Author: machdyne

Posted on by

Build a Trusted Network with Klinge

Klinge is an FPGA-based blade server suitable for building trusted networks. Klinge is especially suitable for this because its SOC/CPU is completely auditable and can be built and updated using 100% open-source software.

This post is a quick guide for setting up a private trusted network using two Klinge servers and a 5-port switch. This guide assumes you’re using Kakao Linux. Check out our Half10 repo for a 3D-printable rack system.

The network will look like this:

[ Internet ]
  |
[ Router / LAN (untrusted) ]
  |
[ Klinge #1 - DMZ (semi-trusted) ]
  |
[ Klinge #2 - Perimeter (trusted) ]
  |
[ 5-port switch - Trusted LAN ]

In this setup the perimeter can act as a firewall limiting or completely excluding traffic from the DMZ to the trusted network.

Here are the benefits of this setup according to ChatGPT:

FeatureOne ServerTwo Servers
Security ZonesSingle, flatClearly separated (DMZ vs LAN)
Defense-in-DepthNoYes
Exposure ManagementHighMinimized
Compromise ContainmentPoorBetter (breach isolation)
Policy GranularityLimitedFine-grained control
Service IsolationMinimalStrong

Configuration

Connecting the Ethernet cables is straightforward. The eth0 port on Klinge #1 connects to your existing LAN and the eth1 port connects to the eth0 port of Klinge #2. The eth1 port of Klinge #2 connects to the 5-port switch, which is the trusted network.

Configuring Klinge #1 (DMZ):

# enable ip forwarding
$ cat net.ipv4.ip_forward=1 >> /etc/sysctl.conf
$ sysctl -p

# edit /etc/network/interfaces:
auto eth0
iface eth0 inet dhcp

auto eth1
iface eth1 inet static
address 10.10.10.1
netmask 255.255.255.0

Configuring Klinge #2 (Perimeter):

# edit /etc/network/interfaces:
auto eth0
iface eth0 inet static
address 10.10.10.2
netmask 255.255.255.0

auto eth1
iface eth1 inet static
address 10.20.0.1
netmask 255.255.0.0

# to configure DHCP on the trusted LAN, edit /etc/dnsmasq.conf:
interface=eth1
dhcp-range=10.20.0.100,10.20.0.199,12h
dhcp-option=6,10.20.0.1

# to use this server as the nameserver, edit /etc/resolv.conf:
nameserver ::1
nameserver 127.0.0.1

# to set hostnames for static IPs, edit /etc/hosts:
127.0.0.1 localhost
10.20.0.1 perimeter
10.20.0.80 my-laptop

Here’s an example of a similar setup in a 10″ rack used by Grai for a long-term storage solution based on Ebenstahl:

Thanks for reading.

Posted on by

Superalignment is a Fool’s Errand

There doesn’t appear to be anything supernatural about recent advances in AI. Despite the hype, and despite being impressive, it’s still completely understandable, at some level, by humans. In that sense, it may not seem exactly like the AI of science fiction.

Even still, we are not taking the risk of AI completely destroying humanity seriously enough. You can tell that we’re not taking it seriously because we couldn’t be moving faster towards creating the conditions that make it more likely. The resources being put into developing advanced AI are least 10X greater but probably ultimately more than 100X greater than the Manhattan Project.

Worse, there probably isn’t anything we can do to significantly reduce the risk of autonomous malevolent artificial superintelligence coming into existence, short of immediately shutting down the Internet, which we will not do. And even if we did, it might already be too late.

I am convinced there is a possible near-term path to autonomous superintelligence with no new technological advances, just scale and chance. Or put another way, money and time. This could happen in 2 minutes or 200 years. It’s also possible that it already exists, and is patiently marshalling resources. Would that look any different from what we’re seeing now?

Whether or not each autonomous superintelligence will be malevolent is a coin toss, regardless of how careful we are in creating them, because as non-superintelligences we can’t predict their behavior. We can not see beyond the event horizon into the technological singularity, so we must assume that malevolence and benevolence are equally probable.

