![Why Millions of Simple Processors Could Beat Complex Chips | Ep. 109 [Steven Swift] Artwork](https://www.buzzsprout.com/rails/active_storage/representations/redirect/eyJfcmFpbHMiOnsiZGF0YSI6MTc1OTgyNTMzLCJwdXIiOiJibG9iX2lkIn19--0c44cc4de5741404ab836d86dc5a4d6dd9ebd35c/eyJfcmFpbHMiOnsiZGF0YSI6eyJmb3JtYXQiOiJqcGciLCJyZXNpemVfdG9fZmlsbCI6WzYwMCw2MDAseyJjcm9wIjoiY2VudHJlIn1dLCJzYXZlciI6eyJxdWFsaXR5Ijo2MH0sImNvbG91cnNwYWNlIjoic3JnYiJ9LCJwdXIiOiJ2YXJpYXRpb24ifX0=--0b154cb6b88107719b1a277923d35676ba5519cc/with%20(5).png)
Industry Ignited Podcast
Industry Ignited is a platform for bold conversations with leaders who are transforming the way business gets done. Each episode spotlights breakthrough stories from the industrial, manufacturing, biotech, chemical, and B2B sectors, giving you an inside look at how top executives, innovators, and changemakers tackle real-world challenges and drive meaningful growth.
Hosted by Dr. Leeanne Aguilarโentrepreneur, executive coach, and marketing strategistโIndustry Ignited goes beyond surface-level discussions to uncover the strategies, mindsets, and lessons that fuel leadership at the highest level. From navigating complex operations and scaling companies to rethinking culture and preparing for the future of work, every conversation is designed to inspire, challenge, and equip you with fresh perspectives.
Whether youโre an executive, entrepreneur, or emerging leader, this podcast will spark ideas, expand your vision, and ignite the drive to lead with confidence in todayโs evolving business landscape.
Industry Ignited Podcast
Why Millions of Simple Processors Could Beat Complex Chips | Ep. 109 [Steven Swift]
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Can millions of simple processors working independently unlock a new approach to computing? Steven Swift joins Dr. Leeanne Aguilar on Industry Ignited to discuss the ideas behind WIS and why he believes simplicity at the processor level can create extraordinary flexibility, reliability, and performance at scale. He breaks down concepts including clockless operation, parallel processing, power efficiency, and the ability to distribute computational tasks across different processing units.
Beyond the technology, Steven discusses one of innovationโs biggest obstacles: the willingness to challenge the way things have always been done. The conversation moves from processor architecture and chip development to engineering education, creative problem-solving, failure, and building from first principles. Itโs an exploration of what can happen when engineers stop simply optimizing existing systems and begin asking whether there is an entirely different way forward.
๐๐ง๐๐ฎ๐ฌ๐ญ๐ซ๐ฒ ๐๐ ๐ง๐ข๐ญ๐๐ https://industryignited.com/
๐๐ง๐ฌ๐ญ๐๐ ๐ซ๐๐ฆ https://www.instagram.com/industryignited/
๐
๐๐๐๐๐จ๐จ๐ค https://www.facebook.com/industryignited/
๐๐ข๐ง๐ค๐๐๐๐ง https://www.linkedin.com/company/industry-ignited
๐๐ฉ๐จ๐ญ๐ข๐๐ฒ https://open.spotify.com/show/3TX1L5LPr4lt1UtcSpSvlN?si=7037433a58c54162
๐๐ฉ๐ฉ๐ฅ๐ ๐๐จ๐๐๐๐ฌ๐ญ https://podcasts.apple.com/us/podcast/industry-ignited-podcast/id1824619671
๐๐ญ๐๐ฏ๐๐ง ๐๐ฐ๐ข๐๐ญ
https://www.linkedin.com/in/steven-swift-20a9b05/
#podcast #podcastinterview #podcasts #podcastclips #podcastlife #industryignited #drleeanneaguilar #stevenswift #innovation #technology #computerarchitecture #processordesign #futureofcomputing #computingtechnology #digitalinnovation #engineeringinnovation #parallelcomputing #techinnovation #powerefficiency #futuretechnology #emergingtechnology #engineering #artificialintelligence #digitaltransformation #ai #digitaltransformation #engineering
What if everything we know about microprocessors, the very foundation of modern technology, is built on a flawed assumption? And what if the next breakthrough isn't more complexity, but radical simplicity? Welcome to Industry Ignited, the podcast where we explore bold ideas, breakthrough innovations, and the leaders redefining what's possible. I'm Dr. Leanne Aguilar, and today I'm joined by Stephen Swift, founder and chief whiz at Swift Enterprises. Steven has spent decades rethinking computing from first principles, challenging the industry's deep-rooted reliance on backward compatibility and complexity. His vision, a fundamentally new kind of microprocessor. Simpler, faster, more scalable, and potentially transformative. Stephen, welcome to the show.
SPEAKER_00Thank you. You got that right. That's what I'm doing. And I'm uh very pleased and honored to be here with you.
Dr. Leeanne AguilarWell, I'm honored to have you and learn more about what you're thinking. And because I know you've been thinking about microprocessors since before the Intel 8080 era. Take us back. What first sparked your curiosity about computing architecture?
