Bad ideas, like death, are inevitable and just as inescapable.
The US-based tech industry is a Pandora’s box of bad ideas, unleashed upon an unwilling and unwitting populace, and indeed world, with reckless abandon, scorching lives and the Earth itself. Never mind, they say, we’re building the future.
The latest bad idea to spread dark wings and take flight is that building a super massive data center for ‘AI’ called ‘Stargate’- a megamachine that will solve all our problems like a resource and real estate devouring Wizard of Oz – is not only good, but essential.
WASHINGTON (AP) — President Donald Trump on Tuesday talked up a joint venture investing up to $500 billion for infrastructure tied to artificial intelligence by a new partnership formed by OpenAI, Oracle and SoftBank.
The new entity, Stargate, will start building out data centers and the electricity generation needed for the further development of the fast-evolving AI in Texas, according to the White House. The initial investment is expected to be $100 billion and could reach five times that sum.
“It’s big money and high quality people,” said Trump, adding that it’s “a resounding declaration of confidence in America’s potential” under his new administration.
Microsoft and OpenAI are reportedly working on a massive datacenter to house an AI-focused supercomputer featuring millions of GPUs. The Information reports that the project could cost “in excess of $115 billion” and that the supercomputer, currently dubbed “Stargate” inside OpenAI, would be U.S.-based.
The report says that Microsoft would foot the bill for the datacenter, which could be “100 times more costly” than some of the biggest operating centers today. Stargate would be the largest in a string of datacenter projects the two companies hope to build in the next six years, and executives hope to have it running by 2028.
[…]
Bad ideas are inevitable but also, apparently, subject to cost overruns.
There are many ways to think and talk about this project, which is certain to fail (and there is news of far less costly methods, making the Olympian spending even more obviously suspicious). For me, the clearest way to understand the Stargate project and in fact, the entire ‘AI’ land grab, is as an attempt to create guaranteed profit for those tech firms who’re at the commanding heights – Microsoft, OpenAI, Amazon, Oracle and co-conspirators. Capital will flow into these firms whether the system works as advertised or not – i.e. they are paid for both function (such as it is) and malfunction.
This isn’t a new technique. The US defense industry has a long history of stuffing its coffers with cash for delivering weapons systems that work… sometimes. The most infamous example is Lockheed’s F-35 fighter, a project that provides the company with funding for both delivery and correction as described in the US Government Accounting Office article, ‘F-35 Joint Strike Fighter: More Actions Needed to Explain Cost Growth and Support Engine Modernization Decision’ May 2023 –
The Department of Defense’s most expensive weapon system—the F-35 aircraft—is now more than a decade behind schedule and $183 billion over original cost estimates.
[…]
That’s a decade and 183 billion of sweet, steady profit, the sort of profit the tech industry has long sought. First there was ‘enterprise software’, then there was subscription-based cloud, both efforts to create ‘growth’ and dependable cash infusions. Now, with Stargate, the industry may have, at last, found its F-35. Unlike the troubled fighter plane, there won’t be any Tom Cruise films featuring the data center. Then again, perhaps there will be. Netflix, like the rest of the industry, is out of ideas.
In part one of this series, I proposed that Trump’s second term, which, as we’re seeing with the rush of executive orders, has, unlike his first, a coherent agenda (centered on the Heritage Foundation’s Project 2025 plan), would be a time of increased aggression against ostracized individuals and groups, a state of exception in which the pretence of bourgeois democracy melts away.
Because of this, we should change our relationship with the technologies we’re compelled to use; a naive belief in the good will or benign neglect of tech corporations and the state should be abandoned. The correct perspective is to assume breach.
In a April, 2023 published blog post for the network equipment company, F5, systems security expert Ken Arora, described the concept of assume breach:
Plumbers, electricians, and other professionals who operate in the physical world have long internalized the true essence of “assume breach.” Because they are tasked with creating solutions that must be robust in tangible environments, they implicitly accept and incorporate the simple fact that failures occur within the scope of their work. They also understand that failures are not an indictment of their skills, nor a reason to forgo their services. Rather, it is only the most skilled who, understanding that their creations will eventually fail, incorporate learnings from past failures and are able to anticipate likely future failures.
