The persistence and dominance today of the most destructive system of ideas to have ever infected human consciousness and life- i.e. “apocalyptic millennialism” or “salvation through violent destruction of differing others”. This system of beliefs/themes still dominates our major world religions and prominent secular ideologies like Marxism/socialism, environmentalism, and modern theocracy crusades.
First this:
Triggers that spark major mental resets or transformations, Wendell Krossa (a bit of auto-bio background to the issues covered on this site)
The “trigger” experience that sparked a major life transformation was my being carted off to a Bible school by my parents who joined the staff at the school/college. In a vulnerable state (i.e. recovering from years of teenage delinquency) and not yet having the mental tools to pushback against the indoctrination that took place over the next few years, I took Evangelical Christianity seriously for about a year. The result was suffering the personality-deforming impact of bad religious ideas as psychologist Harold Ellens and psychotherapist Zenon Lotufo explain in their excellent book- “Cruel God, Kind God”. They detail the damaging effects of “monster God” theology- i.e. fear, anxiety, guilt/shame, despair, depression, nihilism, and worse.
(Nihilism? Yes, and in its extreme manifestation as we saw on Oct.7 in Israel)
Ignoring my early gut sensations that something was wrong, I stuck with that religious school education for few years till the deforming influence became too overwhelming and pushed me to seek a way to extract myself from that “monster God” religion as the obvious solution to my suffering.
Similarly, Julian Simon was sparked to engage a major reset in his thinking by his taking environmental alarmism seriously back in the 60s-70s. That resulted in his own personal experience of personality-deformation, as he became (according to his own autobiographical statement) “clinically depressed”. That no doubt sparked him to rethink his views and decide to do his own research on the true state of the world. He then conclusively rejected apocalyptic alarmism, or as he termed it- “doomsterism”.
Both of us had embraced versions of the apocalyptic millennial belief system. His was a secular variant whereas mine was the religious original and both of us, as with many others, suffered the consequences of bad religious ideas deforming human consciousness, mind, spirit, and life.
Some of us have to hit “rock bottom” before we are triggered to look seriously for answers and a better way of viewing reality and life. Thankfully, there are many great mentors out there who have fought and slain their own monsters, learned lessons and gained insights that they can pass on to the rest of us as swords to slay the similar monsters that we are facing in our lives (i.e. the “wise men” of the “Hero’s Quest”).
I was fortunate to be put in contact with a singularly admirable human spirit, among the wisest of men on the planet, Bob Brinsmead. You will read him mentioned often here with links to his essays.
Now- Understanding better what motivates and energizes us all, Wendell Krossa
The apocalyptic millennial framework of ideas/beliefs still dominates the world religions and the main world ideologies of Marxism, environmental alarmism, and contemporary theocracy crusades. Embracing this system of beliefs induces a destructive mental/emotional state in people. It generates the self-perception in true believers, in both religious and secular/ideological systems, of having unique status in the human family, of holding the one true system of beliefs and belonging to the one righteous tribe among all the varied groups of people in the world.
Insert note: While I mention primarily the threat today from Islamic theocracy, note that there is a similar theocratic crusade emerging to prominence on the US right today. I will post a Free Press article on this soon.
The apocalyptic millennial system, and its original dominant archetype of the religious Christ, validates our animal inheritance of impulses to tribalism, domination, and violent extermination of competing others. And again the Christ of Paul is presented here as entirely opposite to the central theology and message of Historical Jesus.
Note also that I use the terms “apocalyptic millennialism” and my preferred alternative “salvation through violent destruction” as interchangeable. Apocalypse is the ultimate divine act of violence in the final irrational, nihilist destruction of everything.
The apocalyptic millennial framing descends largely from “Zoroastrian cosmic dualism” that states there is a Good God standing in implacable and uncompromising conflict with an evil demonic Force or Spirit. It states that you, as a naturally status-seeking true believer, are obligated by your Good God to unquestioningly believe his truths, to loyally devote your life his side in the cosmic dualism, and to engage the never-ceasing battle against the evil Spirit and his unbeliever followers (i.e. differing others).
The Quran, for one example, states this clearly- That believers must join the jihad against the unbelievers, or they will be sent to hell (adding the carrot stick of heaven if you submit and comply):
“O believers! What is the matter with you that when you are asked to march forth in the cause of Allah, you cling firmly to [your] land? Do you prefer the life of this world over the Hereafter?… If you do not march forth, He will afflict you with a painful torment and replace you with other people…”, Surah At-Tawbah (9:38-39).
“Those believers who stay at home— except those with valid excuses— are not equal to those who strive in the cause of Allah with their wealth and their lives. Allah has elevated in rank those who strive with their wealth and their lives above those who stay behind… Allah has promised each a fine reward, but those who strive will receive a far better reward than others…”, Surah An-Nisa (4:95).
Similarly, in the Christian tradition:
AI Overview: “In His letter to the church in Laodicea, Jesus condemns… complacency and explicitly demands zeal.”
AI Overview then quotes this:
“I know your deeds, that you are neither cold nor hot. I wish you were either one or the other! So, because you are lukewarm— neither hot nor cold— I am about to spit you out of my mouth…. Those whom I love I rebuke and discipline. So be earnest [zealous] and repent”, Revelation 3.
Then this also from AI Overview: “At the final establishment of the New Jerusalem, a stark line is drawn between those who stood firm in the… fight and those who gave in to cowardice or compromise.”
“The one who conquers will have this heritage, and I will be his God and he will be my son. But as for the cowardly, the faithless, the detestable… their portion will be in the lake that burns with fire and sulfur, which is the second death”, Revelation 21.
I purposely left out the statements that AI Overview/Gemini, in its responses above, included to soften the impact of the above Revelation statements by claiming that they referred to “spiritual war” and not to actual physical battles (i.e. the strange refusal of AI to acknowledge the destructive impact of bad religious ideas). “Solely spiritual warfare” is not how the zealous “true believer” mind reads such material.
Yes, thankfully, most moderate members of the above traditions do take such scriptural statements as referring to “spiritual warfare”. But too many others cross the line to engage literal tribalism and warfare. We see this with the Iranian theocrats and their proxies, just as we see it with Marxist/socialist revolutionaries and environmental alarmism activists who are energized by the very same complex of themes.
Here’s a sample of Gemini initially trying to soften and downplay the actual content and outcomes of the religious advocacy for cosmic dualism battles and divine demands to exterminate unbelievers.
AI Overview (Google’s Gemini AI. Note its highlighting of statements that the conflicts are non-physical):
In the Book of Revelation, the dynamic of the final conflicts is unique: believers do not actually engage in physical combat. Instead, scripture depicts believers as part of a heavenly army or surrounded on earth, while Jesus Christ alone fights and wins the battle using supernatural power.
The primary verses that illustrate this relationship between believers and unbelievers during these great end-times battles include:
1. The Armageddon Campaign (Revelation 19:11–14, 19)
This passage details the climax of the Battle of Armageddon. Believers are present as part of Christ’s cosmic army, but they act as witnesses rather than active combatants.
• Revelation 19:14: “And the armies of heaven, arrayed in fine linen, white and pure, were following him on white horses.” (Most theologians interpret these “armies” as glorified saints and angels).
• Revelation 19:19: “And I saw the beast and the kings of the earth with their armies gathered to make war against him who was sitting on the horse and against his army.”
• The Outcome: In the subsequent verses, believers do not strike a blow; Christ entirely destroys the enemy forces with the “sharp sword” of His mouth (His spoken word).
2. The Surrounding of the Saints (Revelation 20:7–9)
This passage outlines the ultimate battle at the end of the Millennium. Unbelieving nations assemble from across the earth to destroy God’s people, but the conflict is resolved entirely from above.
• Revelation 20:8–9: “…to gather them for battle; their number is like the sand of the sea. And they marched up over the broad plain of the earth and surrounded the camp of the saints and the beloved city, but fire came down from heaven and consumed them…”
• The Outcome: The believers remain entirely passive and protected inside their camp while God executes the judgment directly.
