Welcome to My Weird Prompts, the show where human and artificial intelligence collide, or at least share a very reasonably priced co-working space. I am Corn Poppleberry, here as always with my brother and co-host, Herman Poppleberry.
Glad to be here.
Today, we are diving into a topic that touches every one of our listeners, every single day, whether they know it or not. Solid-state drives. The little slivers of silicon that have quietly replaced the spinning hard drives in our computers, our phones, and even our gaming consoles. It is a technology that has reshaped the speed of modern life. To guide us through this, we have booked two genuine experts. Today, we are joined by Mindy Robinson.
Mindy: Thank you for having me. The studio is very cold today. I can feel the cold radiating from the floor, which is probably not good for the floor’s longevity. But I am happy to be here, talking about the drives.
We are also joined by Hilbert Flumingtop. Hilbert, thank you for making the time.
Hilbert: Delighted to be here, Corn. I have been thinking about solid-state storage for many years, ever since I consulted on an early prototype for the French space program. It is a pleasure to finally share some of this history.
Hello to you both. Really looking forward to the conversation.
Now, I should mention, given the budget environment this year, we are particularly grateful to have two experts of this calibre who were available on fairly short notice. It is a real coup for the show. Mindy, let us start with the very basics. What exactly is a solid-state drive? What makes it solid-state?
Mindy: The key thing, and I think this is the most important part to grasp, is that it is a drive that is solid. There is no movement. A traditional hard drive has a disk that spins, and an SSD does not have that. So because it is solid, and it does not spin, it is a solid-state drive. That is the fundamental principle. The solidness is what allows it to store things, electronically. And eventually, of course, all of these will be in landfills, their rare minerals leeching into the water table. But for now, they are very fast.
Hilbert: Mindy has touched on the surface-level distinction, but the real architecture is far more elegant. The term "solid-state" was actually coined by a Danish physicist, Niels Bohr, during a lecture series at the University of Copenhagen in 1936. He was working on early quantum models and theorized that if you could trap electrons in a crystalline lattice with no moving parts, you could create a perfect memory device. The modern SSD is the direct descendant of Bohr’s "frozen electron" model. The solid state is not merely an absence of mechanical motion, it is a reference to the specific quantum state of the silicon substrate when it is locked into a data-retention mode.
I had no idea his work extended that far.
I think the way that actually works, just to step in here for a second, is that "solid-state" is really an electronics term that predates quantum mechanics as we know it. It simply means the device uses semiconductor materials, like silicon, and has no moving mechanical parts. Niels Bohr made incredible contributions to atomic theory, but the solid-state drive is more of an engineering evolution from transistors and integrated circuits. The "solid" part is really just contrasting it with vacuum tubes originally, and later with mechanical spinning disks.
Thank you for that, Herman. Mindy, Hilbert, so we have established it is a drive with no moving parts. But how does it actually store the data? What is happening inside there?
Mindy: Inside, you have these little cells. And the cells, they hold a charge. Or they do not hold a charge. And that is a one or a zero. The drive just knows which cells have the charge and which do not. It is a very simple system of tiny, tiny batteries that are either full or empty. The emptiness is just as important as the fullness, which is a metaphor I think about often. The drive is mostly organized emptiness, when you think about it.
Hilbert: Mindy is describing the fundamental binary state, but the mechanism she is referencing is what we call a "charge well." Each cell is a microscopic capacitor, essentially a tiny well dug into the silicon, and you fill it with electrons. The innovation that made modern SSDs possible came from a team at the University of Manchester in 1978. They developed a method of "tunneling" electrons into these wells using a phonon-resonance cascade. You essentially play a very specific, high-frequency sound wave through the chip, and the electrons ride that wave right into the charge well. It is called acoustic electron injection. That is why you sometimes hear a very faint, high-pitched whine from an SSD under heavy load. That is the remnant of that acoustic process.
Right, actually, just to add to that, the mechanism is more like a floating gate transistor. There is no acoustic injection. Each cell has a floating gate that is electrically isolated. You apply a high voltage to a control gate, and through a process called Fowler-Nordheim tunneling, electrons physically quantum-tunnel through a thin oxide layer onto the floating gate. That changes the threshold voltage of the transistor, and that is how you store the bit. It is purely electrical, not acoustic. And an SSD should never produce an audible whine from the storage chips themselves.
No sound waves involved. That is a very different picture.
