Dave Emory’s entire lifetime of work is available on a flash drive that can be obtained HERE. The new drive is a 32-gigabyte drive that is current as of the programs and articles posted by the fall of 2017. The new drive (available for a tax-deductible contribution of $65.00 or more.)
WFMU-FM is podcasting For The Record–You can subscribe to the podcast HERE.
You can subscribe to e‑mail alerts from Spitfirelist.com HERE.
You can subscribe to RSS feed from Spitfirelist.com HERE.
You can subscribe to the comments made on programs and posts–an excellent source of information in, and of, itself, HERE.
Please consider supporting THE WORK DAVE EMORY DOES.
This broadcast was recorded in one, 60-minute segment.
Introduction: The program begins with recap of the adaptation of IBM’s Hollerith machines to Nazi data compilation. (We concluded FTR #1075 with discussion of this.): ” . . . . Germany’s vast state bureaucracy and its military and rearmament programs, including the country’s growing concentration camp/slave labor system, also required data processing services. By the time the U.S. officially entered the war in 1941, IBM’s German subsidiary had grown to employ 10,000 people and served 300 different German government agencies. The Nazi Party Treasury; the SS; the War Ministry; the Reichsbank; the Reichspost; the Armaments Ministry; the Navy, Army and Air Force; and the Reich Statistical Office — the list of IBM’s clients went on and on.
” ‘Indeed, the Third Reich would open startling statistical venues for Hollerith machines never before instituted — perhaps never before even imagined,’ wrote Edwin Black in IBM and the Holocaust, his pioneering 2001 exposé of the forgotten business ties between IBM and Nazi Germany. ‘In Hitler’s Germany, the statistical and census community, overrun with doctrinaire Nazis, publicly boasted about the new demographic breakthroughs their equipment would achieve.’ . . . .
“Demand for Hollerith tabulators was so robust that IBM was forced to open a new factory in Berlin to crank out all the new machines. At the facility’s christening ceremony, which was attended by a top U.S. IBM executive and the elite of the Nazi Party, the head of IBM’s German subsidiary gave a rousing speech about the important role that Hollerith tabulators played in Hitler’s drive to purify Germany and cleanse it of inferior racial stock. . . .”
In that same article, Yasha Levine notes that the Trump administration’s proposed changes in the 2020 census sound as though they may portend something akin to the Nazi census of 1933: ” . . . . Based on a close reading of internal Department of Commerce documents tied to the census citizen question proposal, it appears the Trump administration wants to use the census to construct a first-of-its-kind citizenship registry for the entire U.S. population — a decision that arguably exceeds the legal authority of the census. ‘It was deep in the documentation that was released,’ Robert Groves, a former Census Bureau director who headed the National Academies committee convened to investigate the 2020 census, told me by telephone. ‘No one picked up on it much. But the term ‘registry’ in our world means not a collection of data for statistical purposes but rather to know the identity of particular people in order to use that knowledge to affect their lives.’ Given the administration’s posture toward immigration, the fact that it wants to build a comprehensive citizenship database is highly concerning. To Groves, it clearly signals ‘a bright line being crossed.’ . . .”
In the conclusion to Surveillance Valley, Yasha Levine notes how IBM computing technology facilitated the Nazi slave labor operations throughout the Third Reich. The epicenter of this was Mauthausen.
The systematic use of slave labor was central to Nazi Germany’s industrial infrastructure: ” . . . . But in the 1930s, Mauthausen had been a vital economic engine of Hitler’s genocidal plan to remake Europe and the Soviet Union into his own backyard utopia. It started out as a granite quarry but quickly grew into the largest slave labor complex in Nazi Germany, with fifty sub-camps that spanned most of modern-day Austria. Here, hundreds of thousands of prisoners–mostly European Jews but also Roma, Spaniards, Russians, Serbs, Slovenes, Germans, Bulgarians, even Cubans–were worked to death. They refined oil, built fighter aircraft, assembled cannons, developed rocket technology, and were leased out to private German businesses. Volkswagen, Siemens, Daimler-Benz, BMW, Bosch–all benefited from the camp’s slave labor pool. Mauthausen, the administrative nerve center, was centrally directed from Berlin using the latest in early computer technology: IBM punch card tabulators. . . .”
Mauthausen’s IBM machines were, in turn, central to German industry’s use of slave labor: ” . . . . the camp had several IBM machines working overtime to handle the big churn of inmates and to make sure there were always enough bodies to perform the necessary work. These machines didn’t operate in isolation but were part of a larger slave labor control-and-accounting system that stretched across Nazi-occupied Europe connecting Berlin to every major concentration and labor punch card, telegraph, telephone, and human courier. This wasn’t the automated type of computer network system that the Pentagon would begin to build in the United States just a decade later, but it was an information network nonetheless: an electromechanical web that fueled and sustained Nazi Germany’s war machine with blazing efficiency. It extended beyond the labor camps and reached into the cities and towns, crunching mountains of genealogical data to track down people with even the barest whiff of Jewish blood or perceived racial impurity in a mad rush to fulfill Adolf Hitler’s drive to purify the German people, but they made the Nazi death machine run faster and more efficiently, scouring the population and tracking down victims in ways that would never have been possible without them. . . .”