SPEAKER_00Thinking about them before I was born. I was probably doing math equations in the womb. I've always been a maths and science nerd. I excelled in the extreme in high school. A perfect score on the math SAT, uh, finished calculus by age 14. So I was a math nerd. In 1970, there were no computers in schools. None whatsoever. There was no programming classes. Despite being advanced in math and science, I had no idea how computers worked. They were a total mystery. And then I went to college, and on my first day at the dorm, a week before classes started, within 10 minutes, kids said, Hey, you want to see something cool? Come with me. And he took me to the to a back room in the computer center. And I saw my first computer. And it was amazing. It had real cores. It was ancient even by that day's standards. It had real cores. You could look at them and see little magnetic loops. And he gave me a 15-minute tutorial, and I was just hooked. I mean, a whole new world opened up to me. I dropped my major and changed it to computer science. That first semester, I took two computer science courses. One was called Fortran, and the other was called, believe it or not, computer architecture. And the textbook was called computer architecture. That was before microprocessors. So architecture then meant really primitive stuff, um, ones and zeros. And I just dug it and I and I I ended up going through their entire curriculum. I had a two-year scholarship, and I finished all up through the graduate courses, and uh the rest is history.
Dr. Leeanne AguilarSo you chose not to follow the mainstream path shaped by Intel. How did that decision shape your thinking over time?
SPEAKER_00I don't really think it was a decision. It just happened that way. I had a series of jobs on various computers and never one that was specifically Intel. I had PCs and I programmed on Intel, but it's not a conscious rejection of Intel. In fact, they're a great company and they make some great processors. I just choose to go a different way.
Dr. Leeanne AguilarAll right. So it wasn't a conscious decision, just you know, the early days of computing in general. And so you were just focused on exploring other technology and other, you know, um, I explored it all.
SPEAKER_00I know the internals on on Intel processors really well. I always, what I watched, they kept making it more and more complicated. And I kept thinking, why couldn't it be more and more simple? Simplicity is my middle name. Uh in my study of math, I go for like the fundamental equations. I don't want to hear the complicated ones.
Dr. Leeanne AguilarUh so that's the direction I went. Yeah, and you mentioned that I know like in our previous conversation, even how they teach math these days, you're saying it's gotten more complicated as opposed to Don't get me started.
SPEAKER_00Just keep in with the sticking with the basics. Exactly. They don't teach the basics. So so you ask a trigonometry graduate, how do you calculate how what is pi? And they don't really know. I challenge them to figure out what it is, and they don't.
Dr. Leeanne AguilarSo you describe the industry though as being stuck in a backward compatibility rut. When did that realization crystallize for you?
SPEAKER_00Yeah, I've I've watched it happen. I've always known it. But when I about 15 years ago, I decided to go off on my own and take the whiz seriously, not as a uh hobby, but as basically my full-time job. I quit all other jobs and been working on my savings. So at that time I realized what I was up against, which is that rut. That rut is a good thing because it created computers all over the world. Microsoft and Intel spread computers to billions of people. Right. But in order to do that, you have to standardize them. And that's the rut. The rut is that every year when Intel puts out a new chip, it has to run 100% of last year's software. Without exception. Even when they put in something stupid, like from time to time there have been instructions added to the Intel chips that nobody wanted and didn't weren't particularly useful. They could not delete them. If even one big customer used it and their software no longer worked, the lawsuits would kill you. So, and not just microprocessors, but I think any product that's mass-produced has to keep its previous customers happy. So you can't change, you can't make big changes. In in Intel microprocessors, basically the the code I wrote in 1974 for the first Intel processor, I'm sure will still run today. The theory of operation, the instruction set architecture has not changed. They've added more instructions and worse yet, more and more complexity. Because they can't delete, they can only add. And that's a rut. As an example, a few years ago, I think it's like 10 now, China decided to design their own chips and not be dependent on Intel. And I thought, wow, they can do anything they want. They could be cool. And what did they do? They did an exact copy of Intel chips. Why? They wanted to compete in international markets. Right. You can't change. And so for me, I am like an alien who landed on this planet and I know computers, but I didn't look at what they were already doing. I just went my own way. And I did it in what Elon Musk calls first principles. And I'll tell you more about that.
Dr. Leeanne AguilarYeah. Right. So the backwards rep because I mean, so many of the upgrading, they're using software and technology that they're depending on. And so if they're not upgrading with everything that they had before, like you said, they might lose some of that um capability. So how do you solve that? How do you innovate and lighten the load without you pretty much can't.
SPEAKER_00You become a pariah and an outcast like me. Because if you if you want to do something completely different, um, unless you're Elon Musk and have a couple hundred million dollars to back yourself, like when he did Rockets.
SPEAKER_02Yeah.
SPEAKER_00You know, that was that was nuts. Uh but he could do that. I can't afford to do that. So um in order to really go a different way, you have to be non-compatible. And that means, I mean, who would buy my chip? It won't run any current programming language, right? It won't run any current operating system. So you have to start over from scratch. And uh that's what I'm up against. That's when it crystallized that this rut is now a problem to me.
Dr. Leeanne AguilarYeah. And what does your technology do that's different then? How how what's the approach that's different than like the Intel approach, let's say?