[…]
For the purposes of this essay, the term, failure, is re-interpreted to mean the intrusion of hostile entities into the systems and devices you use. By adopting a technology praxis based on assumed breach, you can plan for intrusion by acknowledging the possibility that your systems have, or will be penetrated.
Primarily, there are five areas of concern:
Phones
Social Media
Personal computers
Workplace platforms, such as Microsoft 365 and Google’s G-Suite
‘Cloud’ platforms, such as Microsoft Azure, Amazon AWS and Google Cloud Platform
It’s reasonable to think that following security best practices for each technology (links in the references section) offers a degree of protection from intrusion. Although this may be true to some extent, when contending with non-state hostiles, such as black hat hackers, state entities have direct access to the ownership of these systems, giving them the ability to circumvent standard security measures via the exercise of political power.
Phones (and tablets)
Phones are surveillance devices. No communications that require security and which, if intercepted, could lead to state harassment or worse should be done via phones. This applies to iPhones, Android phones and even niche devices such as Linux phones. Phones are a threat in two ways:
Location tracking – phones connect to cellular networks and utilize unique identifiers that enable location and geospatial tracking. This data is used to create maps of activity and associations (a technique the IDF has used in its genocidal wars)
Data seizure – phones store data that, if seized by hostiles, can be used against you and your organization. Social media account data, notes, contacts and other information
Phone use must be avoided for secure communications. If you must use a phone for your activist work, consider adopting a secure Linux-based phone such as GrapheneOS which may be more resistant to cracking if seized but not to communication interception. As an alternative, consider using old school methods, such as paper messages conveyed via trusted courier within your group. This sounds extreme and may turn out to be unnecessary depending on how conditions mutate. It is best however, to be prepared should it become necessary.
Social Media
Social media platforms such as Twitter/X, Bluesky, Mastodon, Facebook/Meta and even less public systems such as Discord, which enables the creation of privately managed servers, should not be used for secure communication. Not only because of posts, but because direct messages are vulnerable to surveillance and can be used to obtain pattern and association data. A comparatively secure (though not foolproof) alternative is the use of the Signal messaging platform. (Scratch that: Yasha Levine provides a full explantation of Signal as a government op here).
Personal Computers
Like phones, personal computers -laptops and Desktops – should not be considered secure. There are several sub-categories of vulnerability:
Vulnerabilities caused by security flaws in the operating system (for example, issues with Microsoft Windows or Apple MacOS)
Vulnerabilities designed into the operating systems by the companies developing, deploying and selling them for profit objectives (Windows CoPilot, is a known threat vector, for example)
Vulnerabilities exploited by state actors such as intelligence and law enforcement agencies (deliberate backdoors)
Data exposure if a computer is seized
Operating systems are the main threat vector – that is, opening to your data – when using a computer. In part one of this series, I suggested abandoning the use of Microsoft Windows, Google Chrome OS and Apple’s Mac OS for computer usage that requires security and using secure Debian Linux instead. This is covered in detail in part one.
Workplace Platforms such as Google G-Suite and Microsoft 365 and other ‘cloud’ platforms such Microsoft Azure and Amazon Web Services
Although convenient, and, in the case of Software as a Service offerings such as Google G-Suite and Microsoft 365, less technically demanding to manage than on-premises hosting, ‘cloud’ platforms should not be considered trustworthy for secure data storage or communications.
This is true, even when platform-specific security best practices are followed because such measures will be circumvented by the corporations that own these platforms when it suits their purposes – such as cooperating with state mandates to release customer data.
The challenge for organizations who’re concerned about state sanctioned breach is finding the equipment, technical talent, will and organizational skill (project management) to move away from these ‘cloud’ systems to on-premises platforms. This is not trivial and has so many complexities that it deserves a separate essay, which will be part three of this series.
The primary challenges are:
Inventorying the applications you use
Assessing where the organisation’s data is stored and the types of data
Assessing the types of communications and the levels of vulnerability (for example, how is email used? What about collaboration services such as SharePoint?)
Crafting an achievable strategy for moving applications, services and data off the vulnerable cloud service
Encrypting and deleting data
…
In part three of this series, I will describe moving your organisation’s data and applications off of cloud platforms: what are the challenges? What are the methods? What skills are required? I’ll talk about this and more.
