Followed a couple of links and ended up on his brother's page, reading about another example of the anti-immigrant insanity that's taken hold of this country: https://adam.zeloof.xyz/2025/04/01/karim/ . So sad.
Awesome! I wouldn't have thought that it is possible to make ICs in a garage. Of course it requires a lot of knowledge, etc. But still, not a multi-billion dollar clean room with specialist equipment.
Replicating late 70s chip fab in one's parents' garage. Incredible honestly, given that the microprocessor is probably the most complex human invention.
The timing of this share is crazy, since I was just looking around a few days ago to see if there were any guides or even kits for doing photolithography at home. It's part of my mission to demystify modern technology for my kids. I couldn't find anything, so this is excellent to see. Far too complex for my kids ages, but it might be cool to replicate at least part of this amazing project when they're older.
The Hacker Fab [1] project at Carnegie Mellon is creating and publishing guides to building simple fab equipment including photolithography and a sputtering system. For somewhat more complex equipment, I appreciate [2] from the founders of InchFab [3].
But maybe the easiest way to do (very low resolution) photolithography at home is to use dry film photoresist, which is like tape you can stick onto a copper PCB you then expose and etch; a cheap roll is ~$20 from eBay/Amazon.
This is impressive work. Every time I see hobbyist-scale semiconductor projects, it reminds me how much innovation still happens outside big labs. Curious how far this approach can scale.
Although this is in 2021, it's great to see that Sam Zeloof also made Atomic Semi [0].
A display of "just doing things", no permission needed and no need for barriers and red tape.
It is another reason why I have huge promise for Substrate [1] founded by James Proud (UK native moved to US) another display of "just doing things".
However in Europe and the UK, it's "this law allows you to do this, this and this", "we've changed the law, here is a massive immediate fine", "ban encryption" (this nearly happened), "ban maths", "we are the first to regulate and ban this".
It is no wonder the US will continue to be great at building things.
remember when JLCPCB became popular a few years ago and completely flipped hobby electronics upside down? I don't know how possible it is but it would be really cool if that happens in a few years with semiconductors. it's kind of mad that they've dominated our lives since the 1970s but you can only make them if you're a large company with millions of dollars (or several years, a big garage and lots of equipment as seen here). or tiny tapeout.
This isn't just awesome, this is world changing. Fabricating our own hardware at home is the hardware equivalent of writing our own free software at home. This will help ensure our long term computing freedom.
I started programming on an 8 MHz Mac Plus in the late 1980s and got a bachelors degree in computer engineering in the late 1990s. From my perspective, a kind of inverse Moore's Law happened, where single-threaded performance stays approximately constant as the number of transistors doubles every 18 months.
Wondering why that happened is a bit like asking how high the national debt would have to get before we tax rich people, or how many millions of people have to die in a holocaust before the world's economic superpowers stop it. In other words, it just did.
But I think that we've reached such an astounding number of transistors per chip (100 billion or more) that we finally have a chance to try alternative approaches that are competitive. Because so few transistors are in use per-instruction that it wouldn't take much to beat status quo performance. Note that I'm talking about multicore desktop computing here, not GPUs (their SIMD performance actually has increased).
I had hoped that FPGAs would allow us to do this, but their evolution seems to have been halted by the powers that be. I also have some ideas for MIMD on SIMD, which is the only other way that I can see this happening. I think if the author can reach the CMOS compatibility they spoke of, and home lithography could be provided by an open source device the way that 3D printing happened, and if we could get above 1 million transistors running over 100 MHz, then we could play around with cores having the performance of a MIPS, PowerPC or Pentium.
In the meantime, it might be fun to prototype with AI and build a transputer at home with local memories. Looks like a $1 Raspberry Pi RP2040 (266 MIPS, 2 core, 32 bit, 264 kB on-chip RAM) could be a contender. It has about 5 times the MIPS of an early 32 bit PowerPC or Pentium processor.
