Electronic computers came on the scene around the 1940s. These early machines were very simple compared to today’s, were built by their end users, and did what they were programmed to do—nothing more, and nothing less.
Today’s computers do what they are wired to do. End users have no say or knowledge about their wiring. As semiconductor manufacturers and governments work in shadows to maintain dominance, the potential for rogues to seize control of our computers is shrouded in trade and state secrets.
Wrongly designed computers aren’t a hypothetical problem. For forty years now, America’s enemies have been hacking into our computers from around the world. History didn’t have to unfold that way, and if we help, the future can offer computers that are radically more secure.
Computers for critical infrastructure and privacy seekers must verifiably do what they are programmed to do—nothing more, and nothing less. When a computer is wired in a way that it might do something it isn’t programmed to do, however tiny the difference, that’s just the opportunity hackers or viruses need to break in and take control. Exploitable defects can lurk in either a computer’s software or its wiring, and when wiring is the culprit, even flawless software can’t protect us.
The tech industry’s focus on money and dominance entice it to manufacture counterfeit simplicity, e.g., machines that offer sleek enclosures instead of straightforward and consistent behavior. To replace this counterfeit simplicity, we need hardware architectures that are simple enough for humans to understand, predict, and control.
In addition to simplicity, computers must also offer transparency of operation, all the way down to the logic gate level, and this transparency must persist for the life of the hardware. To achieve this, not only must all details of a computer’s design be publicly disclosed, but the components and wiring that comprise a machine must be fully decidable, visible, and inspectable by its owner.
Mainstream computers can’t achieve the transparency we need because of the opaque, proprietary semiconductors they contain. We call such components complex VLSI, meaning very large scale integration (VLSI) devices of sufficient complexity to conceal exploitable defects. Examples of complex VLSI include microprocessors, field programmable gate arrays (FPGAs), programmable logic devices (PLDs), and application specific integrated circuits (ASICs).
Complex VLSI is built at nanometer scale in multi-billion dollar semiconductor foundries. Today’s microprocessors may contain tens of billions of transistors. But even vintage microprocessors with only 100,000 transistors can’t function transparently. We need technology that can be assembled at millimeter scale instead of nanometer scale, using maker-scale tools that normal people can afford instead of billion-dollar factories.
Most people who are familiar with electronics probably believe the adoption of microprocessors in the 1980s “enabled” the modern computer revolution. But this is isn’t the whole story, and it turns out that microprocessors are not as essential for traditional computing tasks as many believe. In fact, it could be argued that advances in surface mount packaging and assembly, programming languages, and semiconductor memories are more important enablers than microprocessors for the traditional computers we use today.
It is practical to build computers that function transparently without using complex VLSI. We don’t have to regress to the vacuum tubes of the 1940s, or even the through-hole glue logic of the 1980s. We do need surface mount assembly, dense memory chips, and sensible programming tools. But we can get along just fine without microprocessors and other complex VLSI for many computers we may use for critical infrastructure and protection of basic human rights.
We design and broadcast radically open computer architectures as a global public benefit.
Our computers do not rely for trustworthiness on foreign countries or semiconductor companies, regardless of where located. Instead, it’s the system owners’ own soldering and firmware that determine their computers’ logical connectivity and operation.
Traditional computers are built as inexpensively as practical using a handful of immensely complex chips. This approach is great for optimizing performance, size, cost, and energy use, but its security consequences are catastrophic. America’s adversaries have infested her computers for 40 years continuously, yet few people even realize that chip complexity is a leading enabler.
Dauug computers are built using a larger number of extremely simple chips. They are useful in many applications where performance, size, cost, and energy use are still important, but much less critical than absolute security and accountability.
Most companies that design computers (or chips they want people to build computers from) show scant interest in end users. They only want something that they can sell. And if they sell a product that turns out to be insecure, that’s more lucrative, because they can coerce users into buying newer models.
At The Dauug House, we sell nothing. We design computers that anyone can build, computers that contain no dangerous silicon-borne complex logic such as microprocessors. Our computers are intended to easily last 30 years without needing any “security updates,” or any dependence on us at all. Then we give away all of our designs, to everyone, free of charge.