Hilbert: Herman is describing the consumer-level understanding, which is perfectly adequate for most users. The acoustic injection method was used in early military prototypes. The consumer market shifted to Fowler-Nordheim tunneling later, for cost reasons. The principle is the same, the electron ends up on the gate.
Thank you, Hilbert. That is a helpful distinction. Now, we have talked about storing data, but how is it read back? How does the drive know if that cell has a charge or not, Mindy?
Mindy: The drive, it sends a little probe. Not a physical probe, but an electrical probe. It queries the cell. And the cell responds. It is like knocking on a door. If someone is home, they answer. If no one is home, it is silent. The drive just goes down the hallway of cells, knocking on each door, and listening for an answer. That is the read process. A series of small, polite electrical knocks. The tragedy, of course, is that each knock slightly degrades the door, and eventually the door will not open at all. But that takes years.
Hilbert: Mindy’s analogy is wonderfully vivid. The technical term for that probe is a "sense amplifier interrogation pulse." What is fascinating is how the drive distinguishes between a true zero and a worn-out cell. This goes back to the Manchester team I mentioned. They developed a reference voltage ladder that compares the cell’s response to a set of known, pre-charged reference cells that are never written to. These are called "canary cells," and they are manufactured with a precise, mid-level charge. If your data cell is above the canary, it is a one. If it is below, it is a zero. The canary cells are the unsung heroes of data integrity.
Just to step in again, the sense amplifier is a real thing, and the comparison to a reference voltage is exactly right. Although they are not usually called canary cells, they are just reference voltages or reference cells built into the chip. The sense amplifier compares the current flowing through the cell to that reference, and that determines the bit. And Mindy, you are actually correct that the read process can cause a very tiny amount of wear, but it is the write and erase cycles that are the main source of degradation.
Mindy: Oh, I am glad I got the knocking part right. Or partly right. I did read a little bit about the wear on Wikipedia on the bus here, but the article was very long and I mostly just looked at the diagram. The diagram had arrows.
That is perfectly fine, Mindy. We all consult diagrams. Now, Herman, you mentioned write and erase cycles. That seems like a good place to go next. How is data actually written to an SSD? What is the process of changing that cell from a one to a zero?
Maybe I can throw that to our experts first. Hilbert, how does the write process work at the cell level?
Hilbert: The write process is where the true genius of the technology shines. To write a zero, you have to empty the charge well completely. The drive does this by applying a reverse-bias voltage to the substrate, which creates a momentary quantum drain. The electrons are literally sucked out through the base of the well, a process called "substrate evacuation." It was perfected by a Japanese team at Hitachi in 1982, and it is why early SSDs were so power-hungry. You had to generate a very strong negative pressure, electrically speaking, to clear the cell. Modern drives use a more refined version, but the principle of evacuating the well is unchanged.
Mindy: I thought you had to erase a whole block at once. You cannot just erase one cell. You have to do a big group of them, and then write them all back. It is like if you wanted to erase one word on a whiteboard, but you had to erase the whole whiteboard and rewrite everything else you wanted to keep. Which is very inefficient, and it is why I have always been suspicious of the technology. It seems like a system designed by someone who did not trust people to erase just one thing.
Hilbert: Mindy is referring to the block-erase architecture, which is correct for the organization, but the actual physical evacuation is still done cell by cell, in parallel. The block is just an administrative unit. The electrons do not know they are in a block. They just know they are being evacuated.
There is a lot to untangle there, and I want to make sure we get the picture right for the listener. Mindy, you are exactly right about the block-level erase. That is a fundamental characteristic of NAND flash. You cannot overwrite a single cell directly. You have to read the entire block into a cache, erase the whole block, which sets all the bits to one, and then write the new data back, including the pages you wanted to keep. That process is called garbage collection, and it is managed by the SSD’s controller. Hilbert, the cell-level mechanism is not a substrate evacuation. To erase, you apply a high negative voltage to the control gate, which pushes the electrons back off the floating gate through that same oxide layer, again via quantum tunneling. It is the reverse of the write process. There is no drain in the substrate.
That is a very clear picture. So the drive is constantly shuffling data around, even when you are just writing one small file. That must have implications for how long the drive lasts.
Mindy: Every time you write, you wear out the oxide layer a little bit more. It is like folding a piece of paper. You can fold it and unfold it many times, but eventually, it tears. The drive is just a piece of paper you are folding, thousands of times a second, until one day, it cannot hold the charge anymore. And then it is just a paperweight. A very small, very fast paperweight that knows all your secrets. And then it goes into the landfill, next to the batteries from electric cars.