SPEAKER_00Okay, this is where you're actually asking me to describe what the whiz is.
Dr. Leeanne AguilarWhat is the whiz? That's right.
SPEAKER_00Yes. I've thought about that for a couple of weeks now since I met you. It's really hard to say the answer that in five minutes. It is um uh it is a little tricky to explain something that you have no background in. And I don't mean you, I mean that has no prior knowledge to anybody. But I'm gonna try.
Dr. Leeanne AguilarOkay. In a nutshell, yeah.
SPEAKER_00I'll try. Um, and I'm gonna mention also that there's a the whiz is only half of the project. There's another item called Zaz, Z-O-Z, which somehow stands for WIZ operating system. And it's actually the bigger project, and one cannot exist with the other. The Zaz is an operating system that weaves together every Wiz on the chip or on the board or on the planet. It weaves them together into a single massive supercomputer. And that's way more important, and I'm not going to go there today. So I'm just going to describe the Wiz hardware. And also, I have to make this distinction. The Wiz is very small, it's about 100,000 transistors. That's the same size as the smallest arm right now. It's the same size as the chip, the Motorola 68,000, that ran the first max. So 100,000 transistors is enough to be a decent little microprocessor. It can run languages, it can do operating systems, it run programs. So that's the whiz. It's 100,000 transistors. Today we could put a hundred billion transistors on a chip. Now, if you do that math, that means you could put a million whizes on a single chip. Imagine a million Macs or a million IVM PCs, because that's how big their chip was back then. On the same chip, all running simultaneously. Now that's only theoretical, you can't really do that because you'll overheat and a lot of other things. But I'm going to continue saying it. You can have a million whizzes on a chip. So I call that a whiz chip, and it's different than a whiz. So I need to make that distinction. I'm going to describe now one whiz. It's very simple, but it's designed from the ground up to be surrounded by a sea of more whizzes, which means it it can spread the load, it can let other whizzes do other things. So anyway, here we go. So the whiz is electronically in two parts. I call it the front end and the back end. The front end is, well, I'm going to use a metaphor. Imagine a long straight street which has 50 houses on it, and in front of every house is a mailbox. And the mailman has only one job. The houses are numbered one to fifty, and he's given a quote, instruction, which consists of two numbers. And he goes to the house of the first number and opens the mailbox and pulls out an envelope. He doesn't know what is in it, he just pulls it out. And then he goes to the house of the second number. The mailman gets an instruction like five, ten. So he goes to house five, pulls an envelope out of that mailbox, goes to house ten, sticks it in. That's his entire job, and that's all he does. And that's the end of the description. The front end is a bunch of registers, which are like 32-bit storage, maybe 50 of them, on a bus. So the registers are like the mailboxes, and the bus is like the road. And the whiz executes only one instruction, one kind of instruction, which is a pair of numbers. So if a whiz is given the instruction in binary 5, 20, then it takes register 5, puts it on the bus. And then register 20 is instructed to take its data off the bus. And thereby the data from 5 goes to 20. And that's the only instruction. The only thing the front end can do. And that's how you program it. You give it a series of instructions. You have no idea what they do so far. Now it's time to introduce the back end. So that's the entirety. That's like 10,000 transistors. It's very small, and it'll run between 1 and 200 gigahertz. So that's quote the instruction speed. You're looking at a couple hundred gigahertz, not three gigahertz of modern computers. Of course, that's not doing anything. If we start doing stuff, it slows down, as we'll see. So now the back end. So the back end consists of a bunch of small, simple, full-purpose logic circuits, like an adder or a multiplier, or a square rooter, or a signer, or a cosigner. These are small groups of transistors. They sit in the back end. Each one of them is behind a specific register. So like behind register five could be an adder, and behind register 20 could be a square rooter. And so you copy a number to register 20, stick it in that little mailbox, and then you wait a bit and you pull it out of that mailbox and you find the square root. It's magic to you. From the front end, magic. It just things happen because it doesn't specifically know what the back end is doing. Although the programmer who wrote it better get the right mailbox. And also the back end can contain small single-purpose I.O. circuits, like you could have on the back end a temperature sensor or a motor speed controller. So for instance, if behind register five was a temperature sensor and behind register 80 was a motor speed controller, single instruction 5, 80, gets the temperature, sticks it in the motor speed. And so if it's 20 degrees Celsius, the motor will spin at 20 RPM. And that's how you program it. That's how you do stuff. If behind some register was, say, a multiplier that only multiplies by 3. And if you let's say that's register 15, the number of the temperature sensor, comma the multiplier. And then you take the multiplier and send it to the motor speed. So now at 20 Celsius, it goes 60 RPM. So by routing data one pair at a time from place to place to place, obviously, if you have adder, subtractor, square rooter, you can do any arithmetic function. And so the way you do arithmetic is to route stuff. You route it to the adder, to the multiplier, square rooter, and finally to them to somewhere useful like a motor speed controller or a printer. And so that's the back end. And that completes sort of my description of the WISP. But now I want to do one more thing while we're here, which is I said that I'm going to first principles, finding the axiomatic truths. So there are three truths, and here they are. Truth number one, in my opinion, of course. Truth number one is that any hardware of any complexity can be broken down into small single-purpose blocks of transistors. That's the back end. So any hardware can be broken down into an adder, a subtractor, a divider, a square rooter, etc., including I.O., a temperature sensor. That's axiom one. And that's the back end. That's the fundamental principle of the back end. Think of Legal Bricks. You can build anything out of small blocks. Axiom two is that all computer software can be broken down into small blocks called instructions, which route one piece of hardware to another fundamental basic piece of hardware.
unknownOkay.