In his book–one of the most important in recent memory–Yasha Levine sets forth vital, revelatory information about the development and functioning of the Internet.
Born of the same overlapping DARPA projects that spawned Agent Orange, the Internet was never intended to be something good. Its generative function and purpose is counter-insurgency. ” . . . . In the 1960s, America was a global power overseeing an increasingly volatile world: conflicts and regional insurgencies against US-allied governments from South America to Southeast Asia and the Middle East. These were not traditional wars that involved big armies but guerilla campaigns and local rebellions, frequently fought in regions where Americans had little previous experience. Who were these people? Why were they rebelling? What could be done to stop them? In military circles, it was believed that these questions were of vital importance to America’s pacification efforts, and some argued that the only effective way to answer them was to develop and leverage computer-aided information technology. The Internet came out of this effort: an attempt to build computer systems that could collect and share intelligence, watch the world in real time, and study and analyze people and political movements with the ultimate goal of predicting and preventing social upheaval. . . .”
In this landmark volume, Levine makes numerous points, including:
- The harvesting of data by intelligence services is PRECISELY what the Internet was designed to do in the first place.
- The harvesting of data engaged in by the major tech corporations is an extension of the data gathering/surveillance that was–and is–the raison d’etre for the Internet in the first place.
- The big tech companies all collaborate with the various intelligence agencies they publicly scorn and seek to ostensibly distance themselves from.
- Edward Snowden, the Electronic Frontier Foundation, Jacob Appelbaum, the milieu of the Tor Network and WikiLeaks are complicit in the data harvesting and surveillance.
- Snowden and other privacy activists are double agents, consciously channeling people fearful of having their communications monitored into technologies that will facilitate that surveillance!
The program notes that counterinsurgency–the functional context of the origin of the Internet–is at the foundation of the genesis of Nazism. At the conclusion of World War I, Germany was beset by a series of socialist/Communist uprisings in a number of cities, including Munich. Responding to that, underground Reichswehr units commanded by Ernst Rohm (later head of the SA) systematically assassinated the leaders of the revolution, as well as prominent social democrats and Jews, such as Walther Rathenau. In Munich, an undercover agent for the political department of the Reichswehr under General Von Lossow infiltrated the revolutionaries, pretending to be one of them.
Following the crushing of the rebellion and occupation of the city by Reichswehr units, that infiltrator identified the leaders of the revolution, who were then summarily executed. The infiltrator’s name was Adolf Hitler.
After the suppression of the rebellion, Hitler, Rohm and undercover Reichswehr agents infiltrated a moribund political party and turned it into an intelligence front for the introduction of the supposedly de-mobilized German Army into German society for the purpose of generating political reaction. That front was the German National Social Workers Party.
The broadcast re-capitulates (from part of Miscellaneous Archive Show M11) Hitler’s speech to the Industry Club of Dusseldorf. This speech, which won the German industrial and financial elite over to the cause of the Nazi Party, equated democracy with Communism.
Manifesting a Social Darwinist perspective, Hitler opined that the [assembled] successful, accomplished were, by definition superior to others. If those, by definition, inferior people were allowed to control the political process, they would structure the social and economic landscape to their own benefit.
This, according to Hitler, would be counter-evolutionary.
1a. The program begins with recap of the adaptation of IBM’s Hollerith machines to Nazi data compilation. (We concluded FTR #1075 with discussion of this.)
” . . . . Germany’s vast state bureaucracy and its military and rearmament programs, including the country’s growing concentration camp/slave labor system, also required data processing services. By the time the U.S. officially entered the war in 1941, IBM’s German subsidiary had grown to employ 10,000 people and served 300 different German government agencies. The Nazi Party Treasury; the SS; the War Ministry; the Reichsbank; the Reichspost; the Armaments Ministry; the Navy, Army and Air Force; and the Reich Statistical Office — the list of IBM’s clients went on and on.
” ‘Indeed, the Third Reich would open startling statistical venues for Hollerith machines never before instituted — perhaps never before even imagined,’ wrote Edwin Black in IBM and the Holocaust, his pioneering 2001 exposé of the forgotten business ties between IBM and Nazi Germany. ‘In Hitler’s Germany, the statistical and census community, overrun with doctrinaire Nazis, publicly boasted about the new demographic breakthroughs their equipment would achieve.’ . . . .
“Demand for Hollerith tabulators was so robust that IBM was forced to open a new factory in Berlin to crank out all the new machines. At the facility’s christening ceremony, which was attended by a top U.S. IBM executive and the elite of the Nazi Party, the head of IBM’s German subsidiary gave a rousing speech about the important role that Hollerith tabulators played in Hitler’s drive to purify Germany and cleanse it of inferior racial stock. . . .”