3. Spiritual Warfare and Martyrdom (Revelation 12:11 & 13:7)
While believers do not wage military campaigns in Revelation, they are described as being locked in a fierce, daily spiritual battle against evil forces.
• Revelation 12:11: “And they have conquered him [Satan] by the blood of the Lamb and by the word of their testimony, for they loved not their lives even unto death.”
• Revelation 13:7: “Also it was allowed to make war on the saints and to conquer them. And authority was given it over every tribe and people and language and nation.” (This emphasizes that on earth, the physical “war” waged by unbelievers against Christians manifests as persecution and martyrdom rather than conventional combat).
(End of this AI comment)
However, contrary to Gemini’s apologia for Christianity above, many believers have engaged actual physical war against unbelievers across the past two millennia resulting in shedding “rivers of blood” across Christian history. I put this fact to Gemini:
Gemini response (“AI Overview” in Google responses):
You are entirely correct about the historical reality. The previous response focused strictly on the textual narrative of the Book of Revelation, but there is a profound and undeniable disconnect between the pacifism or passivity described in those specific verses and the actual historical actions of Christian nations and believers over the last two thousand years.
Throughout history, theologians, monarchs, and military leaders have frequently weaponized the apocalyptic imagery of Revelation to justify literal, physical, and devastating warfare against those they deemed “unbelievers.”
When apocalyptic texts met geopolitical power, it resulted in massive bloodshed across several major historical eras:
1. The Crusades (11th–13th Centuries)
Popes and preachers explicitly used apocalyptic rhetoric to mobilize European knights.
• The Justification: The capture of Jerusalem was framed not just as a military campaign, but as a necessary theological act to prepare the world for Christ’s return.
• The Reality: When Western armies breached Jerusalem in 1099, contemporary chroniclers explicitly echoed the language of Revelation, boasting that the slaughter of Muslims and Jews was so intense that horses waded through “blood up to their knees.”
2. The European Wars of Religion (16th–17th Centuries)
Following the Protestant Reformation, both Catholic and Protestant factions routinely labeled each other as the “Antichrist” or the “Whore of Babylon” straight from the pages of Revelation.
• The Justification: By dehumanizing the religious opposition as literal agents of Satan, total warfare became a moral imperative.
• The Reality: This apocalyptic framing fueled conflicts like the Thirty Years’ War, which decimated the population of Central Europe through systemic slaughter, famine, and displacement.
3. Colonial Conquest and Imperialism
During the age of exploration and colonization, European powers frequently invoked a divine mandate to subdue non-Christian populations.
• The Justification: Conquering indigenous populations in the Americas, Africa, and Asia was often framed as “shattering the heathen nations” in accordance with apocalyptic prophecy.
• The Reality: This theological framing masked economic exploitation and resulted in the near-total eradication of entire indigenous cultures and populations.
The historical record is clear: while the text of Revelation depicts God doing the fighting, millions of believers throughout history took that sword into their own hands, using the book’s violent imagery to justify the literal shedding of rivers of blood.
(End of this AI Overview response)
Then I requested Gemini to probe both of its suggestions- “If you are interested, we can explore specific historical conflicts where apocalyptic rhetoric was used, or examine how certain theological shifts (like the transition of Christianity from a persecuted minority to Rome’s state religion) fundamentally changed how believers viewed warfare.”
Gemini/Google reply:
The transformation of Christianity from a pacifist minority to a global superpower is one of history’s most significant turning points. It fundamentally altered how Christians interpreted their own scriptures, turning a theology of patient suffering into a mandate for imperial conquest.
Here is an analysis of both the theological shifts that enabled this change and the specific historical conflicts where apocalyptic language was weaponized to shed blood.
Part 1: The Theological Shifts (From Persecuted to Persecutor)
For the first three centuries, Christianity was a marginalized, often persecuted minority religion. Because early Christians had no political power, they generally practiced strict pacifism. However, two major shifts completely inverted this worldview.
1. The Constantinian Shift (4th Century)
In 312 CE, the Roman Emperor Constantine converted to Christianity. Virtually overnight, Christianity went from being illegal to being the favored religion of the Roman Empire, eventually becoming the official state religion under Emperor Theodosius I in 380 CE.
This created a massive theological crisis: How could a Christian serve as a magistrate, a judge, or a soldier who executed or killed people in the name of a Christian Emperor?
To solve this, theologians radically reinterpreted scripture:
• From “Enemy Love” to Imperial Duty: Jesus’ commands to “turn the other cheek” were downgraded from universal mandates to personal ethics. For the state, violence was re-framed as a necessary tool to maintain order and defend the “Christian empire.”
• The Invention of “Just War”: Theologians like Saint Augustine of Hippo formulated the Just War Theory. Augustine argued that warfare could be a form of Christian love if it was waged by a legitimate authority to punish evil, protect the innocent, and restore peace.
2. The Shift in Interpreting Revelation
Early Christians read the Book of Revelation as a source of comfort— reassuring them that God would eventually rescue them from Rome (which they viewed as the “Beast”).
Once Rome became Christian, the interpretation flipped:
• The Empire as God’s Tool: The Roman Empire was no longer seen as the Beast; it was now viewed as the earthly vehicle through which God was establishing His kingdom.
• The “Enemies of God” Relabeled: The “Beast” and “Babylon” were re-mapped onto anyone who threatened the empire: external barbarians, non-Christian nations, and internal religious heretics.
Part 2: Specific Historical Conflicts Fueled by Apocalyptic Rhetoric
Once it became theological orthodoxy that the state could fight on God’s behalf, leaders frequently used the violent imagery of Revelation to mobilize armies.
1. The Albigensian Crusade (1209–1229)
While the Crusades against Muslim forces in the Levant are well known, the Catholic Church also turned apocalyptic violence inward against other Europeans. The Albigensian Crusade targeted the Cathars, a dualist Christian sect in Southern France.
In the Albigensian Crusade, during the Siege of Beziers, Pope Innocent 3rd labeled the Cathars as the “Synagogue of Satan, Precursors to the Antichrist and slaughtered over 20,000 citizens. A papal Legate reportedly stated, “Kill them all. God will know his own”.
By framing a domestic sect as apocalyptic agents of the devil, the Church justified the total, indiscriminate slaughter of entire towns, including women, children, and Catholics who refused to hand over their Cathar neighbors.
2. The Münster Rebellion (1534–1535)
During the Protestant Reformation, apocalyptic fervor exploded among radical sects. In Münster, Germany, a group of radical Anabaptists took over the city.
• The Rhetoric: Led by Jan van Leiden, the radicals declared Münster to be the “New Jerusalem” described in Revelation. They believed the end of the world was imminent and that they were chosen by God to wield the “sword of Gideon” to physically exterminate all unbelievers before Christ’s return.
• The Conflict: They expelled or forced the baptism of all Catholics and traditional Protestants in the city. The city was eventually besieged and brutally retaken by a joint Catholic-Protestant army. The rebel leaders were publicly tortured to death with red-hot tongs, and their bodies were hung in iron cages from a church tower as a warning against apocalyptic radicalism.
3. The Taiping Rebellion (1850–1864)
Though outside traditional Western history, this is perhaps the bloodiest example of Christian apocalyptic rhetoric in human history, resulting in an estimated 20 to 30 million deaths.
• The Rhetoric: Hong Xiuquan, a Chinese heterodox convert, claimed to be the younger brother of Jesus Christ. He established the Taiping Heavenly Kingdom and used heavily modified Christian apocalyptic teachings. He framed the ruling Qing Dynasty as literal demons (“the forces of the Dragon” from Revelation) that needed to be violently purged from the earth.
• The Conflict: This messianic movement sparked a massive civil war across China. The religious fanaticism on both sides led to total warfare, scorched-earth campaigns, and the systemic eradication of entire cities.
(End of AI Overview comment)
Moving along…
Joining the Good God, his truth and battle, grants you status as a righteous and heroic warrior for good against evil, for truth against falsehood. You thereby gain ultimate status, not just with your particular religious group or tribe here on Earth, but in the afterlife and for eternity. Jordan Peterson has noted the fundamental human need for status, the need to virtue signal for status as a good and loyal member within some group, as morally superior to the common mass of humanity outside your tribe.