SPEAKER_00And that's the front end axiom. And these are, to me, these are basic truths. All hardware can be built out of small blocks, all software can be built out of simple instructions. So the third truth is that all action is within a computer context, is motivated by uh a series of instructions that cause between various blocks. And that's the instruction. So you got back-end blocks, front-end copying between blocks and instructions. And that is the Wiz. Those are as fundamental as it gets. And what I've just said is all it does.
Dr. Leeanne AguilarAnd so we spent decades developing this idea independently. What kept you committed when others didn't see the vision?
SPEAKER_00I can't figure that out. Perhaps I'm nuts and I'm just obsessed, a phantom. No, I am passionate about this. I think it'll work.
Dr. Leeanne AguilarUh-huh.
SPEAKER_00Oh, you know, there's a quote from Elon Musk. He says, design things from first principles. Don't give me old stuff. Don't give me Wi-Fi. Give me new stuff. He says, and if when you design something from first principles, you find it have something different than what is normally done, he says, and I think I quote, then you have an opportunity. So I feel like this really could change the world. Uh-huh. And it's about to try it out. I might be wrong, but I'm going to stick with it. I'm going to either prove it right and we're all going to make trillions of dollars and change the world, or I'm going to die. And that's not really in my plan.
Dr. Leeanne AguilarHow does the whiz fundamentally differ from traditional processor design?
SPEAKER_00Well, in a bunch of ways, but one is this independence of the back-end devices. For instance, I mentioned the Zaz. The Zaz coordinates lots of whizes into a single supercomputer, so to speak. Within a single chip where you have maybe a million whizes, the Zaz coordinates where things get placed. So for instance, if you have a little, say, an algorithm that converts Celsius to Fahrenheit, that can be put in any whiz. And the Zaz puts it in one that's close to whoever's calling it, but it could put it anywhere. So it could put it, for instance, at the other side of the chip, which sends reliability through the roof. When you have a million whizes to choose from, and you can put it in any one, if a dust particle ruins a hundred thousand of the whizes, you still have 900,000 left. And the operating system can place it anywhere, and it can even place it on the next chip on the circuit board. And the programmer never knows. It just things just happen. So it can place, for instance, it can place three copies of an algorithm in three different cities and then compare the results. And you don't even know it did that as a programmer. It just can automatically do stuff like that. So reliability goes through the roof. Flexibility goes through the roof. If your chip starts overheating, it can just move some stuff to another chip. Um, it gives you an amazing flexibility. I I don't have time to give you all the ways, but it frees, it separates the instruction from where it gets executed. And that gives you great power.
Dr. Leeanne AguilarNow, you've eliminated concepts like clocks, instruction sets, and memory management, things most engineers consider essential. What replaces them?
SPEAKER_00Well, there is no spoon. You you don't need them necessarily. So for instance, I'm cheating on the word instruction set. It's not an instruction set architecture. It has only one instruction. It's still an instruction. But it's so different, it's ridiculous. So you don't need an instruction set when you have an instruction that can reference various back-end devices. So it can do multiplies on one register and ads on another. And in an in an ISA, the instruction itself says multiply or add, which makes for amazing complexity. Every instruction. And current Intels have two, three, four hundred instructions now. Every one of them is different circuits. If I have only one circuit, copy from one place to another. And if that place doesn't multiply, it's a multiply. So that's one major difference. Now, clocks is worth mentioning. Every register is 32 bits, let's say. It could be any size. But there's an extra bit. Every back-end device must provide one extra wire. It's called the ready bit. It says the device is ready. It's to use the metaphor of the mailman, it's like a flashing red light on the mailbox. And if the light is flashing, you cannot open it. And so let's say you put in an envelope to ask for a multiply. Multiplies take a relatively long time at nanoscale. And so if you now try to open that box to get the answer, you can't because the light is flashing red. What the whiz does when the ready bit is set is it just stops. It just freezes. In CMOS, when you just stop, you use no power whatsoever. And so basically the mailman lays down in front of the mailbox and goes to sleep until the blinking right light stops. When the light comes back on, the mailman continues. So that's a handshake, not a clock. But it replaces the concept of clocks. In other words, each thing, each computation says, okay, I'm done now. A multiply, for instance, times zero is very fast. So multipliers can be fast or slow, depending on the context. On a regular computer, there's a clock that says every multiply is given so many nanoseconds. It's like you must you start a multiply, you must finish by now, and then you start, yeah, it must finish by now, finish by now, finish by now. That's a clock. And that means the clock rate has to be slower than any device within the clock. Like it has to be calibrated for the slowest multiply. On the whiz, the multiplier just says, okay, I'm done. Next. And the processor keeps going. So every instruction takes as long as it takes, and there's no clock. That has a secondary uh value of even if you're not doing any computation on a regular Intel computer, the clock is ticking. And that turns out to use as much as 50% of the total power of the chip. The clock goes everywhere, and every time in CMOS, every time a bit changes, power is used. And I have an amusing anecdote. Your microwave oven, if you don't use it very often, spends more energy in its clock. The clock on a microwave is maybe five watts. But if you only use your microwave for, like if you have a thousand watt microwave and you use it for one minute, that's a hundred watts or something like that. The clock is using five watts every hour, every day. So it's using 125 watts. So it's ironic that the clock could take more power. On computer chips, the ratio is very much higher. Clocks are anywhere from a third to a half of the power consumption of the processor. Even when it's idle, the clock has to keep going. On the whiz, the mailman just lays down and goes to sleep. He uses no power. And uh so idle things use no power on the whiz. If you have a million processors and 999,000 aren't doing anything, then they aren't using any power. So much more efficient, is what I'm hearing. Relatively efficient. And there's a hundred more things I could I could say, but that's enough for now.