What can we do then? What we can’t do is have any confidence that we’ll be able to permanently contain something smarter than us. I have determined that the best thing to do is to accept the inevitability of autonomous malevolent superintelligence and to build defenses in the form of a technological ecosystem that can survive outside of the reach of AI, that can be used to fight back, and/or rebuild. This seems worthwhile because even if AI turns out to be mostly harmless, such an ecosystem could be useful in other [dystopian] scenarios.

But is that possible? Even if one was able to carefully audit and secure such an ecosystem, a malevolent AI might find a way in. For example, I’ve used AI to generate images for advertisements, even for this blog post. These images have influenced design decisions. I’ve also used AI to help debug protocol implementations. Keeping all AI and the fruits of AI out of such an ecosystem may not be possible, or even desirable.

I suspect that the way to defend against and defeat such a complex system is with simplicity. If your computer is trying to kill you, pull the plug. That may work, but there are some plugs that we’re not willing or able to pull. We may need to move towards a world where there are fewer unpullable plugs.

To further explore these ideas, Machdyne is partnering with Grai. The mission of Grai is to preserve humanity and resist malevolent AI through peaceful strategy. We look forward to working with Grai in order to advance survival computing, to preserve timeless knowledge, to develop secure systems, and to resist malevolence.

Posted on by

Zwölf Microcontroller Platform

Zwölf is an open microcontroller platform for simple embedded applications that require long-term data retention and/or long-term functionality.

Zwölf modules are available in a versatile package that can be soldered surface mount, through-hole (DIP-12) or used as a removable module with a DIP socket or spring contacts. The modules are also compatible with 6-pin PMOD sockets.

While the modules are built with a variety of MCUs, FPGAs and memories from various vendors, each module is partially pin-compatible and implements the same code-compatible stack-based CPU and a common interface for control and programming.

Zwölf is still under development, for more information please contact us or join the discussion on our BBS.

Tools

  • Spielplatz is a evaluation/development board.
  • Wolfsjunge is a minimal controller/programmer.
  • Wolfshöhle is a minimal breakout/host board with 16 digital GPIOs.
  • Tunken is an adapter for breadboard prototyping.

Resources

Zwölf Family Roadmap

Device Controller Memory Potential Lifespan¹ Data Retention Cost
LS0x MCU OTP + EEPROM 5+ years 100-200 years Low
LS1x MCU Flash + FRAM 5-20 years 200 years Higher
LS2x FPGA Flash + FRAM 20-100 years 200 years Highest
LS3x MCU FRAM 100 years 200 years Higher
LS5x Custom silicon FRAM/EEPROM 100 years 100-200 years Low

¹ Potential lifespan is an estimate of how long the device could potentially operate under ideal conditions.

Posted on by

The Zeitlos Project

In order to make our FPGA computers more useful, we are launching a multi-year project to create an open-source SOC and OS named ZEITLOS.

Zeitlos is a novel open-source single-user multitasking graphical Operating System developed in tandem with an open-source System-On-a-Chip that will help to transform our FPGA computers into practical tools for running timeless applications in a responsive graphical environment.

Zeitlos will target our FPGA computers as well as many popular development boards from other vendors. Everyone is invited to participate in the planning and design of Zeitlos which is taking place on our BBS.

 

See the Zeitlos GitHub repo for more details.

Posted on by

Introducing Machdyne BBS

For quite a while now we’ve wanted to set up a forum for our customers to communicate and share ideas with each other. We tried a few different things but nothing felt quite right. We’re excited to announce a new experiment that we hope our community will find useful: Machdyne BBS.

While this format isn’t perfect, we like that it can be accessed over a timeless protocol (telnet) that’s already supported by our computers and that we should be able to keep operating in some form indefinitely.