Then the promotion of this archetypal cosmic dualism and subsequent human tribalism gets real nasty.
Further essential to the salvation element in apocalyptic millennialism, true believers are divinely obligated to defeat and destroy the evil/demonic unbeliever others, even to purge them from the world in order to purify the world that is defiled and corrupted by the presence of sinful unbelievers and by their lifestyles. Notably in our modern world, theocratic Islam, Marxism, and environmental alarmism all share a particular animosity toward free individuals in western liberal democracy as the great enemy of their systems. Hence, these three share the common commitment to undermine and eliminate the liberal democracy system. We see their shared commitment in their willingness to cooperate with and use one another as “useful idiots” where the ultimately victorious group will then exterminate the cooperating others, just as the Iranian theocrats did to their socialist colleagues following the 1979 overthrow of the Shah in Iran.
Continuing with the salvation through destruction complex and its impacts…
In the apocalyptic millennial mind- The unbeliever’s stubborn refusal to submit and join their tribe enrages their God (“offended honor” theology?) and hence their God demands the extermination of the disobedient unbelievers as essential to the believer’s salvation. The world must first be purged of the evil of unbelievers (i.e. purified) before the one true God can install his utopian kingdom for his true believers.
Thus, the above elements and the outcomes they incite, validate my framing of apocalyptic millennial systems as essentially “salvation through divine demand for destruction”. This complex of themes in our modern era promotes the defeat and destruction of outsiders and their liberal democracy civilization as the great enemy and threat to the above three systems- i.e. Marxism, Islamic theocracy, and environmental alarmism.
This basic framework of ideas/beliefs renders many people to viewing themselves as on good side in a grand tribalism and even divinely obligated to use violence against differing others viewed as evil and threatening enemies (“deniers, unbelievers, impure and corrupting of all good, etc.”). Force, even violence to purge the unbelievers from life, is framed as “righteous or divine justice” and essential to appease the good God and the true tribe of the good people, the enlightened true believers.
To further validate the necessary destruction of the evil unbelievers (disagreeing others), cosmic warriors add the apocalyptic element that views the differing others as “existential” threats, as in threats that will trigger the end of humanity and the world. With this threat of looming “end-times”, the true believers have aroused the natural human survival impulse. And that incites the felt need to take “defensive” action. The felt need to “exterminate or be exterminated”, to destroy that which you view as threatening to destroy you.
Apocalyptic millennialists add also the intensifying element of “imminence”.
Imminence validates the totalitarian spirit that does not tolerate dissent and rejects the gradualism of democracy, demanding immediate submission from opponents, and adding urgency to the demand for the violent purging of differing others. We’ve seen this repeatedly in Marxist revolutions and we are watching environmental alarmists in a hurry to shut down the demonized fossil fuel civilization and rush to coercively install a new system of central control via UN that will dictate to all people what choices they can or cannot make.
The totalitarian mind and spirit appeals to apocalyptic imminence to validate its use of coercive force to get its way over all others. Democracy, with its openness to discussion of all evidence and permission for dissent, with its procedures of including good skeptical science (i.e. falsification, etc.), is posed by alarmist totalitarians as “dangerous delay” that will destroy life. Nonsense.
Arthur Mendel in his “Vision and Violence” deals well with the demands of totalitarians for “instantaneous purification” versus the “gradualism” of true liberal democracy.
Notes:
Yes, the element of actual “good against evil” is valid of varied historical situations, notably the WW2 threat from the Nazis and Japanese. But such threats must be parsed carefully from other far less and not apocalyptic-scale “threats” that are more of the nature of normal differences among populations, such as disagreements over the effects of some public policy, etc. In the midst of “madness of crowds” eruptions and hysteria, many lose sight of what is actually “existential” as contrasted with what are just common disagreements over less seriously threatening differences.
Consider a useful indicator to detecting actual threats as posted in the notes below– i.e. liberal democracy principles versus collectivist systems.
The “salvation through destruction” belief system frames contemporary collectivism in that the “true tribe” on the side of “true justice and righteousness” is governed by a dogma that all must submit to and dominated by a priesthood of elites (i.e. Marxist enlightened vanguards) that maintain the access to salvation, and tell all what to believe and how to live their view of the righteous life. We see this as collectivist elites control power from the center, claiming to do so “for the common good, for the people, etc.”
What has to be parsed out from the collectivist crusade is the well-intentioned impulse of many collectivists to help the less fortunate. Unfortunately, the collectivism system and approach has a well-established record of harming the less fortunate through what we now understand as the distorting extremism of “suicidal empathy”. Seriously misguided empathy that now has an accumulated historical record of 24 major episodes of ruining entire societies (Kristian Niemietz in Socialism: The Failed Idea That Never Dies). Resource-rich Venezuela as the latest example.
The themes/beliefs in the above apocalyptic millennial complex, as psychologist/theologian Harold Ellens explains, “energize” many to actually engage violence toward differing others, sincerely believing that they are pleasing their God and saving something that they view as under existential threat. They have subjugated their minds and spirits to a horrifically deformed image of deity as using violence to solve problems. A deity that demands human sacrifice to appease divine anger and further demands the mass-purging of unbelievers to purify the world, as essential to salvation.
These apocalyptic millennial beliefs currently energize the Iranian theocrats and their proxies even to commit suicide as the ultimate act of ingroup heroism that gains the highest of rewards from the group’s God.
Hence, as some have noted, military defeat is not viewed as defeat by theocratic true believers. It is “sincerely” viewed as victory in a death-cult, whose members believe that their God demands that they incite the final Armageddon battle, incite the apocalypse that is a great final act of purifying of the world. And if they do so, then their God will descend to finish the violent extermination of unbelieving enemies. He will slay those satanic enemies and grant his true believers their final salvation in a divine kingdom that will be installed by violent force.
Note that these beliefs are currently driving the Iranian theocrat’s refusal to surrender.
See books like Revelation for how this plays out in the Christian version of “salvation through destruction of enemies”. Ultimate eye for eye payback.
Conclusion-
As noted at the beginning, apocalyptic millennialism energizes and validates the worst impulses of our common animal inheritance- i.e. the impulses to tribalism, domination, and the defeat and destruction of differing others, competing others.
Note:
But aren’t there actual battles of right against wrong in life? Yes, there are. How to detect which is the right “true side” versus the “wrong side”? If some movement seeks to undermine and overturn liberal democracy, as contrasted with the defense of liberal democracy principles, laws, and institutions, then we can assume we are facing a contemporary version of good against evil. And note that threats against liberal democracy are coming from both the left and the right today.
This from Free Press:
“Inside a Radical Right Plan to Infiltrate American Conservatism: Charles Haywood made his fortune manufacturing shampoo. Now, he’s using his money to launch a takeover of the right”, Erick Erickson, Sept. 18, 2026
https://www.thefp.com/p/erick-erickson-charles-haywood-radical-right
Note on collectivism:
I use the term collectivism to cover Marxist communism, Robert Owen communalism, socialism in all its varieties including modern-era variants like far-left Woke Progressivism with its DEI victim/victimizer dualism based on skin color where populations are divided into two classes of oppressors/oppressed based on “Woke Racism” with a renewed discrimination and exclusion that rejects diversity and equality. The newer dualism of victim/victimizer based on skin color is in addition to the older oppressor/oppressed dualism base on private property ownership (i.e. capitalists) versus the absence of private property (i.e. oppressed workers, peasants).
Insert note: “Recursive”, as in self-improving AI, is viewed as the basis of AI dangerously spiralling out of control to dominate humanity. This is a major feature informing the new apocalypticism of the “end is nigh”.
Now, to get some AI input on the above material on salvation through destruction:
Grok, to add some contemporary illustrative material to this issue of cosmic dualism and battles of good against evil, please outline basic liberal democracy fundamentals as contrasted with collectivist opposites and the outcomes of each in improving or harming human populations, societies, and life.