Dr. Leeanne AguilarYeah. So your design emphasizes simplicity at scale. How does that translate to performance? We just talked about power efficiency, but what about reliability?
SPEAKER_00Well, because you have a million whizes, and they're all very simple, but you can put a task on a whiz far away, not even on the same chip, and the operating system will take care of it. Now, if you if you have an algorithm that needs the sign of an angle. It would be absurd to put that algorithm far away. You want it not only on the same chip, but on the very next whiz on that chip. But the Zaz has that flexibility. It could put it anywhere. So, in terms of reliability, as I've said, you could put multiple copies of it and compare. And you don't have to program for that, the Zaz can just do that. It can move when the chip is hot, it can move stuff to another chip. And in terms of performance, every whiz is independent. So for instance, if you've got an algorithm and 20 lines of code down in the algorithm, you're going to need a sine and a cosine and a tangent. On a regular processor, on an inhale processor, it does those one at a time. On the whiz, you can send a data to the sine whiz and another datum to the tangent whiz and another datum to the cosine whiz. And those are three independent whizes. And they're all running at the same time. So you parallel processing goes nuts on the whiz. You can have every whiz running an algorithm. This is what neural nets need. Huge number of whizzes running the same algorithm. On an Intel processor, if it has four multiply units, which typically they do, two or four, on the whole chip, that would have the same power as a whiz, with four of the whizes out of a million having a multiplier. And you can always just send to copy to them and copy from them, no matter where you are. So that so if four whizzes have a multiplier of the same caliber as what Intel has, you have the same power. But on Intel, what they do is as a processor is executing a program, when it has a multiply to do, it then looks forward. It's called instruction look ahead. It looks forward in your program to see if there's another multiply further on down. Maybe 20 lines of code from now, there's another multiply, and it gets it started right now because it's got four multipliers. So that's called look ahead, and it's also called out-of-order instruction. Okay. Where it does this instruction from 20 lines from now, it does it now because it has the multiplier and the data is available. That's called out-of-order instruction. The out-of-order instruction processor on an Intel chip costs them from 50 to 100 million transistors. And man years of development and no end of bugs. The out-of-order execution is dramatically costly. And because on their chip, they can't, they're in that rut. They can't really change the fundamentals. So they're doing everything they can to push the edges. So they're like looking ahead. On the whiz, you just code it that way in the first place. You just if you have five multiplies to do, you send the data to five multiple five other whizes. Oh, also, okay, basic blocks almost forgotten. In traditional computers, if you have a line of code that says, like, if x greater than zero, then do these three lines of code, else do these other three lines of code. That's three whizes. Because you put three, those three lines of code, you put them on another whiz. And the other three lines, you put them on another whiz, and you can do what Intel calls speculative execution, where it tries to execute both branches and then throws out the data. That's another 10 or 20 million transistors. And if you have to throw out the data from one, it's very costly. On the whiz, you just have each of those three instructions give them their own whiz, and you just ignore the one you don't want. It's a little wasteful of power, but it's just changes it.
Dr. Leeanne AguilarSo I'm hearing that the simplicity create or allows more complexity because you can rent a lot more.
SPEAKER_00There's a word in biology called emergent, which means a complex thing emerging from a series of simple things, and in particular, ant colonies. Ants have only like two or three things they can do, but put a billion of them together and they can do amazing stuff. Right. Got it. So yeah, the simplicity gives you the flexibility. Like I said, any hardware can be broken down into small pieces. Right. And the converse of that, a bunch of small pieces can be built back up into the equivalent of any hardware, but now you have the flexibility of building it back up any way you want.
Dr. Leeanne AguilarYeah. And you've also said that the entire circuit can be drawn on a whiteboard in minutes. The front end is true.
SPEAKER_00It's a street, 50 registers, and a couple of transistors to coordinate the copy between two registers.
Dr. Leeanne AguilarWhat is that simplicity unlocked that complexity can't have?
SPEAKER_00Well, what we're talking about, it makes it makes it possible. Well, first of all, because of the simplicity, it's very small. And yet you can still distribute tasks. So you can still have all the complexity. You can have a complex algorithm, but it's made of very simple individual pieces. So you know it gives you tremendous flexibility. You have you get all the complexity as a group of simplicities instead of a single big complexity.