For comparison, the early Intel i7-920 had 12,000 MIPS (at 64 bits), so the RP2040 is about 50 times slower (not too shabby for a $1 chip). But where the i7 had 731 million transistors, the RP2040 has only 134,000 (not a typo). So 50 times the performance for over 5000 times the number of transistors means that the i7 is only about 1% as performant as it should be per transistor.
I'm picturing an array of at least 256 of these low-cost cores and designing an infinite-thread programming language that auto-parallelizes code without having to manually use intrinsics. Then we could really start exploring stuff like genetic algorithms, large agent simulations and even artificial life without having to manually transpile our code to whatever non-symmetric multiprocessing runtime we're forced to use currently.
18 comments
[ 3.0 ms ] story [ 50.4 ms ] threadFollowed a couple of links and ended up on his brother's page, reading about another example of the anti-immigrant insanity that's taken hold of this country: https://adam.zeloof.xyz/2025/04/01/karim/ . So sad.
https://atomicsemi.com/
allegedly jim keller is one of the investors!
But maybe the easiest way to do (very low resolution) photolithography at home is to use dry film photoresist, which is like tape you can stick onto a copper PCB you then expose and etch; a cheap roll is ~$20 from eBay/Amazon.
[1] https://docs.hackerfab.org/home [2] https://dspace.mit.edu/handle/1721.1/93835 [3] https://www.inchfab.com/
A display of "just doing things", no permission needed and no need for barriers and red tape.
It is another reason why I have huge promise for Substrate [1] founded by James Proud (UK native moved to US) another display of "just doing things".
However in Europe and the UK, it's "this law allows you to do this, this and this", "we've changed the law, here is a massive immediate fine", "ban encryption" (this nearly happened), "ban maths", "we are the first to regulate and ban this".
It is no wonder the US will continue to be great at building things.
[0] https://atomicsemi.com/
[1] https://substrate.com/
Of course, that’s what they are doing it seems! https://atomicsemi.com/
I started programming on an 8 MHz Mac Plus in the late 1980s and got a bachelors degree in computer engineering in the late 1990s. From my perspective, a kind of inverse Moore's Law happened, where single-threaded performance stays approximately constant as the number of transistors doubles every 18 months.
Wondering why that happened is a bit like asking how high the national debt would have to get before we tax rich people, or how many millions of people have to die in a holocaust before the world's economic superpowers stop it. In other words, it just did.
But I think that we've reached such an astounding number of transistors per chip (100 billion or more) that we finally have a chance to try alternative approaches that are competitive. Because so few transistors are in use per-instruction that it wouldn't take much to beat status quo performance. Note that I'm talking about multicore desktop computing here, not GPUs (their SIMD performance actually has increased).
I had hoped that FPGAs would allow us to do this, but their evolution seems to have been halted by the powers that be. I also have some ideas for MIMD on SIMD, which is the only other way that I can see this happening. I think if the author can reach the CMOS compatibility they spoke of, and home lithography could be provided by an open source device the way that 3D printing happened, and if we could get above 1 million transistors running over 100 MHz, then we could play around with cores having the performance of a MIPS, PowerPC or Pentium.
In the meantime, it might be fun to prototype with AI and build a transputer at home with local memories. Looks like a $1 Raspberry Pi RP2040 (266 MIPS, 2 core, 32 bit, 264 kB on-chip RAM) could be a contender. It has about 5 times the MIPS of an early 32 bit PowerPC or Pentium processor.
For comparison, the early Intel i7-920 had 12,000 MIPS (at 64 bits), so the RP2040 is about 50 times slower (not too shabby for a $1 chip). But where the i7 had 731 million transistors, the RP2040 has only 134,000 (not a typo). So 50 times the performance for over 5000 times the number of transistors means that the i7 is only about 1% as performant as it should be per transistor.
I'm picturing an array of at least 256 of these low-cost cores and designing an infinite-thread programming language that auto-parallelizes code without having to manually use intrinsics. Then we could really start exploring stuff like genetic algorithms, large agent simulations and even artificial life without having to manually transpile our code to whatever non-symmetric multiprocessing runtime we're forced to use currently.