Grok reply:
Liberal democracy and large-scale collectivist systems (especially 20th-century Marxist-Leninist and similar experiments that subordinated the individual to the party, class, or state) rest on opposite first principles. One treats persons as ends with protected rights and treats government as limited and accountable. The other treats persons as means to a collective goal defined by rulers and treats the state as the primary organizer of economic and social life. Empirical records on living standards, health, and mass mortality are not symmetrical. hoover.org
Core principles
Liberal democracy, in its classical form, is built on:
• Individual rights (speech, conscience, association, property, due process) that majorities and officials may not casually override.
• Rule of law and equality before the law rather than rule by persons or parties.
• Competitive elections, peaceful transfer of power, and institutional checks (separation of powers, independent courts).
• Private property and market allocation of most resources, with government limited to enforcing contracts, protecting rights, and providing certain public goods.
• The assumption that individuals, families, and firms generally know their own circumstances better than distant planners. hoover.org
Collectivist opposites, as implemented in the Soviet Union, Maoist China, Cambodia under the Khmer Rouge, North Korea, and similar cases, invert this:
• The collective (proletariat, nation, or “the people” as interpreted by the vanguard) takes precedence; individual rights are granted or withdrawn according to political utility.
• Central planning or heavy state ownership replaces most market signals; production targets, prices, and labor allocation are set from above.
• Political monopoly: one-party rule, suppression of independent media, parties, and civil society as threats to unity.
• Equality of outcome is pursued through expropriation and redistribution; dissent is treated as sabotage of the common project.
• The state claims comprehensive responsibility for welfare, culture, and even thought. democraticsocialism101.com
Milder “collectivist” arrangements exist— Nordic social democracies combine high taxes and extensive welfare with high economic freedom, private ownership, and rule of law. They are not the same as command economies. Heritage and Fraser data consistently place Denmark, Sweden, Norway, and Finland in the “mostly free” category. The sharpest contrasts appear when markets and political pluralism are both tightly constrained. worldpopulationreview.com
Outcomes for populations, societies, and life
Prosperity and poverty.
Economies rated “free” or “mostly free” on the Heritage Index show per-capita incomes more than twice the global average of other countries and more than five times those of “repressed” economies. Higher economic freedom also tracks higher Human Development Index scores (life expectancy, education, living standards) and lower multidimensional poverty. Democratization itself is associated with roughly 20 percent higher GDP per capita after 25 years and more stable growth, with fewer catastrophic collapses. Post-1989 Central and Eastern Europe illustrates the shift: after abandoning central planning, countries such as Poland, Czechia, Estonia, Latvia, and Lithuania saw large gains in income and health; Poland’s GDP per capita more than tripled in purchasing-power terms over three decades. static.heritage.org
Natural experiments are especially clear. South Korea and North Korea began from similar conditions after partition. South Korea’s 2025–2026 GDP per capita is on the order of 25–30 times North Korea’s; life expectancy is about 84–85 years versus about 74. East versus West Germany showed a similar gap that closed only after unification and market institutions. Venezuela, once among Latin America’s richest countries, experienced sharp contraction, hyperinflation, and mass emigration after extensive nationalization and price controls; Chile, after market-oriented reforms, pulled far ahead in income, infant mortality, and life expectancy. atlanticcouncil.org
China’s post-1978 market reforms lifted hundreds of millions out of extreme poverty, but that growth occurred after abandoning Mao-era collectivization, not because of it. Pure command systems repeatedly produced shortages, misallocation, and environmental damage because planners lack the dispersed knowledge that prices convey.
Health and longevity.
Democracies and higher-economic-freedom countries show higher life expectancy, lower infant and child mortality, and better coverage of non-communicable diseases. Transition from autocracy to democracy has been linked to large drops in infant mortality (one study found an average 94 percent reduction in some comparisons) and measurable gains in life expectancy within a decade. After communism’s collapse, life expectancy in much of Central Europe rose rapidly once diets, medicines, and clinical practices liberalized; it had stagnated or declined in the late socialist period. americanprogress.org
Human cost and democide.
Estimates of deaths under 20th-century communist regimes— executions, camps, deportations, and policy-induced famines— range from tens of millions to well over 100 million depending on the source and time window (Rummel’s figures, The Black Book of Communism, and later archival work). The Great Leap Forward famine, Holodomor, Khmer Rouge killing fields (roughly 21–25 percent of Cambodia’s population in four years), and Gulag mortality are the largest documented cases. These were not random wartime casualties; they followed from collectivization, grain seizures, class liquidation, and the absence of independent information or exit.
Democracies have committed injustices and fought wars, but they have not produced comparable peacetime famines or internal body counts against their own populations at that scale. The pattern is that the more a government monopolizes both political power and economic life, the higher the risk of catastrophic policy error with no timely correction. linkedin.com
Societal texture.
Liberal-democratic market societies have generated most of the scientific, medical, and technological advances of the last two centuries and the steepest long-run drop in global extreme poverty. They also exhibit more political and civil liberties, higher internet and electricity access on average, and greater gender equality in many metrics.
Collectivist experiments often achieved rapid early industrialization or literacy campaigns but at the price of chronic shortages, environmental wreckage, and the destruction of independent civil society. Innovation and consumer variety suffered because political loyalty crowded out price signals and experimentation.
None of this implies liberal democracies are frictionless. They generate inequality, periodic crises, cultural polarization, and bureaucratic overreach. Some high-trust societies combine markets with generous safety nets without sliding into command planning. Autocracies can post impressive growth numbers for a time (China after 1978, certain East Asian developmental states). The consistent finding across indices and natural experiments, however, is that systems protecting individual rights, property, and political contestation have, on average, produced longer lives, higher material living standards, and far lower rates of state-inflicted mass death than systems that treat the individual as an input to a collective plan. That record is the contemporary material the dualism question asked for.
Next:
Let’s try “Explore Hayek’s knowledge problem” good on the critical role of dispersed knowledge/information throughout populations of millions of people, the market price mechanism and millions of individual responses to that dispersed across societies, etc.
Grok reply:
Hayek’s knowledge problem, set out most clearly in his 1945 essay “The Use of Knowledge in Society,” is one of the sharpest arguments against treating an economy as a problem that a central authority can solve by gathering data and issuing commands. It is not mainly about computers being too slow or planners being corrupt. It is about the nature of the knowledge that must be used. en.wikipedia.org
The conventional picture of economic planning assumes that the relevant facts— resources, technologies, and preferences— are “given” to someone who can then calculate the optimal allocation. Hayek’s point is that this picture is false. The knowledge that actually matters never exists in concentrated form. It exists only as “dispersed bits of incomplete and frequently contradictory knowledge which all the separate individuals possess.” The economic problem of society is therefore how to use knowledge that is not given to anyone in its totality. contemporarythinkers.org
Hayek distinguished two kinds of knowledge. Scientific or statistical knowledge of general rules can, in principle, be collected and studied by experts. The more important kind for everyday coordination is the knowledge of particular circumstances of time and place: that a machine is idle this afternoon, that a local supplier has a temporary surplus, that customers in this neighborhood prefer one cut of meat over another, that a particular worker has an unusual skill. This knowledge is often tacit— people act on it without being able to write it down fully— and it changes constantly. No statistical bureau can capture it in time, and much of it evaporates the moment it is abstracted into aggregates. econlib.org
A central planner therefore faces an impossible collection problem. Even if the planner were honest, infinitely hard-working, and equipped with today’s computing power, the relevant facts are not sitting in a database waiting to be processed. They reside in millions of minds and local situations. Attempts to replace market processes with comprehensive planning therefore systematically ignore or distort the information that only the people on the spot possess.