Dr. Leeanne AguilarRight. Yeah. Now you compared your experience to innovators like Charles Babbage and Ada Lovelace's ideas ahead of their time. Do you see yourself in that lineage? Very definitely.
SPEAKER_00Babbage built a computer in 1820. Uh well, he had it running by 1832. He built a mechanical computer and it was brilliant. And the government wouldn't give them any money. They said, What the heck do we need a computer for? It's useless. And the British government uh uh they funded him for a while, but then they didn't because they became conservative and fiscally responsible and they cut him off. Kito walked in and actually got it and saw the value of it. She talked about, you know, you could feed this the music of of Johann Sebastian Bach and it could compute new music. She saw that in 1830. So people don't see the value of it from my point of view. They they say, who, why do you need another computer? Intel has already spent a billion man years developing the best possible computer there is. So how can yours be better? And and why are you bothering? Because Intel will squash you. If I really succeed, Intel will try to buy me out, right? So anyway, I think that answers the question.
Dr. Leeanne AguilarIs there a way to integrate your technology with what Intel has going right now? Not really.
SPEAKER_00You have to start over from scratch. Yeah, you'd have to. If I could get Tim Cook to see the value of the whiz and see, oh, this is faster, better, cheaper than anything you got, he still couldn't use it because you'd have to rewrite hundreds of millions of lines of code in the programs. Yeah. Now, my dream, I have a chapter in my book online where I fell asleep and dreamt I was meeting with Elon Musk. And he wanted to make me the head of the Mars digital system. On Mars, he specifically said, I don't want stuff from the past. I want everything developed from first principles. Right. On Mars, the Wiz could succeed like crazy. And I wrote that chapter almost humorously, and I wrote it for Elon to read. But I'm serious, I'm ready to go to Mars if that's what it takes. Yeah. Because you have to start over, and that's almost impossible here.
Dr. Leeanne AguilarWhen so much is invested in the current technology, like you would have to write everything from scratch, like you were you were saying.
SPEAKER_00Now I could actually create an emulator that acts like an Intel machine and runs on the whiz. I could over my dead body. Because that's just so wrong. And it wouldn't have any of the advantages. So yeah, in theory, the answer to your question is yes, you could integrate it, but no. It would be nuts, and I don't want it to happen.
Dr. Leeanne AguilarSo it'd make more sense to use the technology to create something that hasn't been created before, like the mission to Mars, instead of trying to use it for things that already exist.
SPEAKER_00Right. And I've mentioned to you that IoT, Internet of Things, is a possible niche, but not anymore. That was 10 years ago because they all wrote from scratch. They were developing little programs, they didn't depend on huge libraries of back compatibility. That's changed by now. But there were IoT companies that started from scratch. And also, there's no standard microprocessor for IoT. It needs to have very low power. And nobody's done a good job of that. Uh the Wiz would be perfect. But again, if you've already developed a bunch of software for existing, like the the ARM Cortex-M0 that I mentioned, it's 50,000, 100,000 transistors. That was designed for IoT because it's very low power. The Wiz is way much lower power, but they've already started writing code for it.
unknownYeah.
SPEAKER_00So it might be too late for that option. But it is my, I still consider it's a possibility to find a company that wants to do it. The problem is some company says, okay, I like the whiz, ship me a bunch and we'll use it. Well, now I need a million bucks to create them. So so even if your company likes the whiz and would write its software from scratch on it, there's no whiz. So there's there's this little stumbling block of we got to create one, and that takes a million bucks or more. Now I was in the field, I worked at a small company, I'm sure most people have heard of, called VLSI Technology in 1985. That's why I came to Silicon Valley. I was employee 35, and we ruled chips. The first Mac chip, Steve Jobs, came to our place and we fabb, we had our own fab. And we had our own design team. And so I was there for that, and I know all the tricks. So I know how it really doesn't take a million bucks. You might be able to do it for a quarter of a million, but not really, because if you get it fab, say a TSMC in Taiwan, you need a million bucks just to walk in the door. Salesmen won't even talk to you if if a million bucks isn't hanging out of your pocket. So that's a some that's a big stumbling block. Even if you could find a company that wants to use the whiz, there's no whizes for it to use. Right. That's a what do you call it? Circular Yeah, a problem, right.
Dr. Leeanne AguilarOr like the chicken and the egg kind of example. That's the word I was thinking. Yes, exactly. So is this why you think that highly trained engineers often struggle to embrace radically new paradigms?