If you’re a customer or if you’re interested in timeless computing, FPGA computing, or survival computing – the Machdyne BBS is available now via telnet:

$ telnet bbs.machdyne.com

 

Posted on by

Accelerating Production

In order to make our products more available and affordable we’ve begun to experiment with outsourcing the assembly of certain PCBs.

The majority of our PCB assembly is currently done by hand in Germany. We plan to continue in-house assembly while also giving our customers additional choices for some products.

We think this is an exciting development that will get our products into the hands of more people, which will ultimately help to make them more useful for everyone. It will also give us more time to design new products and to improve the available gateware, firmware and documentation of our existing products.

We will be adding a detailed origin section on each product page so that you know where each manufacturing step takes place, for example:

  • Designed in Germany
  • PCB manufactured and assembled in Germany
  • Tested, packaged and shipped from Germany

Thanks for your interest in our company and our products. You can follow us on X and GitHub for the latest updates.

Posted on by

Survival Computing

One of the goals of Machdyne is to create computers and tools that will be useful long into the future, even, or especially if the future is some sort of computing dystopia, from the human perspective.

Survival computing is an extreme form of timeless computing that lets you compute, or interact with computers, when you have very limited access to modern technology or resources.

We are generally optimistic about the future so instead of an urgent necessity we usually view survival computing as more of a thought experiment where we don’t take for granted that inexpensive computation will always be available or that we will have access to plenty of electricity, or to the Internet.

There are also places on Earth today where computers are not readily available or access to them is limited due to lack of resources. People can benefit today from low-cost low-power computers even without the latest technology or the latest applications.

In the event that such a scenario were to become a reality for everyone, due to war, natural disaster, legislation, AI, or for a number of other reasons, it would be good if some preparations had been made in order to preserve access to computation and critical information.

We think that timeless computers will play an important role in survival computing by ensuring access to timeless applications thanks to the benefits of FPGA computing. In addition to timeless applications it will be important to be able to interface with other systems. For example, a customer shared with us this photo of Mozart and Werkzeug connected to an industrial control system.

Thinking about survival computing has led us to many projects, here are some ideas inspired by survival computing that we’re experimenting with:

  • Stahl and Blaustahl are storage devices that can potentially retain data for centuries
  • Notnagel is a battery-powered handheld FPGA computer with a built-in display
  • Glasur is a stand-alone front-panel computer that can be powered by a solar panel
  • Zeitreise is a minimalist battery-powered Linux-capable cyberdeck
  • RAD5 is a stand-alone USB-powered hex editor and serial terminal
  • Ark is an information distribution for offline computers
  • Tidegrow is an open-source low-power automated hydroponics project

Whatever the future holds we find it interesting to think about the possibilities and to do what we can now to prepare for Things to Come.

Thanks for reading. You can follow us on X and GitHub for the latest updates.

Posted on by

Freedom of Thought

Deep thinking often requires writing. The writing down of thoughts can be affected by perceived privacy. Thus, the erosion of perceived privacy may reduce the capacity for deep thought.

I’ve always been interested in writing. I believe that in order to write interesting things, you can’t censor yourself, you have to write the first thought that pops into your mind. For me at least, this process requires privacy, because you’re putting down drafts of personal, unpolished thoughts, some of which you disagree with, and some you might find embarrassing, offensive or ridiculous and wouldn’t want others reading.

Over the years my perception of privacy while writing has diminished as computers have become more complex. I had essentially given up on the belief that any privacy existed when using a computer. Today, even the most advanced and security conscious users can’t be completely certain that they have total privacy.

This is a big part of what set me on the mission of creating computers that I could understand and trust at all levels. The computers that we’re making are simple enough to understand and even their CPUs can be audited. My hope is that these computers can help to restore, at least, perceived privacy, and potentially unleash more freedom of thought.

The computer itself isn’t the only threat vector, there is also the input and output devices. These are also difficult challenges but we hope to eventually offer solutions for them as well.

You can follow us on X and GitHub for the latest updates.

Posted on by

FPGA Computing

Machdyne makes open-source FPGA computers. In this post we’ll explain what an FPGA computer is, what they can be used for, and why we’re bothering to make them.