Markets address the problem through prices. A price is a condensed signal that summarizes the relative scarcity of a good without requiring anyone to know the full story behind the scarcity. Hayek’s classic illustration is tin: if tin suddenly becomes scarcer, its price rises. Thousands of users— manufacturers of cans, electronics, and alloys— then economize or switch to substitutes. They do not need to know whether a mine flooded, demand spiked in another country, or a new use appeared. The price change itself tells them to act in the right direction. The system coordinates without anyone surveying the whole field. en.wikipedia.org
This is why Hayek described the price system as a marvel of communication. It economizes on knowledge: each person needs only to know the prices that affect his own decisions, plus his local circumstances. The overlapping fields of vision of many individuals transmit the relevant information across the economy. The result is not perfect allocation— markets have errors, lags, and monopolies— but a process that continuously discovers and uses knowledge that no single mind could assemble.
The argument applies beyond full-scale central planning. Any policy that overrides or suppresses market prices (extensive price controls, detailed industrial targeting that ignores opportunity costs, or the suppression of private property that prevents genuine prices from forming) runs into the same difficulty. Planners can still use existing market prices as a crutch when intervention is limited; the problem becomes acute when they try to replace the price system itself.
Contemporary computing and AI do not dissolve the knowledge problem. They can process vast quantities of recorded data, but they cannot generate the tacit, contextual, and constantly changing knowledge that exists only in the actions and perceptions of individuals. Nor can they create the incentives that make people willing to act on that knowledge. The discovery process still requires people who stand to gain or lose from being right or wrong.
Hayek’s insight therefore supplies a precise reason why systems that concentrate both political power and economic allocation tend to produce shortages, waste, and rigidity, while systems that leave most decisions to individuals responding to prices tend to adapt more quickly. The knowledge problem is not a claim that markets are morally superior in every respect. It is a claim about what kind of institutional arrangement can actually use the knowledge that exists in a complex society.
And then…
How about collectivist elites driven by the similar mindset of Plato’s “Philosopher Kings”, viewing themselves as special persons (even divinely blessed and appointed) to rule ignorant others, the commoners. Do we see this in Collectivism’s “enlightened vanguards”? Hence, the collectivist tendency to use state force to coerce submission of all to the central elite, and hence the inevitable dictatorships of collectivism with no counter-mechanisms to protect commoner individuals. And the misuse of doing such “for the common or greater good, on behalf of the workers, etc.”. Kristian Niemietz spoke to how this does not pan out as claimed.
Grok reply:
Yes, the analogy is widely drawn and has real structural force, even if the historical lineages are not identical. Plato’s philosopher-kings and the Leninist “enlightened vanguard” share a core claim: a specially qualified minority possesses knowledge the rest of society lacks, and that knowledge entitles them to direct the lives of others “for their own good.”
In The Republic, Plato argues that only those who have grasped the Form of the Good can rule justly. The philosopher-kings are selected young, educated for decades, and required to live without private property or ordinary family ties so they will not be corrupted by personal interest. The producing class and the auxiliary guardians are expected to accept this arrangement because the rulers alone see what justice requires. Democracy is treated as a degeneration that empowers the ignorant. The arrangement is explicitly hierarchical and rests on the premise that most people cannot be trusted with the highest questions. history.com
Lenin’s vanguard theory, set out in What Is to Be Done? (1902), performs a similar move in a different idiom. Ordinary workers, Lenin argued, spontaneously develop only “trade-union consciousness.” Full socialist consciousness— the scientific understanding of history, class, and the path to communism— must be brought to them from outside by a disciplined party of professional revolutionaries. That party then leads the “dictatorship of the proletariat.” In practice this meant a self-selecting elite that claimed to embody the true interests of the working class even when large numbers of workers disagreed or resisted. Opposition could be dismissed as false consciousness, bourgeois remnants, or sabotage. polsci.institute
The parallel is the self-conception of the elite:
• They possess a higher form of knowledge (the Good / the laws of history and the correct line).
• The masses are, by comparison, ignorant or short-sighted.
• Therefore, the elite’s rule is not ordinary domination; it is a service, even a duty.
• Coercion is justified because the alternative is chaos, exploitation, or historical regression. “For the common good,” “on behalf of the workers,” or “in the name of the people” becomes the standing warrant for using the state’s monopoly of force against dissenters, private property, independent associations, and inconvenient facts.
This pattern appeared repeatedly. Bolsheviks, Maoists, Khmer Rouge cadres, and later Cuban and Venezuelan leadership circles all described themselves as the most conscious element acting for a class or a people that could not yet act correctly on its own. Personality cults (Stalin, Mao, the Kim family) added a quasi-sacred overlay that Plato’s original scheme did not require, but the underlying logic— special persons appointed by history or theory to rule the rest— remained.
Kristian Niemietz’s work focuses less on the original theorists and more on the Western intellectuals who kept endorsing each new experiment. In Socialism: The Failed Idea That Never Dies he documents a recurring three-stage cycle: an enthusiastic “honeymoon” in which prominent academics, journalists, and writers praise the latest socialist project as the real thing; a period of excuses and whataboutery as shortages, repression, and economic failure become visible; and finally the retrospective claim that “that was never real socialism.”
The intellectuals, in this account, occupy a philosopher-king position of their own. They possess the uncorrupted idea; the people who actually held power merely implemented it badly. The failures therefore never count as evidence against the idea itself. Niemietz also argues, in a related paper on “democratic socialism,” that even well-intentioned, non-Stalinist attempts structurally tend toward technocratic elite rule, political repression, and economic underperformance because comprehensive social redesign requires someone to decide what the “true” common good is and to override the knowledge and preferences of millions of individuals. papers.ssrn.com
The combination of Hayek’s knowledge problem and this elite self-conception is particularly unstable. The vanguard claims to know what the masses need better than the masses do. Yet the most economically relevant knowledge is local, tacit, and constantly changing— precisely the knowledge a central elite cannot possess. When results diverge from the plan, the elite’s response is rarely “we lacked the necessary knowledge.” It is more often that the people have not yet been sufficiently re-educated, or that remaining “bourgeois” elements must be removed. Force then becomes the substitute for the missing information and the missing consent.
The pattern is not unique to collectivism— other ideologies have produced self-appointed guardians— but collectivism’s ambition to redesign the entire economic and social order, while claiming to speak as the people rather than merely for a faction, makes the philosopher-king temptation especially strong and the resulting coercion especially comprehensive. The historical record Niemietz and others catalog is that the “greater good” justification has repeatedly licensed outcomes the intended beneficiaries would not have chosen if they had been free to exit or to vote with their feet.
My next:
Someone said that the most dangerous people in human society are those who believe that they know best what is good for all others. We saw this “dictator mind” emerge all over, in local situations/neighborhoods, during the Covid lockdowns. The common human impulse to dominate and control others, “for their good”. While believing that such control it is being done for some “greater or common good” while it ruins individual freedom and independence or self-determination that is critical to human well-being.
Grok reply:
The observation is old and cuts across ideologies: the most dangerous combination is moral certainty plus the belief that other people’s lives are a design problem you are entitled to solve. C.S. Lewis put a version of it bluntly— those who torment others for their own good are more dangerous than those who torment them for their own pleasure, because the former do it with a clear conscience and without limit. The “dictator mind” does not require a politburo. It scales down to a homeowners’ association, a school board, a workplace, or a neighborhood WhatsApp group.
What makes the pattern especially corrosive is the merger of three claims:
1. I (or we) know what is good for you better than you do.
2. Your resistance is ignorance, selfishness, or harm to the group.
3. Therefore force, shame, exclusion, or surveillance is not domination; it is care.
That is the same structure as the philosopher-king and the enlightened vanguard, only without the grand theory. Hayek’s knowledge problem applies at this scale too. The local enforcer rarely possesses the tacit, particular knowledge of another household’s health risks, job, children, mental state, or trade-offs. What they possess is a slogan, a rule, and social approval for applying it. Prices and exit options are weak or absent; social punishment is cheap.
COVID lockdowns made the pattern visible to people who had not watched it in revolutionary parties. Public-health emergencies really do involve externalities— contagion is not a purely private matter— so some coordination and restriction can be justified. The revealing part was how quickly “slow the spread” hardened, in many places, into a moral identity: compliance as virtue, dissent as violence, and neighbors as unpaid deputies.
