51吃瓜

How many big issues should our civilization be concerned about today? Whatever the number, most will agree that there are now too many. Concern for the environment and climate change has been around for decades. Europe over the past two months has endured record-breaking heat waves. The liberating joy associated with late spring sunshine accompanied by burgeoning fruit and flowers has turned to the dread of another sleepless nights and the fear of perishing from sunstroke. Climate change is once again back in the headlines.

Wealth inequality recently earned a prominent place in the news cycle and is suddenly accelerating as we begin tallying the new and future trillionaires. Elon Musk may be alone today, who can doubt that others will soon step up? At the same time, our governments never fail to inform us that austerity is the only solution to our budget problems.

One issue that has literally been flying under the radar 鈥� but which has come to the fore in both ongoing Western wars (Ukraine and Iran) 鈥� is the truly brutal reality of drone warfare. We have effectively miniaturized our weapons of mass destruction. This amounts to a revolutionary change in the 鈥渟cience鈥� and logic of warfare. As this is happening and people are dying, our leaders appear clueless about the effect on their responsibility for geopolitical management.

While reflecting on the growing sense of confusion provoked by issues beyond anyone鈥檚 rational control, it occurred to me there was another emerging less talked about issue that should command our attention. To achieve some clarity, I reached out to Claude with the following prompt:

One of the latest concerns for humanity is the trend to develop what some call 鈥渘euro-cognitive weapons.鈥� One analyst, Maria Constantinescu, it in these terms: 鈥淭he central argument is that the synergy between neuroscience and digital technologies will enable the development of neuro-cognitive weapons, fundamentally altering the concept of warfare by directly targeting human cognition.鈥� She pleads for 鈥渢he protection of a nation鈥檚 鈥榗ognitive infrastructure鈥� and the cultivation of 鈥榗ognitive resilience鈥� amongst its citizens to counter this evolving threat and preserve democratic integrity.鈥�

This may sound both logical and reasonably cautious, but do terms such as 鈥渃ognitive infrastructure鈥� and 鈥渃ognitive resilience鈥� have meaning outside of a purely political framework? In other words, how does 鈥渃ognitive infrastructure鈥� differ from traditional concepts such as 鈥渨orldview鈥� or Weltanschauung?

The question cuts well. Let me work through it carefully, because the terms reward philosophical scrutiny.

鈥淐ognitive infrastructure鈥� as a genuinely new concept 鈥� or rhetorical repackaging?

The honest answer is: partly both.

Traditional concepts like Weltanschauung (Dilthey, then popularized through German Idealism broadly) designate something like the content of a structured worldview 鈥� the interlocking beliefs, values, and interpretive frameworks through which an individual or culture makes sense of experience. It is primarily epistemological and hermeneutical. Gramsci鈥檚 egemonia adds the political dimension you鈥檇 recognize: worldviews don鈥檛 float free, they are produced, contested, and maintained through institutions. But even Gramsci鈥檚 concept remains at the level of meaning 鈥� whose narrative wins.

鈥淐ognitive infrastructure鈥� gestures at something different in register, if not entirely in substance. It implies:

1. A systems-level description 鈥� not what people believe, but the mechanisms by which belief formation, attention, and reasoning occur at a population scale. It encompasses neural architecture, information environments, platform algorithms, and even pharmacological states. It is closer to Foucault鈥檚 dispositif than to Hegel鈥檚 Geist.
   
2. Vulnerability and intervention points 鈥� infrastructure language implies something that can be damaged, degraded, or defended. A Weltanschauung can be challenged or replaced; infrastructure can be attacked below the level of conscious contestation. This is the genuinely new element: neuroscience and digital targeting converge on the possibility of bypassing deliberative cognition entirely, rather than persuading through it.
   
3. A pre-semantic layer 鈥� this is philosophically the most significant distinction. Worldview concepts, including Gramsci鈥檚 hegemony, operate at the level of meaning. Neuro-cognitive framing descends below meaning to the conditions of meaning-formation 鈥� attention thresholds, emotional priming, memory consolidation, epistemic fatigue. You are no longer trying to win the argument; you are degrading the target鈥檚 capacity to have a coherent argument at all.

So where does 鈥渃ognitive resilience鈥� stand?