In his 2005 published book, State of Exception, Italian philosopher Giorgio Agamben (who, I feel moved to say, was an idiot on the topic of Covid 19, declaring the virus to be nonexistent) wrote:
“The state of exception is the political point at which the juridical stops, and a sovereign unaccountability begins; it is where the dam of individual liberties breaks and a society is flooded with the sovereign power of the state.”
The (apparently, merely delayed by four years) re-election of Donald Trump is certain to usher in a sustained period of domestic emergency in the United States, a state of exception when even the pretense of bourgeois democracy is dropped and state power is exercised with few restraints.
What does this mean for information technology usage by activist groups or really, anyone?
…
In Feb of 2024, I published the essay, Information Technology for Activists – What is To Be Done? In this essay, I provided an overview of the current information technology landscape, with the needs and requirements of activist groups in mind. When conditions change, our understanding should keep pace. As we enter the state of exception, the information technology practices of groups who can expect harassment, or worse, from the US state should be radically updated for a more aggressively defensive posture.
Abandon Cloud
The computer and software technology industry is the command and control apparatus of corporate and state entities. As such, its products and services should be considered enemy territory. Under the capitalist system, we are compelled to operate on this territory to live. This harsh necessity should not be confused with acceptance and is certainly not a reason to celebrate, like dupes, the system that is killing the world.
The use of operating systems and platforms from the tech industry’s primary powers – Microsoft, Amazon, Google, Meta, X/Twitter, Apple, Oracle – and lesser known entities, creates a threat vector through which identities, data and activities can be tracked and recorded. Moving off these platforms will be very difficult but is essential. What are the alternatives?
There are three main areas of concern:
Services and platforms such as social media, cloud and related services
Personal computers (for example, laptops)
Phones
In this essay, cloud and computer usage are the focus.
By ‘cloud’, I’m referring to the platforms owned by Microsoft (Azure), Amazon (Amazon Web Services or, AWS) and Google (Google Cloud Platform or GCP) and services such as Microsoft 365 and Google’s G Suite. These services are not secure for the purposes of activist groups and individuals who can expect heightened surveillance and harassment from the state. There are technical reasons (Azure, for example, is known for various vulnerabilities) but these are of a distant, secondary concern to the fact that, regardless of each platform’s infrastructural qualities or deficits, the corporations owning them are elements of the state apparatus.
Your data and communications are not secure. If you are using these platforms, your top priority should be abandoning usage and moving your computational resources to what are called on-premises facilities and use the Linux operating system, rather than MacOS or Microsoft Windows.
On Computers
In brief, operating systems are a specialized type of software that makes computers useful. When you open Microsoft Excel on your computer, it’s the Microsoft Windows operating system that enables the Excel program to utilize computer hardware, such as memory and storage. You can learn more about operating systems by reading this Wikipedia article. This relationship – between software and computing machinery – applies to all the systems you use: whether it’s Windows, Mac or others.
Microsoft Windows (particularly the newest versions which include the insecure by design ‘Co-pilot plus PC’ feature) and Apple’s MacOS should be abandoned. Why? The tech industry, as outlined in Yasha Levine’s book, Surveillance Valley, works hand in glove with the surveillance state (and has done so since the industry’s infancy). If you or your organization are using computers for work that challenges the US state – for example, pro-Palestinian activism or indeed, work in support of any marginalized community, there is a possibility vital information will be compromised – either through seizure, or remote access that takes advantage of backdoors and vulnerabilities.
This was always a possibility (and for some, a harsh experience) but as the state’s apparatus is directed towards coordinated, targeted suppression, vague possibility turns into high probability (see, for example, UK police raid home, seize devices of EI’s Asa Winstanley).
The Linux operating system should be used instead, specifically, the Debian distribution, well known for its secure design. Secure by design does not mean invulnerable to attack; best practices such as those described in the article, Securing Debian Manual 3.19, on the Debian website, must be followed to make a machine a harder target.
Switching and Migration
Switching from Microsoft Windows to Debian Linux can be done in stages as described in the document ‘From Windows to Debian’. Replacing MacOS with Debian on Mac Pro computers is described in the document, ‘Macbook Pro’ on the Debian website. More recent Mac hardware (M1 Silicon) is being addressed via Debian’s Project Banana.