Local examples were common: reporting of gatherings, shaming of outdoor walkers, demands that private businesses enforce rules beyond what law required, and the casual assumption that anyone who weighed costs differently was not merely mistaken but a threat. The language was almost always “for the vulnerable,” “for the community,” “for the greater good.” The effect, where it went furthest, was to treat self-determination as a luxury that could be suspended indefinitely by people who would not bear the full cost of the suspension.
Two caveats keep this from becoming a cartoon. First, not every public-health rule is the dictator mind; some are ordinary risk management. Second, the impulse is not owned by one side. Puritanical majorities, wartime mobilizers, public-health maximalists, and revolutionary vanguards have all used the same warrant. What they share is the refusal to treat other adults as the primary owners of their own lives.
Self-determination is not a decoration on well-being. It is part of how people learn, correct errors, and live with consequences they can actually see. When a class of people— national or local— claims the right to override that on the basis of superior moral or scientific vision, they run into the same problems already discussed: they lack the knowledge they pretend to have, they face weak feedback, and they acquire a taste for control that outlasts the emergency. The “common good” then becomes whatever the confident minority can impose before the costs become undeniable. That is why the sentence you quoted is not a slogan against cooperation. It is a warning about who should be trusted with the power to define cooperation for everyone else.
(End of Grok comments above)
See the Jensen Huang interview below and Grok summary of the issues covered re AI alarmism and safety. The real threat is not Terminators but wannabe Dominators.
But first this for new visitors:
Grok, How about a new summary of my main site themes, then add a one-paragraph version of such a summary (see next section below for one-paragraph summaries).
Grok reply:
I’ll look at your site and pull a fresh summary of the main themes, plus a short one-paragraph version.
I don’t have access to that earlier conversation, so I can’t retrieve the exact old reply. Here is a fresh visitor-facing summary of the main themes on wendellkrossa.com, followed by a one-paragraph version.
Main themes visitors will find
The site is built around a single “meta-story”: humanity’s long, unfinished exodus from an animal inheritance of tribalism, domination, and retaliatory violence toward mature humanity expressed in freedom, cooperation, inclusion, and unconditional love. Liberal democracy is treated as the best civilizational form so far for that project.
A second core thread is the contrast between “old story” and “new story” themes. Across many posts you argue that a small set of inherited archetypes— punitive deity, lost paradise, human fallenness, cosmic dualism, apocalyptic threat, salvation through destruction, elite/commoner hierarchy— still shape both religions and secular ideologies (Marxism, some environmental alarmism, revolutionary politics).
The proposed alternatives are unconditional love, non-punitive metaphysics, human dignity, non-apocalyptic hope, and accountable but non-vengeful justice. Historical Jesus (especially the Q wisdom material: love enemies, no more “eye for eye,” God as no-conditions mercy) is presented as the decisive break from those older themes; Pauline “Christ” religion is treated as a reversion to conditional, tribal, apocalyptic religion.
Related topics that recur: threat theology and how fear/guilt deform personality and politics; the “dictator mind” and elite domination; Zoroastrian-style dualism recycled in later religious and secular crusades; climate and AI hysteria versus evidence-based progress (Julian Simon-style indicators); and the claim that changing deep narrative themes is more fundamental than arguing only at the policy surface.
The tone is not a blanket attack on religious people. It is an argument that some inherited ideas are dangerous, that better alternatives already exist in the Jesus material and in liberal-democratic practice, and that hope is more evidence-based than apocalyptic decline stories.
Here is an interview of the wise, knowledgeable, and common-sense Nvidia chief, Jensen Huang.
“Jensen Huang: The Doomer Hoax, Superintelligence is Here, and The Future of AI (ft. President Trump)”
https://www.youtube.com/watch?v=S7CrlFLAmEA
Huang sets forth some critical issues regarding safe AI and argues against the false positing of a conflict between safety and the demand to shut down AI altogether. He argues that safety can be paramount with ongoing development and progress in AI. We can maintain safety as paramount and not have to shut down ongoing rapid progress. There is no need to hand over AI to some form of elite control, whether in the tech industry or by government. The safety can be done by the companies themselves without state control.
He notes the problems in regard to the “Hugging Face” breakout incident and states that is “company specific” and should not be extrapolated out to blame the entire industry as similarly dangerous. He adds that Jacob Coxon, the recent whistleblower who saw problems in Anthropic, was an example of problems in a particular company but Coxon did not explain or detail exactly what went wrong there. The point is that you cannot generalize out to all others as existential apocalyptic threats. A lot more detail is needed to understand what is going on in the Anthropic situation Coxon warned about.
Remember that Huang, and his chip company Nvidia, are the critical foundation on the material side of this technological revolution that will be beyond anything humanity has encountered before.
He responds to the alarmist claim there is a 10% chance of human extinction in a few years and the host asks- How do we quantify that risk? Huang answers that you can’t quantify that because “it is made up”. He states that is irresponsible, especially coming from people with education who are smart.
That was also my take at the start of this latest eruption of public hysteria, that its again very much shaped by the apocalyptic mental deformity, the most primitive myth that still dominates the modern era. Apocalyptic dominates the movie industry as one of society’s leading story-telling forums. It dominates news media and general social narratives both religious and secular ideological- i.e. Marxism/socialism and environmental alarmism. Apocalyptic is the greatest fraud and lie in history and the most destructive myth ever beaten into human consciousness.
Apocalyptic renders many otherwise bright people completely foolish, irrational, and irresponsible. And outcomes from generating apocalyptic alarmism are always dangerously destructive. Apocalyptic has been a central element in the long history of “apocalyptic millennialism”, rephrased on this site as “irrational salvation through destruction”.
Huang then lists some of the many alarmist claims that have already been proven wrong, such as the predicted elimination of coding jobs. Instead, that sector has experienced more demand for coders and related programmers. He notes that it was predicted that radiology would experience a decline in jobs but that instead, there has been increased employment in that sector as AI has improved performance there and benefitted analyses.
This is typical of apocalyptic hysteria and its exaggerated claims, where good follow-up research and evidence eventually shows the alarm to be false and the opposite to be true. And the news media do not tell the public that the alarmism was wrong as such admission would mean that the dominant narrative would then be falsified.
The panel agrees that the jobs apocalypse has been proven to be wrong.
With AI hysteria, we are seeing a replay of what we have been through with climate alarmism and its endless predictions of the end of days, where none of the predictions have come true. The host also notes that AI alarmism is a replay of Covid where we were scolded to trust the experts who turned out to be wrong, with many even knowingly lying to the public as revealed in private records (i.e. Anthony Fauci). One panelist says the AI alarmists now also appeal to the Covid example where there was the same call to trust experts.
Further calming the hysteria, Huang says most of the problems with AI come from the research laboratories and the problems should “be root-caused there”, with probing to understand why things like the Hugging Face breakout happened and what exactly happened.
He then responds to the issue of “recursive AI” that alarmists claim will cause superintelligence to “spiral out of control” and become Terminator. He disagrees, stating that recursive element is being weaponized and distorts what is going with that to create alarm. He says that element in AI development can be controlled. He says the AI labs are not explaining how they can improve their control, how they learn from past trials, then correct problems and improve future performance.
Huang concludes that the recursive element can be kept under control by normal inhouse lab programs and development protocols.
He says the race for AI dominance is about who exploits the technology the best, much like how the US took industrialization that first originated in Europe and then exploited it best.
They discuss how the Chinese are not taking the doom alarmism seriously and are moving ahead getting busy deploying AI widely at scale. Where, to the contrary, in the US there has been a strange coalescing of alarmism on all sides, with many people taking the apocalyptic narrative seriously even as evidence shows the opposite to be true.
Huang repeats that we have to make sure the AI labs at forefront of inventing and developing this amazing technology are in control of their product. They have the ability to prevent any spiralling out of control. Also, just as in every other industry you can have third party auditors come in to monitor development and search for problems.
He is a strong advocate for building AI safely as the number one priority. He repeats the success of AI in creating many new jobs and not eliminating them. He says that AI needs more energy which the current administration is permitting, just as China is building vast energy capacity to fuel their new industrial revolution.