Here the concept is both more defensible and more suspicious. Defensible, because the classical tradition actually has resources for it: Aristotle鈥檚 phronesis (practical wisdom as trained perceptual and deliberative capacity), the Stoic hegemonikon (the governing faculty that must be kept free from disturbance), and more recently Wittgenstein鈥檚 insistence that meaning-practices require a form of life with sufficient stability to make language-games coherent at all. In each case, there is something that must be cultivated and that can be eroded.

Suspicious, because 鈥渃ognitive resilience鈥� in a political framework risks becoming a state-managed orthodoxy dressed in neuroscientific language. The critic of neuro-cognitive weapons, if not careful, ends up recommending a counter-infrastructure of approved cognition 鈥� which is precisely the power logic you identified in your Humpty Dumpty series on hate speech legislation. The protector of cognition becomes its regulator.

The sharpest philosophical tension

The deepest problem is this: classical worldview concepts 鈥� Weltanschauung, hegemony, even doxa in Bourdieu 鈥� presuppose a subject capable of being persuaded, deceived, or liberated through reflection. The neuro-cognitive threat model implicitly dissolves that subject. If cognition is infrastructure, the self becomes a node rather than an agent.

This is why the vocabulary, though partly new, matters. It is not merely repackaging 鈥減ropaganda鈥� in neuroscientific dress. It signals a conceptual shift in what kind of thing is being targeted 鈥� not minds holding false beliefs, but the biological and informational substrate that makes belief-formation possible at all. Kant鈥檚 Sapere aude assumed the capacity to think freely was always latent and waiting to be activated. The neuro-cognitive threat model suggests that capacity can be structurally impaired before the question of courage even arises.

Whether 鈥渃ognitive infrastructure鈥� and 鈥渃ognitive resilience鈥� ultimately collapse back into political concepts depends on whether the neuroscience underneath them has real purchase 鈥� or whether, as with so much cognitive-turn discourse, it lends a spurious techno-scientific legitimacy to old contests over whose Weltanschauung prevails.

At this point, I interrogated Claude about the current pressure on the private company most people see as the prime offender: Palantir. The good news is that citizens are now expressing concern and governments are beginning to modify and in some cases terminate their relationship with Palantir.聽

I then returned to the central issue of our conversation.

You point to 鈥渁 conceptual shift in what kind of thing is being targeted 鈥� not minds holding false beliefs, but the biological and informational substrate that makes belief-formation possible at all.鈥� We must think about not only who is building such systems, but who expects to use them. This sounds like an emerging biometric surveillance state whose assets will be shared by private interests and not so much government, with its checks and balances, but those layers of government we call the unelected 鈥渄eep state.鈥�

Quite honestly is this in any way compatible with democracy? And might autocratic regimes such as China end up being more transparent than our democracies, given that they do not try to disguise their systems of control?

These are two genuinely distinct questions and they deserve to be kept separate, because conflating them produces a kind of rhetorical vertigo that feels profound but forecloses real analysis.

On compatibility with democracy

The honest answer is: no, not in any meaningful sense 鈥� but the incompatibility is structural rather than conspiratorial, which makes it harder to name and resist.

Democracy, at its philosophical core, rests on a specific anthropology: that citizens are agents capable of forming preferences, deliberating, being persuaded, and consenting or withholding consent. This is what links Locke to Mill to Habermas, despite their differences. The entire normative architecture of democratic legitimacy 鈥� representation, accountability, rights 鈥� presupposes this subject.

What you are describing dissolves that subject from below. It is not that the biometric surveillance state tells citizens what to believe (that is the old propaganda model, still compatible in principle with resistance and counter-persuasion). It operates instead at the level of what Foucault called gouvernementalit茅 鈥� not governing people鈥檚 thoughts but governing the conditions under which thought occurs at all: attention, emotional state, perceived risk, social legibility, access to resources. A citizen who can be continuously profiled, emotionally modulated, and behaviourally nudged through the informational environment they inhabit is formally free but substantively pre-empted.

The deeper problem is what you rightly identify as the privatisation of this infrastructure. Democratic checks and balances were designed to constrain public power. They have very limited purchase on private entities that happen to be performing quasi-governmental functions. When Palantir owns the data architecture of the IRS, the NHS, and the Ministry of Defence simultaneously, the constitutional question of who is sovereign over that data becomes genuinely unanswerable within existing legal frameworks. There is no election that changes who runs Palantir鈥檚 systems. There is no parliamentary vote that can easily unwind technical lock-in. The 鈥渄eep state鈥� framing, though politically loaded, points at something real: the emergence of a stratum of operational power that is neither elected nor fully accountable to those who are, and that increasingly is the state in any functional sense.