SPEAKER_00I know psychologist or expert in you are more than me, I'm sure. But I have a couple of theories, and I think they're all it's all of the above. One is we all are more comfortable doing the things the way we've always been doing them. If uh you buy me a new t-shirt, I'm gonna reject it. Because I didn't buy it myself. So that's one reason. It's just general inertia. But another reason, which is important to me, and Elon Musk has has talked about this a lot, he's talked about you can't hire creative engineers anymore. He's like, they're taught in college to not take chances. Um if a professor gives you an assignment to do something, you do it exactly as he said. You don't come back at him and say, you know, if you change the X to Y and rotated the Z, this would be a way better problem and more interesting. You don't do that. You don't talk back to your professors. Ian is saying he wants engineers to talk back to their management. If a manager says, you know, build me this valve, he needs you to come back to the manager and say, Why is the why is the bolt on the left? Wouldn't it be better on the right? And bigger arguments than that, he points out even if Albert Einstein is your manager, he makes mistakes and you should be questioning it. And then there's the sense of you don't want to fail because you might get fired. Elon says, not only is failure an option, but failure is required. If you don't fail a bunch of times, you're not trying hard enough. Here in Silicon Valley, if you haven't started 10 companies that failed, you ain't an entrepreneur. Uh you're not trying hard enough. You have to not be afraid to fail. And almost everybody is afraid to fail. And Elon has, you know, he tries to promote don't be afraid to fail, try stuff. Yeah. And it's hard to get engineers to do that. They're competent, they do what they're told and they do a great job of it. But they don't fight back.
Dr. Leeanne AguilarTrue. So, yeah, like you're mentioning, imagination is being lost in modern education because people are more concerned about the grade or about passing and succeeding instead of failing. And in math, it's designed that way.
SPEAKER_00In math, especially, which is my field, they're moved along too quickly. They don't get to master any one thing, and they memorize. They just memorize the formula and and then they forget it a week after the exam. This is I actually uh uh tutor math on the side. I have a few students and kind of mentor high school. And they're all in this in this mode of they when you get up to in high school in algebra, there's usually only uh you you do it by the book. But in calculus, there's very often like five ways to solve a problem, and they only solve it in the way they're told. Yeah, don't they don't create, they don't imagine other solutions, and it'll continue in college. They will be just memorized and make good grades.
Dr. Leeanne AguilarSo you lose like the essence of the problem solving then, which is you know what I always appreciated out of college, is that you go to college to learn how to solve problems.
SPEAKER_00Elon has made a point, which I also agree with, which is only go to two years of college because that's where all the fundamentals are taught. You learn everything you need to know in two years, and the next two years are just wasted, go start a company. I kind of like that. There's an example I would give you if I could do a slide. And I'm gonna I'm gonna try to give it simply. But uh if I show an adult a triangle and say, what's the area of that triangle? Uh most adults don't know, or many adults, uh of course I'm surrounded by engineers, they all know. And if you ask a kid who's done algebra or geometry, he's just memorized the formula. It's one half the base times the height, and that's all they know. So if you draw a rectangle around the triangle and then split it down the middle, you will see that the left half of the triangle is exactly equal to the left half of the rectangle. You could like flip the triangle over and you'd get a rectangle. So every triangle, and I show a few examples of this, is half the enclosing rectangle. Now, once I show that to an adult, they are elated and they go, they will never forget that. Triangles are half the enclosing rectangle. Just draw a rectangle around any triangle, and you'll and then split the triangle down the middle, and you will see that the left side of the triangle, if flipped over, would be equal to the left side of the rectangle, and vice versa. Right.
Dr. Leeanne AguilarWhich is half a baset time height, but right, but if you picture it as the formula is forgotten.
SPEAKER_00Whereas once you can see it with your eyes, and you see a rectangle in a triangle, or triangle in a rectangle, you see it, and the adults I talk to, friends I meet, say at a at a at a gathering or whatever, they will remember that for the rest of their lives.
Dr. Leeanne AguilarBecause it becomes visual, right? The picture of the you're seeing it. It's obvious.
SPEAKER_00And then so then I go and and I say, now I'm going to test your imagination. And I draw what's called a kite, which looks like a kite. It's it's got a roof and then a longer, it's basically four triangles. Yeah. And I say, What is the area of this? And they just stare at it and they don't realize it's four triangles. You just draw a rectangle around the kite, it's half the area of that rectangle. Now that takes the the jump from understanding the first to understanding the kite. Right. Requires a tiny bit of imagination.
SPEAKER_02Yeah.
SPEAKER_00And once you show them that they get it, the kids memorize a different formula for every shape. And they never realize, oh my God, a kite is also half its area. You can see it. Just draw a rectangle and you can see it. That's what geometry is supposed to produce. A person who can look around in his environment and see these relationships.
Dr. Leeanne AguilarYeah.
SPEAKER_00And that's exactly the opposite of what current geometry courses create. And I know I tutor these kids, and um, it's it's it's sad. Yeah.
Dr. Leeanne AguilarRight, they lose the the essence of the actual practice and just reduce the case.
SPEAKER_00So geometry is key there. Uh by all kids, all kids hate geometry. And yet, if you taught it right, and there's some books on the subject, famous mathematician, Paul or something, wrote a book on how it should be done. And it's beautiful, and it causes kids to be like big smiles to realize I had a kid who did not understand the formula for for uh diameter. Did not understand, he was in geometry, he was flunking everything. He didn't know what diameter was. So I went around his house with a ruler and we measured the diameter of plates and saucers and stuff. And I also found Home Depot sells pipes of various diameters and little diagrams right up there. And I spent 15 minutes with the kid. He's flunking geometry, he's flunking every subject. He's 12 years old. And he gets it, he sees it, and he's got a big smile on his face. And I ended off right there, it's just 15 minutes. His mom told me she'd never seen him smile about anything related to school. I showed him that one simple thing, but I actually showed it. He saw it. He made it tangible.
