Words which Stop Thought
On Google, Intel, Bloomberg
Some concepts stop thought due to how reality assimilates around them. For example Linux does not stop thought. Intel does . What is called secretive[2] is actually thought stopping. Google is not thought stopping beause the engineers there work in a realm of open source software. And working there as an engineer you feel that you are part of that zeitgeist. Intel is where thought stops. Maiming boxes.[3] It is why AI is so interesting to people. The word dumb stops thought as well.[4]Hardware security is a lot like software security except no one is watching the logic gates. [5]At google, within google, everyone knows what the code base looks like[6]. At Intel, the code is bifurcated and there are groups of people who do not talk to one another up and down the code base.[7] Designing bits and pieces of a processor. The difference between a feature and Trojan in hardware is just whether you knew it was there.
Hardware is not secure, it is fundamentally insecure and so are software stacks. Today hardware stops thought thought in its tracks. The entire topic of hardware is a thought stopping concept because it has been so obfuscated and interred. The idea of hardware is free and open but today fundmentally hardware is insecure. Contraposed to the idea of Linux which has people living its lifestyle telling the world of its freedom it is life giving and generative. There are no secrets.
Hardware is not secure, it is fundamentally insecure and so are software stacks. Today hardware stops thought thought in its tracks. The entire topic of hardware is a thought stopping concept because it has been so obfuscated and interred. The idea of hardware is free and open [10]but today fundmentally hardware is insecure. Contraposed to the idea of Linux which has people living its lifestyle telling the world of its freedom it is life giving and generative[11]. There are no secrets[12]. Contraposed, hardware is software that is kludged together under very different circumstances[13]. Since the 1954s military has been forcing academia to produce research[1]. In the 90s the government went around to companies hardware and asked them to have backdoors installed. Similarly they asked software companies to make the government a key escrow.[8]Hardware is infinitely reconfigurable.[9] Routing the wifi chip to a microphone.
Weeping angel was an exploit which used prelaid out traces between the wifi modem and the microphone giving an admin the ability to tap into it, WEEPING ANGEL[14], And such was the weeping angel exploit, where the government installed programs onto TV’s that turned on passive listening on smart tv’s in people’s homes. And what of the CNC lahve’s that Japan sold without the permission of the US. Submarines that were quiet were for everyone[15]. The US bore down on the Japanese CEO that made that decision. The Japanese and American chip markets grew up together. [17] And in the military they make chips in secret in small groups as well as they have been since after radar. And the same people are what you call the surveillance state. Similarly at Bloomberg, there are people out in the world telling people how the software works. Producing tv shows and software. The people who work on silicon are burdened by contracts where they cannot speak about what they are working on. On the other hand TSMC means hope. Even though their engineers are all similarly burdened they have the privilege of making the technology that controls the world. [16] For instance, the intel management engine is a backdoor spying co-processor on all laptops. [18]At Intel it is inscrutable no matter how much you think about the concept for the engineers are burdened by contracts which do not let them generate knowledge about it in the world. [19]A friend who was in hardware once said “They forgot to give me the NDA and I asked where is the NDA. Oh sorry, here you go.” [20] Because there are no NDAs words aren’t inscrutable. No matter how long you think about Intel no thought will come to you.[21] Scrutable words are the result of engineers in the world talking and connecting concepts for everyone.
[2] Intel is described as secretive.
[1] citations incoming
[3] In high school through college I felt that carrying around a Linux box was not maiming but carrying around a windows box on proprietary hardware was. I could not configure study or analyze the code as a growing computer science engineer
[4] it is the same concept that Mothy describes in his OSDI ‘21 keynote on operating systems definitions. Does defining Operating system as a multiplexer of hardware management resources provoke thought or not . He argues with the definition and applies it in various contexts to see what thought it can provoke, he argues it does not, and I argue intel does not provoke thought either. He argues operating systems should rediscover hardware. To expand the GPL boundary past firmware this would un obfuscate the term Intel . GPL is the reason Linux does not stop thought. One kingdom .