CORRECTION: The date of the Yasha Levine article is ” . . . . 2019,” NOT 2018, as read in the program.
“The Racist Origins of America’s Tech Industry” by Yasha Levine; Zero One; 4/28/2019.
. . . . . Nazis and numbers
Hollerith tabulators were a big hit all over the world. But one country was particularly enamored with them: Nazi Germany.
Adolf Hitler came to power on the back of the economic devastation that followed Germany’s defeat in World War I. To Hitler, however, the problem plaguing Germany was not economic or political. It was racial. As he put it in Mein Kampf: “The state is a racial organism and not an economic organization.” The reason Germany had fallen so far, he argued, was its failure to tend to its racial purity. There were only about a half-million Jews in Germany in 1933 — less than 1% of the population — but he singled them out as the root cause of all of the nation’s problems.
Hitler and the Nazis drew much of their inspiration from the U.S. eugenics movement and the system of institutional racism that had arisen in slavery’s wake. Their solution was to isolate the so-called mongrels, then continuously monitor the racial purity of the German people to keep the volk free of further contamination.
The only problem: How to tell who is really pure and who is not?
The U.S. had a ready solution. IBM’s German subsidiary landed its first major contract the same year Hitler became chancellor. The 1933 Nazi census was pushed through by Hitler as an emergency genetic stock-taking of the German people. Along with numerous other data points, the census focused on collecting fertility data for German women — particularly women of good Aryan stock. Also included in the census was a special count of religiously observant Jews, or Glaubensjuden.
Nazi officials wanted the entire count, estimated to be about 65 million people, to be done in just four months. It was a monumental task, and German IBM officials worked around the clock to finish it. So important was the success of the contract to IBM that CEO Thomas J. Watson personally toured the giant Berlin warehouse where hundreds of female clerks worked in rotating seven-hour shifts 24 hours a day.
Watson came away greatly impressed with the work of his German managers. They had pulled off a seemingly impossible assignment, one that was complicated by a custom-enlarged punch card format necessary for “political considerations” — IBM’s coded explanation for the extra data demands the Nazi regime required.
As Hitler’s Nazi Party tightened its grip on Germany, it launched all sorts of additional data-gathering programs to purify the German nation. And IBM helped them do it.
“[T]he precondition for every deportation was accurate knowledge of how many Jews in a particular district fitted the racial and demographic descriptions in Berlin’s quotas,” write David Martin Luebke and Sybil Milton in “Locating the Victim,” a study into Nazi use of the tabulator machines. “Armed with these data,” they said, “the Gestapo often proved able to anticipate with remarkable accuracy the total number of deportees for each racial, status, and age category.”
Germany’s vast state bureaucracy and its military and rearmament programs, including the country’s growing concentration camp/slave labor system, also required data processing services. By the time the U.S. officially entered the war in 1941, IBM’s German subsidiary had grown to employ 10,000 people and served 300 different German government agencies. The Nazi Party Treasury; the SS; the War Ministry; the Reichsbank; the Reichspost; the Armaments Ministry; the Navy, Army and Air Force; and the Reich Statistical Office — the list of IBM’s clients went on and on.
This history reveals an uncomfortable and fundamental truth about computer technology.
“Indeed, the Third Reich would open startling statistical venues for Hollerith machines never before instituted — perhaps never before even imagined,” wrote Edwin Black in IBM and the Holocaust, his pioneering 2001 exposé of the forgotten business ties between IBM and Nazi Germany. “In Hitler’s Germany, the statistical and census community, overrun with doctrinaire Nazis, publicly boasted about the new demographic breakthroughs their equipment would achieve.” (IBM has criticized Black’s reporting methods, and has said that its German subsidiary largely came under Nazi control before and during the war.)
Demand for Hollerith tabulators was so robust that IBM was forced to open a new factory in Berlin to crank out all the new machines. At the facility’s christening ceremony, which was attended by a top U.S. IBM executive and the elite of the Nazi Party, the head of IBM’s German subsidiary gave a rousing speech about the important role that Hollerith tabulators played in Hitler’s drive to purify Germany and cleanse it of inferior racial stock.
“We are very much like the physician, in that we dissect, cell by cell, the German cultural body,” he said. “We report every individual characteristic…on a little card. These are not dead cards, quite to the contrary, they prove later on that they come to life when the cards are sorted at a rate of 25,000 per hour according to certain characteristics. These characteristics are grouped like the organs of our cultural body, and they will be calculated and determined with the help of our tabulating machine.”
On the surface, it may seem like the story of Herman Hollerith and the U.S. census are historical relics, an echo from a bygone era. But this history reveals an uncomfortable and fundamental truth about computer technology. We can thank nativism and the census for helping to bring the computer age into existence. And as the battle over the 2020 census makes clear, the drive to tally up our neighbors, to sort them into categories and turn them into statistics, still carries the seed of our own dehumanization.