On software
If you’re using Microsoft Windows, it’s likely you’re also using the MS Office suite. You may also be using Microsoft’s cloud ‘productivity’ platform, Microsoft 365. Perhaps you’re using Google’s Workspace platform instead or in addition to Microsoft 365. In the section on ‘Services and Platforms’, I discuss the problems of these products from a security perspective. For now, let’s review replacements for commercial ‘productivity’ suites that are used to create documents, spreadsheets and other types of work files.
In the second installment of this essay series I will provide greater detail regarding each of the topics discussed and guidance about the use of phones which are spy devices and social media, which is insecure by design.
I came of age – emerging into young adulthood, liberated, it seemed, from teenaged concerns by entering my 20s – during what we were told was the Cold War’s end. A year before the Soviet Union fell in 1991, President George H.W. Bush, in a speech that was once infamous but is rarely discussed today, delivered to a joint session of the US Congress September 11, 1990 declared that a ‘new world order’ was born (Bush went on to repeat this phrase – a leitmotif of his foreign policy – during a speech at the UN in 1991). For those of us who grew up under the shadow of nuclear annihilation, what macabre war planners called Mutual Assured Destruction or MAD, this provided a form of comfort or at least, the prospect of release from modernity’s prime terror.
Fear inspires a variety of reactions, among them, real or pretended ignorance of danger or the opposite: a desire to know more, to feel, some sense of, not control, always an illusion, but awareness. If I was destined to die, vaporized by a luminous ball of atomic fire, at least there’d be a millisecond of knowing the infernal mechanism’s workings. Growing up in the latter stages of the MAD era, to get this sense of awareness, I studied nuclear weapons and nuclear war doctrines (at least, what was made public). If, on a given Sunday, during lunch, you wanted to know about hydrogen bombs and turned to me for an answer, I could take a sip of vodka and give a solid, well studied non-specialist’s reply.
In the collective imagination, there was a fixation on the scale of devastation. Whether in fictional depictions such as the Terminator films or grimly matter of fact Pentagon strategy documents, the total destruction of major cities – millions dead from blast, heat, radiation and fallout – was a common theme. When anyone said, ‘nuclear war’, it meant the end of the world. What most of us did not know was that just as rust never sleeps, war planners do not cease working to sharpen their blades. New types of nuclear weapons were in the minds of designers, expressed via mathematics and simulated using computation. There is evidence these abstractions have recently taken solid form to be unleashed on Syria’s tortured soil.
On December 23, 2024, Swiss physicist Hans-Benjamin Braun posted the following to his Twitter account:
Nuclear attack in Tartus (Syria):
Radioactive fingerprint of nuke (Tartus) measured in Cyprus within ~16 hours after the attack.
[Note that the dose rate peak cannot be ascribed to precipitation as higher precipitation occurred on Dec 5 with no discernible radiation increase]
The post, first in series that read like urgent dispatches, was based on an analysis of several data points: seismic, radiation and blast effects, used to present a dark conclusion: a new class of nuclear weapon, called Fourth Generation Fusion Nuclear Weapons (FGNW) by US Air Force researcher James Denton in a report titled ‘The Third Nuclear Age: How I learned to Start Worrying About the Clean Bomb’ was deployed in Tartus, Syria.
As the days wore on, more evidence appeared. In a post made on December 26, Dr. Braun posted this update::
Tartus nuke:
DoD data yield for a 99.9% clean weapon (e.g. “Housatonic”) with 0.3kt yield at 110 miles a (max) fallout of 0.035 mR/h. With the obs. time decay at 15 to 20h this yields a dose rate of 9-12nSv/h.
This agrees with observation of 11.55 +/-1.27 nSv/h (>8 sigma signal).
A great deal of technical terrain is covered in this brief post so let’s walk through it.