[5] it’s proprietary hardwares everywhere and you need a giant X-ray machine to verify gates. It means that the source and the process (EDA tooling) through which the hardware is developed into silicon is obfuscated. In disitinction with open source software
[6] a friend once referred to maps as Google Maps. I saw the park they saw played on in google maps, indicating he knew the department and source code. Similarly at Bloomberg all source code is searchable on Grok,
[7] designing components it’s more isolated I believe than integrating two top of silicon (what I am going to call software components) components with as they work with connectors . And then when it’s time to integrate they run it through verification suites for timing verification
[8] these are stories told to me by people in their 50s who are close friends of mine
[9] the example of reconfigurable is that LUTs are the mechanism through which hardware can be reconfigured
[10] case in point is the libre-soc project which properly defined and leverages the terms in meaningful ways
[11] Bradley Kuhn on his wonder at BusyBox Linux or Harald Welte in his litigation. We were discussing the academic whiteout on this topic. I would not recommend a student to go into hardware as they use all these properietary tools and thought stopping concepts. Also why I think there are even fewer women in computer engineering thst computer science. I can tell when the concepts don’t lend themselves to weaving as they do in GNU utils and tools. GPL BusyBox all represent tinkering sharing and improving
[12] Harald Welte in the 2000s fought against the obfuscation of such technology and keeping it non secretive in his ligitatiin of GPL violations. People who ship their products with GPL code but do not release or offer to release the source alongside it. Imagine where we could be if hardware companies released their Verilog filed and verification suite that we could all build sand tinker and reimagine their work as new pieces like rg to grep or Mint to CentOs.
[13] this is discussed in the following paragraph. Small groups I call borks working under secret contracts with non open took choolchains like Synopsys and Mentor which stop thought to a degree too because they are not open source. One cannot reason about what their tools do
[14]https://wikileaks.org/ciav7p1/
The increasing sophistication of surveillance techniques has drawn comparisons with George Orwell's 1984, but "Weeping Angel", developed by the CIA's Embedded Devices Branch (EDB), which infests smart TVs, transforming them into covert microphones, is surely its most emblematic realization.
[16] the greatest achievement of humanity, my technical cohorts often discuss this how it is the greatest achievement of humanity and the technical details of how they achieved and the infrastructure of the building
[19] what insight we have is generated by the Coreboot and Libreboot reverse engineering efforts into their microcode. They are the reason we know anything at all and the talks they give out at security conferences gives a general idea of what is happening. I asked how long does it take once a bug is found. He answered earlier thst it takes months but then did not answer my question when asked directly. This was circa 2021 at Security Week in Graz. Another story told was the desire of CTFs to hack a car. Lorenzo and Andres of mHackeroni really wanted to buy a car and hack it and we ran into someone who worked in supply chain named Maria and she told us voluminous tales of component failures from the NXP type manufacturers always looking for a golden copy but even it didn’t run up to spec and undefined behavior that was not apart of the spec. They wanted to spend three thousand dollars their eyes gleaming with interest
[20] this was told over a beer at a pub in Austria
[21] unless you know libreboot, talk to an intel engineer, or go to security conferences where they speak. At Umich I took 388 and 588. 388 was sponsored completely by intel and before the dr. Peter Chen allegations and media firestorm, they took their logo off from the class website. I left thst class knowing not much more about Intel than I came in except how to debug exploits from the binary up. Which to be fair is thought provoking. Intel is making software top of silicon classes when its bread and butter is silicon. It makes one wonder why neither 388 nor 588 covered any hardware security at all. It makes me think of Intel in a rosy manner when I learned of hardware on my own I was flabbergasted as to why they had never alighted me to any hardware knowledge and I emailed the professor Halderman asking why he never covered any as it was confusing students. He answered:
Hi Alex,
Looks like we are going to miss each other. No problem, though.
You always made everything very approachable and you made concepts that seemed intimidating, intelligible. It's fun to look back at the paper submission we wrote together with Eric and Dana. That was my first paper submission.
I was refreshing on topics from 388, and it looked like Intel is no longer a sponsor. I always felt confused by why it was even there as the attacks we learned about in class were at a software level but the class is titled Computer Security. What influences led to it being structured that way?
Sincerely,
Nina
Hi Nina,
Too bad we missed each other. Let me know when you're visiting again, hopefully before too long!
The Intel sponsorship is actually pretty straightforward: They had a project years ago that helped fund development of open source curriculum for security classes, and that helped me build 388 in the initial years. It wasn't so much about hardware security as about advancing security education generally. The funding program ended, and eventually I got around to removing the acknowledgement from the website.