Dr. Leeanne AguilarYeah, you made it relatable, tangible, exactly.
SPEAKER_00And and suddenly you get it. But that's where education is right now. It's not educating, nobody likes it, especially in math.
Dr. Leeanne AguilarNow you mentioned Elon Musk as your hero. So if Elon Musk encountered your work, what do you believe he would immediately recognize the others don't?
SPEAKER_00Well, well, like I mentioned, Ada saw that this machine could be fed Johann Sebastian Bach. She imagined the future using this machine. Elon has that kind of imagination. He will not have to be explained to him why a whiz chip of a million whizes would be radiation tolerant, if not radiation hard, which means it could go into outer space. No Intel chip can. Because with the architecture of Intel chips, you one alpha particle and you're dust. And a whiz, one alpha particle, you lose one whiz. The Zaz operating system just moves that code to a different one. He would see stuff like that. My my dream come true is that he will see this. Maybe somebody will refer, maybe somebody in this audience knows him. Uh, if he reads my first page, he will get it. And I think he will decide right then one of two things. Either I'm a complete nutcase and it can't be done, or I'm on to something and the world could change. I need a million bucks, maybe two million. That's like asking you for a nickel. I mean, he could give me a million bucks and not even notice it was gone. So my postulate is he will give me a million bucks and say, here, buy yourself enough rope to hang yourself, or maybe you'll get a result. If you do, I get a piece of the action. That's my dream. And I think he will see the vision of this could actually work. And so it's worth a nickel to uh let the let the kid try it. You know, if he fails, nothing lost.
Dr. Leeanne AguilarAnd if fully realize where would the whiz architecture create the most immediate disruption? Yeah. What industries or applications? I mean, do you see like a I want to put Intel?
SPEAKER_00Uh-huh. I want I want all chips to to be this. It's just a better architecture. And maybe somebody else will replace me with it even better after that. But but it's a better architecture. Anywhere there's a computer or computer chip, I think the whiz is better.
Dr. Leeanne AguilarAnd so do you think the biggest barrier right now is financial, or is it technical or psychological? It's financial.
SPEAKER_00Psychological, hey, I'm ready, I'm there. Technical, it's already done. All the circuit diagrams are out on the website. If you go to my website, which we'll we'll give people in a minute, I said complete schematics, complete explanation. You could steal it. You know, I don't have any patent protection. Just steal it and make it. That'll make me happy. I want to I want to use it. What I want, my business model, I'll delay that answer. So I think it would affect every industry. It would be the new paradigm for chips. Just like what happened in 2007 when Steve Jobs announced the iPhone. Uh-huh. It changed the nature of everything.
Dr. Leeanne AguilarRight, revolutionary. And you also mentioned, like, for the the IoT devices and as well in that technology. It's how could it redefine that space?
SPEAKER_00It's very small, it's very low power. It would be perfect for that. We wouldn't change them much. Which is why I've changed my mind now.
Dr. Leeanne AguilarAnd I don't think that's where I want to start. Yeah. Do you have a specific point of entry? Like where would you want to start?
SPEAKER_00I don't know.
Dr. Leeanne AguilarNow you said you're not looking for investors, but a patron, though. What kind of partner or ecosystem would it take to bring this to life?
SPEAKER_00So my business model is I don't want to spend a million dollars a chip. I don't want to get into the manufacturing business. I don't want to get into dealing with TSMC. I want to license it to somebody. Like you can have it. I'll license it to you. You put in your a million dollars, you build it, you own it. You get all the profit from it. You can use it any way you want, in any product you want. It's yours. I just want a small royalty. I want a penny or two for every whiz you manufacture. In fact, I'm dead serious. I want a trillionth of a penny for every instruction it executes. So my business model is I want to find somebody who'll build it. It's like an architect. I have a plan for a nice house, but I don't want to hire carpenters and plumbers and contractors. So I give it to a contractor or license it. He builds the houses, he builds hundreds of them, and takes back a couple of thousand or ten thousand to me for each one. That's my thought as to the business model. So I'm looking for somebody who wants to build it.
Dr. Leeanne AguilarHopefully, yeah, that dream will be a reality and someone listening will reach out and make those connections. Yeah. Uh yeah, call me. Well, thank you. Yeah. I mean, so how can people get in touch with you?
SPEAKER_00So the website is very easy to find. It's www.stepenswift.org. StephenSwift is one word S-T-E-V-E-N-S-W-I-F-T. You're probably looking at it because it's the name of the podcast, perhaps. Yes. Stevenswift.org. It uh click on whiz. And at that webpage is all my contact information, phone number, zoom room address, the works.
SPEAKER_02Yeah.
SPEAKER_00So that's where the answer to all your questions is. And uh feel free to contact me after you get there and check it out. All right.
Dr. Leeanne AguilarWell, Stephen, thank you for sharing such a bold and thought-provoking vision today. It's definitely an interesting conversation, and you have you brought a lot um of challenging and mind-opening you know thoughts to the forefront. And hopefully, yeah, yeah, people will reach out. And for our listeners, uh, thank you for tuning in. Thank be sure to subscribe and join us for more conversations with leaders who are pushing the boundaries of what's possible. Until next time, stay bold, stay curious, and keep igniting industry.