By “DoD data yield” Dr. Braun is referring to the calculations of nuclear weapon outcomes derived from a US Department of Defense document, ‘The Effects of Nuclear Weapons’ (originally published in 1977). Using these calculations, Braun determined that the Tartus detonation’s characteristics were in line with what was calculated for the last of the 31 test explosions in the 1962, Operation Dominic series, the “Housatonic” detonation of Oct 30, 1962. This explosion was declared 99.9% ‘clean’ , that is, the amount of radiation was significantly less than what is usually produced by nuclear explosions. The reduction of radiation, while retaining other nuclear effects, was the result of the use of a design approach called Ripple.
The impetus for the Ripple program is described in the document, ‘Ripple: An Investigation of the World’s Most Advanced High-Yield Thermonuclear Weapon’. Here is an excerpt:
Operation Redwing and “Clean” Weapons
To help explain the significance of the Ripple concept and the context in which it was devised, we begin with this 1955 letter from then Secretary of Defense Charles E. Wilson:
Until the CASTLE (1954) tests confirmed the feasibility of megaton yields at comparatively small cost, military economy in the atomic weapons field had been largely dominated by blast effects and means of maximizing these (effects) in relation to design and delivery costs. As important as these blast considerations still are, we are now confronted with perhaps even more important considerations in the radioactive by-products field. Stated broadly, the problem appears to be that of maximizing the military effect at the desired time and place, and minimizing such effects where they are not desired. While blast effects are essentially instantaneous and local, the radioactive effects may cover very large areas and may persist for very long periods ranging, in fact, from days in the local fallout effects to many years in atmospheric contamination effects. In other words, radioactive effects force us to bring time in as an additional dimension in dealing with this problem. Moreover, the areas subject to lethal radiation are so large, that in planning the use of these weapons we must carefully weigh the damage to friendly as well as enemy installations.
[…]
“Stated broadly, the problem appears to be that of maximizing the military effect at the desired time and place, and minimizing such effects where they are not desired”.
Unsurprisingly and appropriately, Dr. Braun has faced objections, which, when offered in a spirit of scientific inquiry, he seems to welcome. Social media is an arena, where attempts at conversation or debate are as likely to come to the attention of people who are uninformed, yet confident in their ignorance as a peer who knows what you’re talking about. Among the informed challenges (as opposed to random objections and, potentially, IDF bots) were questions about the radiation levels; shouldn’t they be much higher? Dr. Braun’s answer, based on his understanding of the effects of more advanced designs – the ‘Housatonic’ class – is no, the Tartus detonation represents the first use of a new type of weapon. If he is right, we have entered a more dangerous phase of the nuclear era in which the use of nuclear weapons becomes more attractive because the goal of ‘maximizing the military effects while minimizing undesired effects’ has been achieved.
The March 6, 2022 edition of the BBC’s ‘Point of View’ radio program featured British novelist and essayist, Will Self, reading his work titled ‘Return of the Bomb’. Self used the Russian invasion of Ukraine, then only a month old, and statements President Putin made at the time about Russia’s readiness to use nuclear weapons to discuss what Self called the ‘60th year of the Arkhipov age’. Arkhipov, as in Vasily Arkhipov, the Soviet naval officer who, at a crucial moment in the Cuban Missile Crisis of 1962, prevented the firing of an atomic torpedo on US naval vessels, which surely would have led to a full nuclear war and an end to all things. This decision, Self accurately tells us, earned Arkhipov a special place of honor (a place he has not been given, certainly not in ‘the West’).
Discussing the contradictions of the MAD doctrine we were told maintained a sort of nervous equilibrium, Self stated:
“One of the curious things about the doctrine [of mutually assured destruction] is that it assumes nation states, and even empires, behave as rational, self interested individuals, while the Arkhipov incident tells us that in fact, armageddon is often only averted by actual individuals who will rebel against groupthink. Another paradox of MAD besides its worrying acronym, is that it relies on hostile powers’ motivations and dispensations being transparent to one another. However, what we know from the record, is that both the possibility of nuclear war and its avoidance during the Cuban crisis were a function of ignorance and misreading of intelligence.”
If Dr. Braun is correct and, a precision type of nuclear weapon was used in Tartus, Syria, a productionized refinement of what was deployed in the ‘Housatonic’ test of 64 years ago, we have exited the MAD era (perhaps, as Self notes, we were never in it) and entered an age in which nuclear weapons become a regular part of military action.
Computational Impacts 