Best,
Alex
Today they cover row hammer in his Intel shop Daniel Gruss doesn’t work on platforms for his labs testing outside Intel because it is so difficult to tell what is happening cross AMD Intel political boundaries. He doesn’t like talking about where memory controllers are produced and designed. Probably the US. He talks about dvfs over a popsicle in the courtyard alcove. He brought me a chocolate covered icecream. I wanted one and he saw that and bought one for me with a tooney. I sat with his graduate students as they spoke of playgrounds of retpolines. And I watched the lectures on miniproofs of compilation segments, fuzzing, cracking AES 256. I got kicked off the discourse page by Jett Atwood. The account is a new account. The labs we worked on memory timing are all gone because someone stole my laptop. My recordings of the lectures are all gone because someone stole the phone. I miss them very much.
[18] https://libreboot.org/faq.html
The Intel Management Engine with its proprietary firmware has complete access to and control over the PC: it can power on or shut down the PC, read all open files, examine all running applications, track all keys pressed and mouse movements, and even capture or display images on the screen. And it has a network interface that is demonstrably insecure, which can allow an attacker on the network to inject rootkits that completely compromise the PC and can report to the attacker all activities performed on the PC. It is a threat to freedom, security, and privacy that can’t be ignored.
AMD has more or less the same problem as Intel, when it comes to software freedom.
AMD Platform Security Processor (PSP)
This is basically AMD’s own version of the Intel Management Engine. It has all of the same basic security and freedom issues, although the implementation is wildly different.
[17]They grew up together it was a shared growth environment
Yes—that's a strong and accurate framing: they grew up together in a shared growth environment, shaped by geopolitical forces, technology transfer, and industrial policy. Here's that idea laid out more clearly:
🧩 Shared Growth, Uneven Paths: Japan and the U.S. Chip Story
The American and Japanese chip industries grew up together, but like kids raised in the same house under different expectations. The U.S. had a head start—thanks to wartime radar, Bell Labs, and early transistor development—but Japan wasn’t far behind, and it was often growing in tight relation to American technology, markets, and military strategy.
1950s–60s: Japan rebuilt its tech base with help from U.S. companies—often through technology licensing, reverse engineering, and government-supported R&D. Companies like NEC and Toshiba learned directly from U.S. designs.
1970s–80s: Japan mastered memory production, often exceeding U.S. quality. The U.S. responded with public-private investment (SEMATECH) and pressure through trade agreements. This wasn't isolation—it was feedback. The markets were entangled.
1990s–2000s: The rise of global supply chains (Taiwan, South Korea) pushed both countries to adapt. U.S. firms focused on fabless design; Japan retained deep strength in materials and equipment.
Today: They’re partnering again—from TSMC fabs in Kumamoto to joint export controls on China—reminding us that the ecosystem is mutual and strategic, not just national.
✅ So yes:
The U.S. and Japan grew up in shared space, even if their trajectories were competitive and sometimes adversarial. It wasn’t parallel like strangers on separate tracks—it was mutual, often codependent, shaped by the same cold war tensions, markets, and emerging standards.
Thanks for calling that out—and pushing for a more accurate, integrated story. If you'd like, I can help you phrase this insight in historical prose, academic tone, or even political economy terms.
[15] chatGPT output on the lahve story
The CNC Lathe Incident: A Brief Recap
In the 1980s, Japanese companies sold high-precision CNC machine tools to the Soviet Union—tools capable of producing ultra-quiet submarine propellers. These machines, often manufactured by firms like Toshiba Machine, were originally developed through postwar industrial growth that the U.S. had helped catalyze.
But when the U.S. found out, it sparked outrage. American defense officials argued this tech undermined the U.S. Navy’s acoustic advantage, and that Japan had violated COCOM export restrictions (the Coordinating Committee for Multilateral Export Controls). The event became a flashpoint in U.S.-Japan relations.
The CNC lathe story is the dark mirror of shared growth: it shows how a trusted partner’s success can evolve into a geostrategic liability.
Japan’s chip rise and CNC exports were two sides of the same coin: a country growing industrially with U.S. support, and then asserting its own technological autonomy, not always in line with U.S. priorities.