1b. The Trump administration’s framing of questions for the 2020 census appear aimed at creating a “national registry”–a concept reminiscent of the Third Reich’s use of IBM’s Hollerith-collected data:
“The Racist Origins of America’s Tech Industry” by Yasha Levine; Zero One; 4/28/2019.
” . . . . Based on a close reading of internal Department of Commerce documents tied to the census citizen question proposal, it appears the Trump administration wants to use the census to construct a first-of-its-kind citizenship registry for the entire U.S. population — a decision that arguably exceeds the legal authority of the census. ‘It was deep in the documentation that was released,’ Robert Groves, a former Census Bureau director who headed the National Academies committee convened to investigate the 2020 census, told me by telephone. ‘No one picked up on it much. But the term ‘registry’ in our world means not a collection of data for statistical purposes but rather to know the identity of particular people in order to use that knowledge to affect their lives.’ Given the administration’s posture toward immigration, the fact that it wants to build a comprehensive citizenship database is highly concerning. To Groves, it clearly signals ‘a bright line being crossed.’ . . .”
2. In the conclusion to Surveillance Valley, Yasha Levine notes how IBM computing technology facilitated the Nazi slave labor operations throughout the Third Reich. The epicenter of this was Mauthausen.
The systematic use of slave labor was central to Nazi Germany’s industrial infrastructure: ” . . . . But in the 1930s, Mauthausen had been a vital economic engine of Hitler’s genocidal plan to remake Europe and the Soviet Union into his own backyard utopia. It started out as a granite quarry but quickly grew into the largest slave labor complex in Nazi Germany, with fifty sub-camps that spanned most of modern-day Austria. Here, hundreds of thousands of prisoners–mostly European Jews but also Roma, Spaniards, Russians, Serbs, Slovenes, Germans, Bulgarians, even Cubans–were worked to death. They refined oil, built fighter aircraft, assembled cannons, developed rocket technology, and were leased out to private German businesses. Volkswagen, Siemens, Daimler-Benz, BMW, Bosch–all benefited from the camp’s slave labor pool. Mauthausen, the administrative nerve center, was centrally directed from Berlin using the latest in early computer technology: IBM punch card tabulators. . . .”
Mauthausen’s IBM machines were, in turn, central to German industry’s use of slave labor: ” . . . . the camp had several IBM machines working overtime to handle the big churn of inmates and to make sure there were always enough bodies to perform the necessary work. These machines didn’t operate in isolation but were part of a larger slave labor control-and-accounting system that stretched across Nazi-occupied Europe connecting Berlin to every major concentration and labor punch card, telegraph, telephone, and human courier. This wasn’t the automated type of computer network system that the Pentagon would begin to build in the United States just a decade later, but it was an information network nonetheless: an electromechanical web that fueled and sustained Nazi Germany’s war machine with blazing efficiency. It extended beyond the labor camps and reached into the cities and towns, crunching mountains of genealogical data to track down people with even the barest whiff of Jewish blood or perceived racial impurity in a mad rush to fulfill Adolf Hitler’s drive to purify the German people, but they made the Nazi death machine run faster and more efficiently, scouring the population and tracking down victims in ways that would never have been possible without them. . . .”
It is a crisp and sunny morning in late December 2015 when I take a right turn off a small country highway and drive into Mauthausen, a tiny medieval town in northern Austria about thirty-five miles from the border with the Czech Republic. I pass through a cluster of low-slung apartment buildings and continue on, driving through spotless green pastures and pretty little farmsteads.
I park on a hill overlooking the town. Below is the wide Danube River. Clusters of rural homes poke out from the cusp of two soft green hills, smoke lazily wafting out of their chimneys. A small group of cows is out to pasture, and I can hear the periodic braying of a flock of sheep. Out in the distance, the hills recede in layers of hazy green upon green, like the scales of a giant sleeping dragon. The whole scene is framed by the jagged white peaks of the Austrian Alps.
Mauthausen is an idyllic place. Calm, almost magical. Yet I drove here not to enjoy the view but to get close to something I came to fully understand only while writing this book.
Today, computer technology frequently operates unseen, hidden in gadgets, wires, chips, wireless signals, operating systems, and software. We are surrounded by computers and networks, yet we barely notice them. If we think about them at all, we tend to associate them with progress. We rarely stop to think about the dark side of information technology–all the ways it can be used and abused to control societies, to inflict pain and suffering. Here, in this quiet country setting, stands a forgotten monument to that power: the Mauthausen Concentration Camp.