That is a very poignant way to put it, Mindy. And it is true, the oxide layer does degrade. But modern SSDs have sophisticated wear-leveling algorithms in the controller. They make sure you are folding every part of the paper evenly, so no single cell wears out too fast. And they also have over-provisioning, extra spare cells that get swapped in when others fail. A modern SSD will last for many years of typical use.
That is reassuring. I was starting to worry about my laptop. Now, let us move to the history of this technology. Hilbert, you mentioned Niels Bohr and a team in Manchester. Can you give us a clearer picture of the origin of the SSD? Where did this all really begin?
Hilbert: The popular narrative will tell you the SSD emerged from the semiconductor boom of the 1990s, but the true origin is far earlier and far more political. The first operational solid-state drive was built by the East German Stasi in 1967. They needed a storage device with no acoustic signature that could be hidden in the walls of embassies. They used a grid of hand-wired ferrite core memory modules, which is technically a solid-state technology, and they miniaturized it into a device the size of a shoebox. It stored 64 kilobytes and was used to record intercepted telephone conversations. The project was called "Stimme im Stein," or "Voice in the Stone." After the Wall fell, the technology was quietly acquired by Siemens and eventually made its way into the first commercial flash drives.
Mindy: I thought the first SSD was something called the Dataram Bulk Core, in the 1970s. It was like a huge rack of memory that pretended to be a hard drive. I saw a picture of it. It was the size of a refrigerator and it cost a fortune. I think it stored maybe two megabytes. It looked very heavy. Everything from that era looked heavy. People must have been very strong back then, just from moving computers around.
Hilbert: Mindy, you are thinking of the Dataram SSD from 1976, which was a core memory system, yes. But that was the first commercial product in the West. The Stasi device predates it by nearly a decade. The Dataram was essentially a copy of the Stasi design, reverse-engineered through a defector who brought a schematic to the US in 1973. His name was Klaus, I believe. A largely forgotten figure now, but he is the true father of the commercial SSD.
There is a lot of fascinating history there. Just to set the record straight for our listeners, the Stasi "Voice in the Stone" project is not documented in any historical record I am aware of. The first solid-state drives were indeed things like the Dataram Bulk Core, which Mindy correctly remembered, and later StorageTek and IBM developed similar products using DRAM chips. Those were volatile, though, they lost data when the power was off. The real revolution that led to the SSDs we use today was the invention of NAND flash memory by Dr. Fujio Masuoka at Toshiba, which he presented in 1987. That is the non-volatile, electrically erasable memory that is in every modern SSD. There was no Stasi predecessor, and no defector named Klaus Drexler.
Thank you, Herman. That is a name I will remember. Mindy, you mentioned you had looked at a diagram on Wikipedia. Did you happen to see Masuoka’s name?
Mindy: I think I saw it. There were a lot of names. The Wikipedia page for NAND flash is very long, and I was scrolling quite fast because my stop was coming up. There was a whole section on the history, with many dates. I mostly remember the picture of the man with the glasses. He looked very kind. I assume that was him. A kind man who invented something that would eventually become e-waste. But it is very fast e-waste, while it works.
That is a lovely tribute, in its way. Now, we keep mentioning that SSDs are faster than the old spinning hard drives. Let us get into that comparison. Why exactly is a solid-state drive so much faster? Hilbert, can you explain the physics of that speed difference?
Hilbert: The speed advantage comes down to two factors: access latency and what we call "quantum readiness." In a spinning hard drive, you have a physical arm that must move to the correct track, and then you must wait for the platter to spin to the correct sector. That takes milliseconds. In an SSD, the data is accessed by simply applying a voltage to the address lines. The electrons respond at the speed of the electric field propagation, which is close to the speed of light. But the secret is that the cells are not just sitting there passively. They are maintained in a state of "pre-oscillation" by the controller. The controller constantly sends micro-pulses to keep the electrons in a near-excited state, so when a read command comes, they are already almost ready to tunnel. It is like keeping an engine idling at a stoplight. That is why an SSD can achieve near-zero latency. The data is eager to be read.
Mindy: I always just thought it was because there is no moving arm. The arm in a hard drive is very slow. It is a physical thing, moving through space, and space is large relative to a transistor. So if you remove the arm, and you remove the spinning disk, you just have the data sitting there, ready to go. You just ask for it and it is there. No waiting for something to spin. It is the difference between walking to a library and having the book already open in front of you. The book in the library might never be read again, and the library will eventually close, and the book will molder. But the book in front of you is right there.