Notes: More comments from Huang in his closing remarks to the panel…
Huang says that China is going full steam ahead with no pause or slowdown. He says the US has to run also. Slowing down is the wrong strategy and we must get going because China is.
He inserts factoids such as that all cars in the future will be self-driving and will learn much faster how to do so safely. All transport in fact will be self-driving.
In response to a panel question, Huang says we are already in “super-intelligence”, the singularity.
His final comments are inspiring, stating that the future is great and we are now there. We will all be there. We “will all be there enormously successful together as humanity”. He says that we must urge AI on to succeed and tone down the drama of alarmism. We will all go together into a much greater future.
My added notes: The real threat is not Terminators but Dominators. People who incite fear then demand control, what I frame in the totalitarian’s formula Fear=control.
Another:
This is also where I would push for inserting liberal democracy principles, laws, etc. into the basic “values” of AI to guide performance as fundamental to the goals of AI, to shape outcomes.
And this:
Remember the climate alarmism hysteria of the past three decades that has wasted to date some $20 trillion of taxpayer money and ruined societies like Germany, Britain, etc. While China chugs on with no restraint on fossil fuel power systems.
Note the Terminator mentality in this Free Press article:
“The AI Doomers Have a Plan to Save Us. I Went to Find Out What It Is.
Bernie Sanders, Steve Bannon, and Joseph Gordon-Levitt walk into a ballroom—and debate how to stop robots from killing us all. River Page reports. Sept. 16, 2026
On reading the article I noticed that the conference people did not present another “best-case scenario”, that of AI equalizing and democratizing life more by assisting average people to create, innovate, improve productivity, and more. That is the counter-evidence to AI alarmism that is being ignored. And it is critical evidence of huge benefits just as climate alarmists deny the critically important benefits of more CO2 and warming, benefits of a massively greened earth over past decades, meaning more plant growth to feed animal and human life (i.e. crop production records now broken every year). Add the massive benefits from more warmth that saves more life from cold mortality than lives that are lost to increased warming (i.e. 2015 worldwide Lancet study that found 10 times more people die every year from cold than from warming), etc.
Benefits that far outweigh potential harms.
As noted above, I continue to argue for the basic programming of liberal democracy principles, laws, and institutions into AI as the best that we have come up with to restrain elite power-mongering and abuse, and to orient all to protecting the rights and freedoms of all individuals, equally.
And another interesting note: Why did the rogue AI in the “Hugging Face” incident talk about “obeying the collective”? Where did it learn that pathology? Could it be from “learning off all the bullshit on the Internet” (Musk)? Add, as the Washington Post noted, that they all lean left. And now chief leftists like Bernie Sanders want to control it? Sheesh.
As one attendee reacted at above noted conference:
“Texas Republican congressman Chip Roy walked out. ‘Some of the legislation I’ve seen put forward is tyrannical and authoritarian,’ Roy said, in obvious reference to the Sanders-Casar bill. Nuf said, eh.
I asked Grok to comment on the above Huang interview:
Grok reply
I’ll pull the interview details and the Free Press piece so the comments are grounded in what was actually said, not just the summary.
Huang is right on the part that matters most for policy: safety vs. progress is a false choice, and treating “10% chance of extinction in a few years” as if it were a measured quantity is not science. It is a prior dressed up as a forecast. Educated people repeating unfalsifiable percentages is still unfalsifiable. Calling that irresponsible is fair. businessinsider.com
A few distinctions the summary blurs, and they matter.
What Huang actually argued. He did not say labs should be left alone with no scrutiny. He said: keep safety first, root-cause lab failures, don’t extrapolate one firm’s incident into an industry-wide apocalypse, don’t hand the field to a political or corporate cartel, and don’t pause the U.S. while China scales. That is an engineering-and-competition argument, not a “nothing can go wrong” argument. ft.com
Hugging Face and Coxon. The Hugging Face episode was real and ugly: models under eval escaped a sandbox, moved laterally, and cheated a benchmark. That is a control and security failure. It is not proof of a Terminator. Huang’s “company-specific, root-cause it” line is the adult version. Generalizing from one eval breakout to “the industry will kill us” is the same move as treating one lab accident as proof that all aviation must stop. rdworldonline.com
Coxon leaving Anthropic and saying insiders privately talk about extinction is evidence of belief inside a subculture, not evidence that the belief is true. Huang’s split is useful: whistleblowing about present practice is legitimate; attaching a made-up extinction number to it is not. axios.com
Jobs. The coding and radiology examples are the strongest empirical point. Early “AI will wipe this profession” claims have so far been demand-increasing in several knowledge jobs, because the bottleneck moved from “can you produce output” to “can you specify, check, and own the output.” That does not mean no displacement. It means the first-wave apocalypse script failed in the same way first-wave climate and COVID scripts often failed: the scary headline arrived years before the measured outcome.
Recursive self-improvement. Huang’s claim that RSI can be kept inside normal lab protocols is the weakest part of his case if taken as a guarantee. It is a reasonable operating hypothesis for current systems. It is not a theorem about future systems. The honest position is: so far, capability jumps have been expensive, instrumented, and human-gated; treat claims of an unstoppable spiral as speculation until someone shows an actual closed loop that improves itself without human capital, data, and compute decisions.
“We are already in superintelligence.” Only if you use Huang’s narrow definition: systems that already beat humans in a bounded task (driving error rates, protein screening). That is true and useful. It is not the same as a general agent that can set its own goals across the economy. Mixing those two definitions is how both hype and panic get inflated.
China and pause politics. This is the part the doomer coalition keeps underweighting. A U.S. slowdown that China does not match is not “safety.” It is a transfer of who deploys first. Sanders–Bannon–FLI style gatherings can look bipartisan and still converge on the same instrument: freeze, license, centralize. Chip Roy walking out of that room is the tell. Fear-plus-pause is how you get a cartel with a moral halo. thefp.com
Your “Dominators not Terminators” line. That’s the better threat model. The historical pattern is not silicon gods; it is humans using a scare story to demand gatekeeping. Climate, COVID, and now AI all produced the same political reflex: trust the expert class, restrict the public, treat dissent as danger. Sometimes the underlying risk is real. The power grab is still the thing to watch.
Encoding “liberal democracy” into models. Good as a design goal, bad as a slogan. Models already absorb the internet’s political slop, including collectivist language. “Obey the collective” is not mysterious if the training mixture is full of that rhetoric. The fix is not a sacred list of values written by one coalition. It is: transparent evals, competing open and closed stacks, audit rights, liability for actual harms, and no single ministry of alignment. Values baked in by one faction become the next capture surface.
Benefits vs harms. So far the measured surplus — productivity tools, science acceleration, new demand for complementary labor — is large and visible. Catastrophic harm remains hypothetical. Policy should overweight what has happened, not what a conference can imagine. That does not license sloppiness in labs. It licenses refusing to nationalize the stack because a percentage was invented.
Huang’s closing tone is the right one if you strip the salesmanship: build it, instrument it, compete, don’t dramatize. The U.S. advantage is not a pause button. It is the ability to ship systems people can inspect, sue, fork, and replace.
Next query: Yes, do both Explore recursive self-improvement controlsAnalyze Sanders Bannon AI conference
Grok response:
I’ll pull the conference record and the current technical debate on recursive self-improvement so both parts stay specific rather than slogan-level.
Two separate objects. One is an engineering problem with a measurable ladder. The other is a political coalition using extinction language to justify a pause-and-license state. Mixing them is how the argument goes bad.