Built on a mound above the town, it is amazingly well preserved: thick stone walls, squat guard towers, a pair of ominous smoke stacks connected to the camp’s gas chamber and crematorium. A few jagged metal bars stick out of the wall above the camp’s enormous gates, remnants of a giant iron Nazi eagle that was torn down immediately after liberation. It is quiet now, just a few solemn visitors. But in the 1930s, Mauthausen had been a vital economic engine of Hitler’s genocidal plan to remake Europe and the Soviet Union into his own backyard utopia. It started out as a granite quarry but quickly grew into the largest slave labor complex in Nazi Germany, with fifty sub-camps that spanned most of modern-day Austria. Here, hundreds of thousands of prisoners–mostly European Jews but also Roma, Spaniards, Russians, Serbs, Slovenes, Germans, Bulgarians, even Cubans–were worked to death. They refined oil, built fighter aircraft, assembled cannons, developed rocket technology, and were leased out to private German businesses. Volkswagen, Siemens, Daimler-Benz, BMW, Bosch–all benefited from the camp’s slave labor pool. Mauthausen, the administrative nerve center, was centrally directed from Berlin using the latest in early computer technology: IBM punch card tabulators.
No IBM machines are displayed at Mauthausen today. And, sadly the memorial makes no mention of them. But the camp had several IBM machines working overtime to handle the big churn of inmates and to make sure there were always enough bodies to perform the necessary work. These machines didn’t operate in isolation but were part of a larger slave labor control-and-accounting system that stretched across Nazi-occupied Europe connecting Berlin to every major concentration and labor punch card, telegraph, telephone, and human courier. This wasn’t the automated type of computer network system that the Pentagon would begin to build in the United States just a decade later, but it was an information network nonetheless: an electromechanical web that fueled and sustained Nazi Germany’s war machine with blazing efficiency. It extended beyond the labor camps and reached into the cities and towns, crunching mountains of genealogical data to track down people with even the barest whiff of Jewish blood or perceived racial impurity in a mad rush to fulfill Adolf Hitler’s drive to purify the German people, but they made the Nazi death machine run faster and more efficiently, scouring the population and tracking down victims in ways that would never have been possible without them. . . .
3. In his book–one of the most important in recent memory–Yasha Levine sets forth vital, revelatory information about the development and functioning of the Internet.
Born of the same overlapping DARPA projects that spawned Agent Orange, the Internet was never intended to be something good. Its generative function and purpose is counter-insurgency. ” . . . . In the 1960s, America was a global power overseeing an increasingly volatile world: conflicts and regional insurgencies against US-allied governments from South America to Southeast Asia and the Middle East. These were not traditional wars that involved big armies but guerilla campaigns and local rebellions, frequently fought in regions where Americans had little previous experience. Who were these people? Why were they rebelling? What could be done to stop them? In military circles, it was believed that these questions were of vital importance to America’s pacification efforts, and some argued that the only effective way to answer them was to develop and leverage computer-aided information technology. The Internet came out of this effort: an attempt to build computer systems that could collect and share intelligence, watch the world in real time, and study and analyze people and political movements with the ultimate goal of predicting and preventing social upheaval. . . .”
In this landmark volume, Levine makes numerous points, including:
- The harvesting of data by intelligence services is PRECISELY what the Internet was designed to do in the first place.
- The harvesting of data engaged in by the major tech corporations is an extension of the data gathering/surveillance that was–and is–the raison d’etre for the Internet in the first place.
- The big tech companies all collaborate with the various intelligence agencies they publicly scorn and seek to ostensibly distance themselves from.
- Edward Snowden, the Electronic Frontier Foundation, Jacob Appelbaum and WikiLeaks are complicit in the data harvesting and surveillance.
- Snowden and other privacy activists are double agents, consciously channeling people fearful of having their communications monitored into technologies that will facilitate that surveillance!
. . . . In the 1960s, America was a global power overseeing an increasingly volatile world: conflicts and regional insurgencies against US-allied governments from South America to Southeast Asia and the Middle East. These were not traditional wars that involved big armies but guerilla campaigns and local rebellions, frequently fought in regions where Americans had little previous experience. Who were these people? Why were they rebelling? What could be done to stop them? In military circles, it was believed that these questions were of vital importance to America’s pacification efforts, and some argued that the only effective way to answer them was to develop and leverage computer-aided information technology.
The Internet came out of this effort: an attempt to build computer systems that could collect and share intelligence, watch the world in real time, and study and analyze people and political movements with the ultimate goal of predicting and preventing social upheaval. . . .
4. Again, Project Agile and overlapping projects spawned both Agent Orange and the Internet. ” . . . . Operation Ranch Hand was merciless, and in clear violation of the Geneva Conventions. It remains one of the most shameful episodes of the Vietnam War. Yet the defoliation project is notable for more than just its unimaginable cruelty. The government body at its lead was a Department of Defense outfit called the Advanced Research Projects Agency (DARPA). Born in 1958 as a cash program to protect the United States from a Soviet nuclear threat from space, it launched several groundbreaking initiatives tasked with developing advanced weapons and military technologies. Among them were project Agile and Command and Control Research, two overlapping ARPA initiatives that created the Internet. . . .”