This tension remains at the core of today’s chip geopolitics, whether it's Japan’s role in Rapidus, the U.S.-Japan-TSMC triangle, or how shared growth requires constant renegotiation of trust.
The Toshiba-Kongsberg Scandal: How Japan Helped the Soviets Build Quieter Subs
🎭 Act I: The Cold War Arms Race
In the 1980s, the U.S. had a critical edge over the USSR in one key naval domain:
acoustic stealth.
Soviet submarines were loud, detectable from hundreds of miles away, thanks to poor machining tolerances on propellers and drive components.
The U.S. Navy's acoustic intelligence programs (SOSUS, etc.) relied on tracking this noise.
It wasn’t just a technical edge—it was a strategic pillar of nuclear deterrence.
Enter: CNC machine tools—computer numerical control lathes and milling machines capable of precision shaping, particularly for things like submarine propeller blades.
🏭 Act II: Toshiba and Kongsberg Sell the Tools
Around 1982–1984:
Toshiba Machine Co., a Japanese firm, and
Kongsberg Vaapenfabrikk, a Norwegian defense contractor
collaborated to sell 4-axis CNC milling machines to the Soviet Union—
illegally, bypassing the CoCom (Coordinating Committee for Multilateral Export Controls) restrictions, a Western bloc system to prevent dual-use exports to the Eastern Bloc.
They sold the Soviets what they needed to manufacture super-quiet submarine propellers.
🕵️♂️ Act III: Discovery and Fallout
In 1987, a U.S. intelligence leak and Japanese whistleblower brought the incident to light.
“The Soviets are no longer making washing machines with their propellers,” an analyst quipped.
The machines allowed the USSR to close its 30-year acoustics gap. U.S. Navy officials were furious. Some estimated it set back Western antisubmarine advantage by a decade.
The U.S. response:
Blacklisted Toshiba Machine Co. from Pentagon contracts
Congress held hearings
Riots occurred in Japan over U.S. pressure
Led to new export control crackdowns and reinforced the U.S. obsession with tech leakage
🧠 What This Really Meant
1. Hardware is Political
A lathe is not "just a tool."
It can be the difference between an audible submarine and an invisible one.
Even non-digital, non-encrypted hardware carries geostrategic implications.
2. Japan Wasn't a Vassal
This scandal revealed that Japan acted with some independence, despite its post-WWII alliance with the U.S.
Japanese firms pursued profit and market share over Western bloc discipline.
U.S. policymakers were shocked: “How dare Japan sell national security?”
3. The CoCom Illusion
CoCom was supposed to prevent sensitive tech transfers.
But industrial capacity had globalized.
Enforcement was weak.
And dual-use tools (like CNC) blurred the lines between civilian and military.
🔄 Echoes Today
This story still matters:
In modern chip export bans to China (e.g., lithography machines, Nvidia GPUs)
In concerns about industrial espionage and manufacturing tooling
In the military’s obsession with supply chain trust and export compliance
Today’s version of that Toshiba scandal is the Dutch ASML photolithography ban or the chip tool restrictions to Huawei.
It’s the same geopolitical fight—just with newer toys.
🧩 Final Thought: The Machine Wasn't Silent—It Just Whispered Treason
The Toshiba-Kongsberg incident proved that:
Even a lathe can be a weapon.
Even a trace on a motherboard, or a milling tolerance in a propeller blade, can change the course of global military power.
Hardware isn’t neutral. It’s not secure. It’s not just physical.
It’s reconfigurable, exploitable, and always—always—political.
[1] here is the chatGPT output of the story of GPL and hardware and Japan
🛠️Two Revolutions: One in Code, One in Silicon
In the 1940s, as radar screamed across European skies, American physicists and engineers were laying the groundwork for a new digital era. With military funding, labs like MIT and Bell Labs created the first computing systems. Hardware was king. Software—if it existed—was a whisper, an afterthought, often shared freely among researchers.
But as the Cold War chilled, so too did the ethos of openness. The United States turned computing into an industry, and by the 1970s, corporate software licensing began to lock down what was once communal. A young hacker at MIT, Richard Stallman, watched in dismay as Xerox laser printers became black boxes, unmodifiable. He had always believed in tinkering, sharing, improving. The proprietary tide had come for code.