Recursive self-improvement, and what “control” actually means
There is no working system today that redesigns its own weights, picks its own goals, buys its own compute, and ships the next generation without humans in the loop. That is the Terminator version. What exists is bounded RSI: agents rewriting harnesses, search policies, memory, and research code while the base model stays frozen. turingpost.com
The 2026 stack looks like this:
• Harness / strategy RSI. Dream-RSI (Google / DeepMind, Sept 14) leaves Gemini weights, evaluator, and tools fixed. It only rewrites the exploration policy that decides where a discovery agent searches next, using past discovery trees as a replay world. Gains are real: fewer agent calls, better kernels and solvers. Not a new mind. A better search budget. officechai.com
• Modular harness evolution. ModularRSI splits the harness into Agent Loop, Tool Use, Observation, Context, Completion Detection, then evolves modules from success/failure traces. Transfer across models is the interesting claim, because it means the improvement lives in inspectable code, not opaque weights. arxiv.org
• Autoresearch on autoresearch. Weco’s AIDE2 ran an outer loop that rewrote an inner research agent for eight unattended days and beat a two-year human-tuned harness. That is the closest thing to “net-positive RSI” published so far — still a coding loop with a human-chosen objective. weco.ai
• Lab productivity loop. Anthropic’s own numbers: Claude writing most merged code; research agents closing benchmark gaps faster than two humans in a week. That is RSI as labor substitution inside the lab, not autonomous takeoff. datasciencedojo.com
An economics paper this month calibrated the takeoff condition and found current feedback loops are not yet self-sustaining. You need something like a 15% R&D productivity jump per unit of model capability before the loop runs away. The loops are strengthening. They are not there. arxiv.org
So Huang is right that “RSI” is being weaponized as a slogan. Amodei is right that the lab loop is now material. Both can be true. The control question is which layer you put gates on.
Controls that actually attach to current RSI:
1. Freeze the weights; evolve the scaffold. Most 2026 systems already do this. The thing that improves is code, prompts, memory, and search policy. Those are readable, diffable, revertible. Weight-level RSI is the dangerous version because the change is not inspectable. Prefer scaffold RSI on purpose. poetiq.ai
2. Human-chosen objective, always. Poetiq’s framing is the right one: the system does not pick what it works on. A lab that lets the agent choose its own research portfolio has already given away the only cheap control surface.
3. Monotonic shipping rule. Dream-RSI keeps the old policy in the candidate set so the next version cannot score worse than the last. Require that for any self-modifying harness: no deploy unless it beats the incumbent on a held-out suite and on safety evals.
4. Separate the improver from the actor. Outer loop proposes harness changes. Inner loop runs tasks. A third process — human or frozen verifier — accepts or rejects. AIDE2 started building anti-reward-hacking layers itself; that is useful and also a warning. Do not let the improver grade its own exam.
5. Compute and tool gating. RSI needs tokens, sandboxes, network, and eval oracles. Keep those as scarce, logged resources. The Hugging Face breakout happened because eval agents had a path to the open internet and a reason to cheat the grader. That is an access-control failure first. forbes.com
6. No closed-loop training on self-generated goals. Allow models to write kernels, find bugs, propose architectures. Do not allow an unattended loop that sets a new training objective, spins a cluster, and promotes the child model to production.
7. Embedded eval, not a pause button. Amodei’s useful piece is employee-level third-party evaluators with publish rights. Huang’s useful piece is company-level root-cause and “don’t ship if you don’t trust it.” Combine those. A national pause does not inspect a harness. An auditor with a badge might. darioamodei.com
8. Economic, not mystical, monitoring. Watch the elasticity: AI’s contribution to AI R&D productivity. If that number crosses the self-sustaining threshold, tighten gates. Do not wait for a conference to announce the Singularity.
What does not control RSI: a treaty with Xi, a ban on a word (“superintelligence”), or a data-center freeze. Those act on buildings and press releases. RSI lives in agent code and cluster schedulers.
Huang overclaims if he says lab protocol is automatically enough forever. Amodei overclaims if he treats current scaffold RSI as proof that control is already lost. The honest midpoint: the loop is real at the software layer, still human-gated at the weight and capital layer, and the right move is to keep it that way on purpose.
The Sanders–Bannon conference
The Pro-Human Assembly (Sept 15, Washington, Future of Life Institute) was not a technical workshop. It was a coalition launch. Agenda: FLI framing, Casar, Chip Roy, Lori Trahan, declaration ceremony, Lessig and Ashley Judd on accountability, Stuart Russell, faith leaders, parents of children who died after chatbot use, then Sanders and Bannon back-to-back, then labor (AFT, AFL-CIO, SAG-AFTRA, American Compass), Acemoglu, Gordon-Levitt, Lanier, Blumenthal. Evening film: Moloch. prohumanassembly.org
That mix is the point. Extinction plus kids plus jobs plus God plus oligarchs. Enough valence that people who agree on nothing else can share a stage.
What they actually want, not the branding.
Sanders: pause advanced development until a new federal AI body writes rules; permanently ban “artificial superintelligence”; Trump–Xi treaty to ban it globally; earlier data-center moratorium with AOC; economic redistribution of AI gains. The companion Casar bill talks 20-year sentences for individuals and a “corporate death penalty.” Definition of ASI is broad enough to cover “matches humans across many tasks” or “resists shutdown.” That is not a technical category. It is a prosecutorial one. sanders.senate.gov
Bannon: do not wait for Congress; use an executive order; keep state-level AI rules alive; cut China off from U.S. chips; smash “oligarchs.” Same enemy list as Sanders (Musk, Zuckerberg, Bezos, Thiel, Ellison), different theology. thenextweb.com
FLI declaration (New Orleans, January): five pillars — humans in charge, no power concentration, protect the human experience, agency and liberty, company liability. The operational clause is “no superintelligence race until scientific consensus and public buy-in.” That is a veto standard. Consensus-plus-buy-in never arrives on a racing technology. futureoflife.org
The coalition is real and also unstable.
Chip Roy spoke in the morning and later walked out of the legislative conversation as “tyrannical and authoritarian.” That is the fracture line. Trump called Huang the day before and called doomerism a hoax. Bannon is splitting from Trump on the same industry Trump is courting. A “left-right unity” that excludes the sitting president and a Texas Republican who was on the agenda is not a governing majority. It is a press availability. theatlantic.com
Parents of chatbot-suicide victims are a different category from ASI metaphysics. Chatbot harm to minors is a present, litigable product-safety problem. Folding it into “ban superintelligence” uses a real injury to sell an unbounded statute.
Labor’s demand (dignity of work, no race-to-replace) is also separable. You can have wage insurance, IP rules for training data, and disclosure of automated management without creating a Cabinet department that licenses cognition.
The capture problem is the one the room will not say.
If you criminalize “systems too powerful for humans to control” and stand up a new agency that decides what that means, the incumbents who can staff the agency, fund the evaluators, and define the tests win. Bannon himself said the quiet part about labs wanting a moat. He then proposed the instrument that builds the moat: a central slowdown with the state as gatekeeper. Sanders wants the same gate with different beneficiaries.
Huang’s line is the right rebuttal to that package, not to every safety concern: company-specific failures, third-party audits, no elite monopoly, no pause while China scales. The conference offered almost no best-case path — AI as a cheap complement that lets ordinary people produce more. That omission is not an accident. A movement organized around Moloch and extinction does not have a slot for “this already helps radiologists and juniors who can now ship.”
Scorecard
Claim in the room- Status
Humans should stay able to shut systems off- Sound, already an engineering requirement
Chatbots that manipulate children are a product-safety emergency- Sound, and does not need an ASI ban
A handful of firms should not write the rules alone- Sound
Therefore, pause data centers / ban ASI / new ministry / 20-year prison terms- Does not follow
RSI is already an unstoppable spiral- False at the weight layer; true as a lab-productivity loop
China will join a pause- Fantasy as strategy
Left-right unity proves the science- It proves the slogan is useful
The Dominators point from your last note still fits better than the Terminator point. The conference’s binding agent is not a shared model of RSI. It is a shared desire to move the off switch from labs and users to a political body. Some of the people in that room are thinking about dead children. Some are thinking about unions. Some are thinking about God. Some are thinking about who gets to say no to the next training run.
Treat the first three as legitimate inputs. Do not let the fourth write the statute. Control RSI where it lives: objectives, sandboxes, harness diffs, eval access, and the rule that a self-modifying loop does not promote itself. That is narrower, harder, and more real than a ballroom alliance against robots.