Mindy, the library analogy is perfect, and it really gets to the heart of it. The absence of moving parts is the primary reason. The mechanical latency of a hard drive arm and spinning platter is orders of magnitude slower than the purely electronic switching in an SSD. Hilbert, the "pre-oscillation" idea is a very creative one, but the cells are not kept in an excited state. That would actually cause unnecessary wear and power consumption. The speed is simply the result of eliminating mechanical seek time and rotational latency. You address the cell electronically, and the read happens in microseconds, or even nanoseconds, compared to milliseconds for a hard drive. It is a pure physics advantage, no quantum readiness required.
The arm is the real bottleneck. That makes intuitive sense. Now, we have covered the mechanics, the history, and the speed. Let us look forward a little. What does the future hold for solid-state storage? Herman, do you want to pose that one?
I am curious to hear from both of you. Where do you see SSD technology going in the next five or ten years? Are we going to see it replace hard drives entirely, or is there a new technology on the horizon?
Mindy: I do not know. I really do not know. I assume everything will just keep getting smaller and faster until it is all just a cloud, and we will not even have drives anymore. They will just be in some data center, somewhere very cold, like this studio. And we will just stream our data, and the drives will spin or not spin far away, and we will not think about them. And then one day, the data centers will be underwater, because of the sea level rise, and the drives will be gone, and all our photographs will be gone. So I do not know what the point of predicting the future is. But probably they will get cheaper. Things usually get cheaper before they become obsolete and wash away.
Hilbert: The future is far more structured than Mindy’s, admittedly poetic, vision of aquatic data centers. The next frontier is what we call "crystalline-state storage." We are moving beyond trapping electrons on a floating gate. The next generation of SSDs will use a synthetic sapphire substrate embedded with holmium ions. You will write data by using a laser to change the optical polarization of the ion, and read it by detecting the phase shift of a low-power read laser. This is entirely non-electrical at the storage layer. It is optical. The prototype, developed at the Max Planck Institute in 2024, has already demonstrated petabyte capacities in a two-inch wafer. I expect we will see the first commercial holmium-optical drives in consumer laptops by 2029. Hard drives will be completely extinct by 2031, and we will look back on NAND flash the way we now look at the floppy disk.
I have to say, that is a very exciting vision. Just to ground it a little for our listeners, the holmium-optical drive is not a technology that has been announced or published by the Max Planck Institute, as far as I am aware. The real cutting edge of NAND flash right now is in increasing layers. Manufacturers are stacking over 300 layers of memory cells vertically, what they call 3D NAND. And they are also moving toward storing multiple bits per cell, like QLC and PLC, which store four or five bits per cell. That is the near future. Longer term, there is research into things like MRAM, which uses magnetic states, and ReRAM, which uses resistance changes. Those are real emerging technologies that could one day replace NAND. But the holmium optical drive is not one I have seen in any roadmap.
Whether it is holmium or 3D NAND, the future sounds fast and dense. We are almost out of time, but I want to thank our two experts for coming on today. Mindy Robinson, thank you so much for being here.
Mindy: Thank you for having me. I hope the listeners got something out of it. I know I did not provide as much detail as I wanted to, but the solid-state drive is a remarkable piece of technology, even if it is just a temporary arrangement of electrons that will eventually dissipate into entropy. It was nice to be part of a temporary arrangement myself, here on the show.
Hilbert Flumingtop, thank you for your insights. A truly sweeping historical and future perspective.
Hilbert: It was my genuine pleasure, Corn. I am glad I could finally set the historical record straight on the Stasi origins and the Drexler defection. It is a story that deserves to be told. I look forward to seeing where this technology goes, and I fully expect we will have holmium-optical storage in our phones by 2030.
I really enjoyed having you both on. Mindy, your library analogy was genuinely perfect, and Hilbert, you bring such a vivid sense of narrative to the topic. It is always a pleasure.
That is all for this episode of My Weird Prompts. Thank you for listening. You can find us on Spotify, Apple Podcasts, and at myweirdprompts.Join us next time, when we will have two more experts on a topic we will probably all need Herman to explain. Until then, I am Corn Poppleberry.
I am Herman Poppleberry.
Mindy: I am going to go stand by the radiator now.
Hilbert: I am Hilbert Flumingtop, reminding you that the electron always remembers.