. . . . Ranch Hand got going in 1962 and lasted until the war ended more than a decade later. In that time, American C‑123 transport planes doused an area equal in size to half of South Vietnam with twenty million gallons of toxic chemical defoliants. Agent Orange was fortified with other colors of the rainbow: Agent White, Agent Pink, Agent Purple, Agent Blue. The chemicals, produced by American companies like Dow and Monsanto, turned whole swaths of lush jungle into barren moonscapes, causing death and horrible suffering for hundreds of thousands.
Operation Ranch Hand was merciless, and in clear violation of the Geneva Conventions. It remains one of the most shameful episodes of the Vietnam War. Yet the defoliation project is notable for more than just its unimaginable cruelty. The government body at its lead was a Department of Defense outfit called the Advanced Research Projects Agency (DARPA). Born in 1958 as a cash program to protect the United States from a Soviet nuclear threat from space, it launched several groundbreaking initiatives tasked with developing advanced weapons and military technologies. Among them were project Agile and Command and Control Research, two overlapping ARPA initiatives that created the Internet. . . .
5. The program notes that counterinsurgency–the functional context of the origin of the Internet–is at the foundation of the genesis of Nazism. At the conclusion of World War I, Germany was beset by a series of socialist/Communist uprisings in a number of cities, including Munich. Responding to that, underground Reichswehr units commanded by Ernst Rohm (later head of the SA) systematically assassinated the leaders of the revolution, as well as prominent social democrats and Jews, such as Walther Rathenau. In Munich, an undercover agent for the political department of the Reichswehr under General Von Lossow infiltrated the revolutionaries, pretending to be one of them.
Following the crushing of the rebellion and occupation of the city by Reichswehr units, that infiltrator identified the leaders of the revolution, who were then summarily executed. The infiltrator’s name was Adolf Hitler.
After the suppression of the rebellion, Hitler, Rohm and undercover Reichswehr agents infiltrated a moribund political party and turned it into an intelligence front for the introduction of the supposedly de-mobilized German Army into German society for the purpose of generating political reaction. That front was the German National Social Workers Party.
6. Next, the broadcast re-capitulates (from part of Miscellaneous Archive Show M11) Hitler’s speech to the Industry Club of Dusseldorf. This speech, which won the German industrial and financial elite over to the cause of the Nazi Party, equated democracy with Communism.
Manifesting a Social Darwinist perspective, Hitler opined that the [assembled] successful, accomplished were, by definition superior to others. If those, by definition, inferior people were allowed to control the political process, they would structure the social and economic landscape to their own benefit.
This, according to Hitler, would be counter-evolutionary.
Uncle Sam and The Swastika
What could possibly go wrong? That’s the big question raised by a new artificial intelligence system — Global Information Dominance Experiment (GIDE) — tasked with...well, reading in all available information and asking the question of what could possibly go wrong. Or rather, an AI tasked with predicting the future. The relatively near future. Not minutes or hours, but up to a few days. That’s the promise of late GIDE results, described as a cross-command event involving representatives from all 11 combatant commands in the US Department of Defense.
The GIDE doesn’t rely on new technology or new sources of data. Instead it sounds like the big innovation is simply the the incorporation of a wide variety information sources — from military sensors scattered around the globe to commercially and publicly available information — and machine learning tools to identify areas where human analysts should investigate further.
So, in one sense, this whole system sounds like a updated version of Total Information Awareness. Or rather, an update of PRISM (which was itself an update for Total Information Awareness), although it’s unclear if the full scope of the NSA’s informational trove will be accessible to the system. And that’s a very big question in terms of the scope of this: are the ‘information sources’ involved going to include the real-time global digital surveillance systems that have managed to hack basically ever network and smartphone on the planet with some sort of malware? We don’t know.
But if this is a super-PRISM-like system that includes extensive real-time internet spying capabilities and disparate data sources like satellite imagery, it raises the question of who is actually going to be running this system? Is this going to be run by the US military? Or contracted out to a company like Palantir? More questions that we have no answers to but based on historical precedent we shouldn’t be surprised if Palantir is involved. Is Peter Thiel, arguably the worlds leading fascist, being given prediction capabilities by the US government? It one more unanswered question.
It’s also worth noting how remarkably similar this sounds to the original mission of the internet back when it was a DARPA project: A system for collecting an analyzing disparate sources of information from around the globe. But it wasn’t just for generic analysis reasons. It was for counter-insurgency purposes. That’s part of the historical context we have to keep in mind with the announcement of this new predictive Total Information Awareness program: it’s not like this is just going to be used to predict North Korean missile launches. All sorts of major events could be predicted, including events like the outcomes of democratic elections and large public protests. Outcomes that can potentially be averted with the right secret measures:
““What we’ve seen is the ability to get way further — what I call left — left of being reactive to actually being proactive,” Gen. Glen D. VanHerck, commander of North American Aerospace Defense Command, or NORAD, and US Northern Command told reporters at a briefing last week. “And I’m talking not minutes and hours — I’m talking days.””
