the LLM sand pile at some point, and all that discounted hardware will need re-homed.
I don't see a lot of work going on in neuromorphic - there was some work at Intel, IIRC. Not saying you're wrong, but just wondering where you think it's going to come from?
> just wondering where you think it's going to come from?
Will likely evolve like any regular biological system, and consume translated LLM weight sets into its initial training condition 3D propagation structure.
The speed at which this occurs will likely initially be measured in weeks due to slower growth state writes, but once bootstrapped the GC is self-propagating.
I normally don't like to speculate, but the barrier to entry would actually be much lower than traditional silicon fabrication processes. It was an old idea from science fiction, that until recently was highly impractical. Asimov was likely wrong about the physical process, but not about how it is made.
https://en.wikipedia.org/wiki/Positronic_brain
Due to theoretical constant morphological changes under GC, one must acknowledge the inherent lack of safety such systems would pose. Have a nice day. =3
I expect you'd carry out most such work in the form of simulations, only moving to hardware once you'd demonstrated an efficient algorithm. If I'm right about that then it would be easy for any corporate R&D on the topic to fly under the radar indefinitely.
On the academic side of things there's a steady drip of papers on things like spiking neutral networks so I'd say the general theme is being explored.
I figure if a breakthrough happens it will be overnight just like what happened with transformers.
The next big shift will be HBF. All that DRAM holding essentially static weights that are read in nice, long linear reads in inference machines is wasted; if you had a proper interface to it you could replace it all with flash for a tenth of the cost.
I buy it from the local Amish.