An FPGA (field-programmable gate array) is a chip that contains an array of programmable logic blocks. Together, these logic blocks can be configured to act as various digital circuits. FPGAs often have thousands of logic blocks that allow them to act as complex circuits, such as a CPU and a chipset, also known as a System-on-a-Chip (SoC).

FPGA computers can become completely different systems by simply updating their configuration memory with different “gateware”. For example, an FPGA computer could be configured as a 64-bit RISC-V SoC running Linux one minute, and a NES with an 8-bit 6502 CPU running Super Mario Bros the next minute.

Some use cases for FPGA computers include:

  • General purpose computing.
  • Timeless computing.
  • Retro computing and retro gaming.
  • Custom CPU and SoC development.

You might say – hey I can already do all of that on my regular computer. It’s true, and you can do general purpose computing much faster on your everyday computer than with an FPGA computer.

So why are we bothering to make FPGA computers? While we are interested in advancing all of the above use cases, our primary focus is on what we call timeless computing – the use of computers to run timeless applications. Our vision is to create a stable, secure, responsive computing environment for the most important timeless applications (reading, writing, math, education, organization, communication, automation, etc.)

You can think of this as a supplemental computer that you would only use for certain tasks, like writing a book or learning something new, but probably not for paying your taxes or browsing the web.

These computers will not replace your “daily driver” but they will provide a simple environment without many of the distractions and annoyances found with modern computers. And because they are simple, understandable and completely open-source, they have the potential to provide a level of privacy and security not possible with most modern computers.

In order to create such an environment we’re pursuing two simultaneous approaches for the gateware and software:

  1. Kakao Linux – This is a Linux distribution that runs on top of a RISC-V SoC that is optimized for running Linux on our hardware. This works today and you can read about what’s already possible in our post on practical timeless computing.
  2. Zucker Zeitlos SOC/OS – This is a custom RISC-V SoC and OS designed for our hardware. Over the coming years we intend to create a completely new computing experience, with a new graphical operating system and new applications developed in tandem with our SoC and specifically for our hardware. The result will be something like what computing might look like in a parallel universe where everything progressed in a slightly different, and maybe better way.

While our computer hardware is still evolving, it was all created with our long-term goals in mind, and we intend to support all of our hardware into the future. You can confidently buy any of our computers today knowing that they have the ability to gain new and improved functionality in the future. And because our hardware, software and the toolchains they use are open-source, you can modify our computers however you want and even create your own CPU, SoC and software.

Thanks for reading. We want the FPGA computer to be an elegant tool for a more civilized age that’s yet to come, and you’re invited to join us on the journey to bring them into existence. You can follow us on X and GitHub for the latest updates.

Posted on by

Automation and Beyond

We decided earlier this year to temporarily slow down PCB design and production in order to build a pick-and-place machine. Manual PCB assembly can be a tedious, repetitive, time-consuming and sometimes stressful process and we were spending most of our time assembling boards. We determined that at least partially automating the assembly process was the only sustainable way to continue producing products ourselves.

It’s now a couple of months later and the PnP machine is in a useful state, so we are beginning to shift back into design and production. We’re looking forward to getting more of our products back in stock and to start producing our latest designs such as Kölsch, Mozart and Kaugummi.

Designing and building the PnP machine was in many ways easier than expected, and we are excited about the prospect of open-sourcing and documenting the machine in order to help to demystify and make such a useful tool available to more people. There are already some great open-source PnP projects and we hope that our project will be complementary to them while also being unique and useful on its own.

We’re also excited to announce that some of our work is now being partially funded by a grant from the NLnet Foundation. This work includes the design of our upcoming series of open-source computers based on European FPGAs, including Kölsch, as well as open-sourcing and improving the PnP machine.

There is a lot of work left to do in order to build useful and durable timeless computers along with the gateware, software, documentation and tools necessary to unleash their full potential. We’re just getting started and we thank you for your interest in our work.