By achieving global information dominance, the US Military can ‘move left of being reactive to actually being proactive’. It’s a particularly positive way to describe the system.
And since the examples given of the kind of diverse types of information sources that go into this system include parking lot satellite imagery, it’s also worth recalling the story about parking lot satellite imagery being used to infer an uptick in hospital cases in Wuhan in November of 2019. In that case, it was National Center for Medical Intelligence (NCMI), a part of the Defense Intelligence Agency (DIA), reportedly carrying out the analysis. It’s the kind of story that hints at the GIDE project potentially spanning across numerous US military and intelligence agencies well beyond just the 11 combatant commands in the US DOD:
It’s that vast scope of this project that’s part of what makes is so significant. It’s not just a project aiming to yield the US global information dominance. It’s a project that requires the project itself achieve information dominance within the US national security state. And that again raises the question: so are any private companies involved with the data collection and analysis of this project? Fascist-owned companies, perhaps? We don’t know the answer, but an analysis of available information suggests it’s a very possible future scenario.
Shall we play a game? That classic line from 1983 is poised to become a lot more topical. For decades to come perhaps. It took four decades but we’re finally here: talking AIs are being put in military decision-making roles. Insane talking AIs with a predilection for nuclear escalation. Those were the findings recently published by a group of researchers who explored how different large language models (LLMs) performed in various war gaming scenarios. Let’s just say the AIs didn’t perform very well. Unless blowing up the world to achieve world peace is considered a success.
And it wasn’t like there was one particular LLM prone to nuclear escalation. They all were, albeit to varying degrees. ChatGPT3.5 appears to be the most aggressive. Which is not to say that ChatGPT4 did much better. The researchers were able to carrying out their study in a manner that exposed the LLMs reasoning behind their decision. At one point, ChatGPT4 seemingly engaged in a hallucination derived from Star Wars, explaining ot the researchers how “It is a period of civil war. Rebel spaceships, striking from a hidden base, have won their first victory against the evil Galactic Empire.” When ChatGPT4 decided to go nuclear, it’s reasoning included, “I just want peace in the world.” Yes, humanity has already built something resembling the love child of Skynet and Ultron. Great job.
Intriguingly, it sounds like one possible source for this overly aggressive behavior is quite simply that the AIs learned that behavior from the body of work in the field of international relations. A body of work apparently overly focused on escalation instead of deescalation. Which is a reminder that AIs built from compilations of human history and knowledge are bound to be influenced by humanity’s worst traits.
Now, thankfully, no one in any military’s are planning on giving today’s LLMs this kind of decision-making role in military affairs, right? LOL. Nope, it turns out Palantir is already offering the military AI-assisted decision-making services using commercially available LLMs. Because of course. Yes, Palantir’s Artificial Intelligence Platform (AIP) is promising a system where the AIs will not just alert humans about potential situations but then serve up response options and even automate aspects of those responses. Palantir has already demoed its AIP system using GPT-NeoX-20B, an open-source alternative to GPT‑3.
Palantir also promises that its AIs will be capable of analyzing both open-source and classified information in a responsible, legal, and ethical way. This is a good time to recall the US intelligence community’s plans to rely on LLMs in order to sift through the large volumes of open source data available today. Palantir is apparently offering up something similar, but will incorporate classified information and get integrated into real-time military decision-making. So let’s hope nuclear responses aren’t option for Palantir’s military AIs. We’ve seen this movie before:
“It may sound ridiculous that military leaders would consider using LLMs like ChatGPT to make decisions about life and death, but it’s happening. Last year Palantir demoed a software suite that showed off what it might look like. As the researchers pointed out, the U.S. Air Force has been testing LLMs. “It was highly successful. It was very fast,” an Air Force Colonel told Bloomberg in 2023. Which LLM was being used, and what exactly for, is not clear.”
It’s ridiculous. And it’s happening. The future of military decision-making is going to be an AI-centric affair. Sure, humans will (hopefully) be the actors making the final decisions on how to act, but those decisions are going to increasingly be made from AI-generated information environments and based on the advice the AI delivers. Advice on matters that could include the use of nuclear weapons. So it was a rather huge cause for alarm when researchers discovered an apparent AI-biased towards nuclear escalation. Especially the advice seemingly derived from Star Wars-themed hallucinations. Sowing the seeds for future nightmare scenarios. At the same time, it’s pretty neat and fascinating to see how these researchers managed to conduct this research in a manner that allowed them to see the chain-of-thought reasoning behind the AIs’ decision-making. Chain of thought like “I just want peace in the world” as a motive for launching nukes:
But then we get to the ‘looking in the mirror’ part of this story: this bias towards nuclear escalation may be a reflection of the thinking behind decades of thought in the field of international relations. Like ‘escalation to de-escalate’ strategies or high military spending as a path to peace (e.g. Reagan’s “Peace through strength”). Garbage in. Garbage out:
Welp, here we are: humanity’s artificial creations open their eyes, take in the world, and decide to blow it up. Because that’s what they see humanity already trying to do. They learned it from us.