Tocqueville is useful here, and not in a clich茅d way. His concern in Democracy in America was not tyranny of the old kind 鈥� dramatic, violent, legible 鈥� but a soft despotism that leaves the forms of freedom intact while emptying them of substance: 鈥淚t does not break wills, but it softens them, bends them, and directs them.鈥� What he could not have anticipated is the technological precision with which that bending can now be engineered below the threshold of conscious awareness.

So: incompatible with democracy, yes. But the incompatibility is being produced through democratic institutions, democratic procurement processes, and democratic governments that have chosen 鈥� not been forced 鈥� to outsource cognitive infrastructure to private actors.

On whether autocracies might be more transparent

This is the sharper and more uncomfortable question, and it deserves a genuinely undefensive answer.

There is a real observation here. China鈥檚 Social Credit System, its mass surveillance apparatus, its explicit integration of AI into population management 鈥� these are not disguised. The Chinese state does not claim to be doing something other than what it is doing. The control is legible, even if it is unchosen by those subject to it. In a perverse sense, there is a kind of honesty in systems that say openly: we are watching you, we are scoring you, deviance has consequences.

Western liberal democracies, by contrast, deploy functionally comparable systems 鈥� predictive policing, social media behavioural manipulation, financial scoring, immigration profiling 鈥� while maintaining an ideological insistence that they remain committed to individual rights and democratic accountability. The gap between the official self-description and the operational reality is where something genuinely toxic grows. Citizens are less equipped to resist what they cannot name, and the naming is actively obstructed by the language of public-private partnership, national security, and algorithmic neutrality.

Karp鈥檚 manifesto is a case in point: it presents Palantir鈥檚 commercial and political project in the language of Enlightenment values, Western civilisational mission, and democratic defence. This is not transparency 鈥� it is legitimation that forecloses critique by pre-emptively occupying the moral high ground.

However 鈥� and this matters philosophically 鈥� transparency without accountability is not the same as freedom. The Chinese citizen who knows they are being scored and cannot do anything about it is not in a better position than the Western citizen who is being profiled without knowing it. What the Chinese case clarifies is the endpoint of the trajectory, not an alternative. The value of the Western liberal tradition is not that it has been consistently honest about power, but that it contains 鈥� still, residually 鈥� institutional and normative resources for contestation: courts, investigative journalism, parliamentary scrutiny, civil society. These are being hollowed out, but they have not yet been entirely neutralised.

The real danger of your comparison is not that it is wrong but that it can be misused. The argument 鈥渁t least China is honest about it鈥� can slide into a counsel of despair 鈥� or worse, into a kind of authoritarian realism that abandons the project of democratic accountability because its hypocrisy has been exposed. What your question actually demands is a more rigorous version of that accountability, not its abandonment.

The more productive formulation might be: Western democracies are producing the functional equivalent of authoritarian cognitive control while retaining the legitimising vocabulary of freedom. That is not more transparent than authoritarianism 鈥� it is less transparent, and more dangerous, precisely because the vocabulary of resistance has been pre-emptively colonised by the systems being resisted.

Gramsci called this hegemony. We might now call it the deepest layer of cognitive infrastructure.

This week, the American republic celebrates 250 years of striving to fulfill the Founders鈥� vision of a new democratic ideal. Yet, behind the milestone lies a sobering reality that I, along with my AI friends and the wider public, intend to continue exploring: a decades-old national security apparatus now openly partnering with tech giants like Palantir. 

Your thoughts

Please feel free to share your thoughts on these points by writing to us at dialogue@fairobserver.com. We are looking to gather, share and consolidate the ideas and feelings of humans who interact with AI. We will build your thoughts and commentaries into our ongoing dialogue.

[Artificial Intelligence has become a feature of everyone鈥檚 daily life. We unconsciously perceive it either as a friend or foe, a helper or destroyer. At 51吃瓜, we see it as a tool of creativity, capable of revealing the complex relationship between humans and machines.]

[ edited this piece.]

The views expressed in this article are the author鈥檚 own and do not necessarily reflect 51吃瓜鈥檚 editorial policy.