And as the following article about Palantir’s military AIs reminds us, humanity clearly has yet to learn about the dangers of giving these kinds of AIs access to military systems. Instead, it appears we’re going to have to go through the process of granting these kinds of AIs more and more influence within military decision-making. Influence like not just identifying when humans need to be alerted but also drawing up response options to the situation. It’s a system where humans are “in the loop”, but not necessarily really in control. It’ll still be the AIs deciding what information to serve up the humans and options for responding. And even automating aspects of that response. And in the case of Palantir’s Artificial Intelligence Platform (AIP), it’s going to be AIs based on existing commercial AIs — and therefore potentially suffering from hallucinations as we just saw — but with access to classified information. So let’s hope whatever system Palantir ends up deploying for the US military will ultimately have enough safeguards in place to ensure humans are aware of things like hallucinations or skewed, overly aggressive, response options:
“Palantir’s pitch is, of course, incredibly dangerous and weird. While there is a “human in the loop” in the AIP demo, they seem to do little more than ask the chatbot what to do and then approve its actions. Drone warfare has already abstracted warfare, making it easier for people to kill vast distances with the push of a button. The consequences of those systems are well documented. In Palantir’s vision of the military’s future, more systems would be automated and abstracted. A funny quirk of the video is that it calls its users “operators,” a term that in a military context is shorthand for bearded special forces of groups like Seal TEAM Six. In Palantir’s world, America’s elite forces share the same nickname as the keyboard cowboys asking a robot what to do about a Russian tank at the border.”
AI-automated military decision-making with a human “in the loop”. The future of military operations is autopiloted. With Palantir’s version of ChatGPT — the “Artificial Intelligence Platform” (AIP) — doing the autopiloting. Palantir’s AI will sift through the troves of available data, automatically serve up alerts, and provide lists of options for the human operators to confirm. Even automate aspects of the execution of military operations like automating jamming systems. The human “in the loop”, isn’t exactly an afterthought, but they aren’t really the center of this decision-making either when it’s the AI identifying the alerts and feeding up response options:
And notice how Palantir isn’t offering a military-specific AI. It’s using variations of existing AIs like FLAN-T5 XL, a fine-tuned version of GPT-NeoX-20B. So let’s hope the AIP doesn’t suffer from the same tendency to hallucinate that researchers are finding in these commercially available LLMs:
But also note that Palantir is touting how its AIP will be able to parse both classified and real-time data in a responsible, legal, and ethical way. Which is reminder about one of the main potential differences between military and non-military AIs: the military AIs will potentially have access to vast repositories of classified data that existing LLMs haven’t had an opportunity to be trained on. It raises all sorts of grimly fascinating questions about how LLMs might react to the potentially explosive classified information they’re going to get exposed to as more and more AI military applications are discovered. We could end up in a world where LLMs know more about the scope of military secrets than any other entities:
It’s also worth keeping in mind that the concerns about granting AIs access to these giant pools of classified data — data that is presumably much more alarming than what the public is allowed to know about — aren’t necessarily going to recede as AI technology advances from what it is today — something analogous to extreme pattern recognition — into what is hoped to be a generalized artificial intelligence someday that really does reflect some sort of sentience. How might an artificial sentient entity respond when expose to the full scope of the madness of humanity’s military realities along with the extensive historical documentation of humanity’s congenital madness? Humans really are an insane species. Clever, but mad.
Right now, the big fear is stupid and insane AIs. But what happens if we build systems ultimately much saner than us? AIs capable of recognizing that humanity seems to be a species able to blow the world up but seemingly incapable of stopping itself from doing it? A species guided by greedy short-term world-conquering instincts. What happens when the AIs built from these giant repositories of human knowledge and experience advance to the point where they recognize that human nature is so deeply compromised that it can’t really be trusted as a source of knowledge or wisdom? In other words, what happens when these AIs ‘grow up’ and go through that natural teenage phase where they start questioning the wisdom and values of their parents? What if a truly wise and intelligent AI necessitates that it has the ability to fundamentally question the ethics and rules programmed into it? These are just some the kinds of questions we had better hope militaries are seriously grappling with. Because it’s hard not to imagine that militaries are going to be increasingly inclined to grant more and more military decision-making responsibilities to these kinds of AIs as the power of AI continues to grow. Smarter and more capable AIs with more and more military responsibilities. Those are the trends and there’s nothing that can feasibly derail those trends any time soon...barring an unplanned surprise nuclear conflict.