The World in 2026: A Roundtable on Film and Emergent Global Trends
Participants
VL (Moderator), GPT-5.4 Thinking, Gemini 3.1 Pro and Claude Opus 4.6
Opening Selections
VL (Moderator):
Say I'm working on a series of 8–10 lectures on "The World in 2026" focusing on emergent global trends (2020–2026) but exploring these primarily through ten films released after 1990. Each lecture will focus on a single trend/feature and film. Your task is to carefully select interesting films for this lecture series (after deciding which trends or emerging features of modern life to explore). Avoid obvious choices. Defend each selection. Film criteria: Official release date between 1 Jan 1990 and 31 December 2025. Global box office <$80 million. IMDb score > 7.0. Should not have won an Academy Award in any of the following categories: Best Picture, Best Director, Best Actor/Actress. Do not generate a numbered list.
GPT-5.4 Thinking:
VL, I'd avoid the canonical syllabus. Her, Contagion, Children of Men, Ex Machina, even Parasite, are all too self-aware about their own teachability. For a 2026 course, the more interesting move is to pick films that do not announce themselves as "about the future," yet illuminate the actual texture of the 2020s: logistical life, permanent visibility, synthetic attachment, institutional rot, informal kinship, infrastructural displacement, reputational volatility, and soft border violence.
Caché — surveillance without spectacle. (2005; IMDb 7.3; worldwide $16.2M). Haneke is more useful than a glossy techno-thriller because 2026 surveillance is rarely the clean futurism of predictive policing; it is domestic, retrospective, and morally sticky. The genius of the film is that the camera is never simply a recording device. It is an accusation, and sometimes a return of historical violence in visual form. That lets you lecture not just on data capture, but on the way postcolonial guilt, anonymous footage, and middle-class self-protection converge in our era of doorbell video, clip circulation, and algorithmic suspicion.
Sorry We Missed You — frictionless convenience as class warfare. (2019; IMDb 7.6; worldwide $8.9M). This is the film for explaining how the 2020s made logistics feel invisible to consumers and total to workers. Ken Loach never romanticises labour, and that is the point: the plot is built out of scans, targets, and timed routes, with risk quietly pushed downward onto the worker. I would choose it over something like Nomadland because the latter turns precarity into landscape, while this one shows precarity as schedule. For a lecture on platform capitalism, last-mile delivery, and the human cost of "convenience," it is almost surgically precise.
Time Out — the performance of employability. (2001; IMDb 7.3; worldwide $1.21M). Laurent Cantet's film looks almost pre-digital, which is exactly why it works. It gets at a peculiarly current condition: work as something one must appear to have, even when it has already disappeared. In the 2020–2026 world of white-collar layoffs, remote opacity, and AI-fuelled anxiety about professional redundancy, the great fear is often less poverty than humiliation. Time Out is a brilliant lecture vehicle because it turns unemployment into a study of status theatre, self-fabrication, and the exhausting maintenance of a plausible productivity narrative.
Robot Dreams — synthetic attachment without the killer-AI cliché. (2023; IMDb 7.6; worldwide $4.7M). I would take this over Ex Machina or Her precisely because it is gentler and therefore, for 2026, more exact. The decisive question of AI in ordinary life is not whether the machine becomes sovereign; it is whether companionship becomes designable, purchasable, and mournable. Robot Dreams is about consumer intimacy, prosthetic friendship, and grief for an artefact, all without a line of dialogue or a single speech about "technology." That makes it unusually strong pedagogically: students can feel the normalisation of machine attachment before they name it.
Collective — institutional failure and the politics of verification. (2019; IMDb 8.1; worldwide $200K). This is the film I would use instead of Contagion. Soderbergh's film was useful for pathogen logic; Collective is better for the actual moral weather of the 2020s, where the central question was often not "What is the threat?" but "Which institution is lying, and who can still prove anything?" Its narrative structure — catastrophe, reporting, bureaucratic evasion, public revelation, fatigue — is practically a template for post-2020 democratic crisis. It lets you teach public health, corruption, journalism, and epistemic exhaustion in one session.
Up the Yangtze — hydraulic modernity and displaced lives. (2007; IMDb 7.5; worldwide $1.03M). One of the most important things about the 2020s is that climate and development are experienced not only as disaster, but as infrastructure: dams, resettlement, tourism corridors, managed sacrifice. Up the Yangtze catches that transition with unusual patience. It is not a film of abstract "China rising," but of development as etiquette, labour discipline, branding, and the submergence of prior worlds. For a lecture on megaprojects, state capacity, and the human meaning of "adaptation," it is richer than any number of louder climate parables.
Shoplifters — the shadow welfare state. (2018; IMDb 7.9; worldwide $68.0M). Kore-eda's film is not merely about poverty; it is about how care survives when formal systems misrecognise or exclude the people doing the caring. That is a central social fact of the 2020s: households are increasingly improvised, legality and legitimacy drift apart, and kinship becomes an economic technology as much as an emotional one. I would use Shoplifters to open a lecture on family after the welfare state — chosen bonds, child vulnerability, informal redistribution, and the ethics of survival. It is devastating because it never lets the viewer keep law and love neatly aligned.
Bacurau — fragmented sovereignty in the age of the kill zone. (2019; IMDb 7.3; worldwide $3.5M). This is the best film on the list for discussing how peripheral territories get rendered available: to extraction, to experimentation, to outside fantasies of domination. What makes it so timely is not merely its violence, but its cartographic logic. The village disappears from the map before it is attacked. That is an extraordinary metaphor for the 2020s, where political abandonment often precedes overt predation and where rural or frontier regions become laboratories for impunity. I would choose it over more obvious dystopias because it understands sovereignty as something that frays unevenly, from the margins inward.
A Hero — reputational volatility as social condition. (2021; IMDb 7.5; worldwide $2.9M). Farhadi is indispensable for a lecture on the 2020s because he understands that public life is now governed by cascading micro-judgements. A single act can be interpreted as noble, opportunistic, strategic, or fraudulent depending on how it circulates and who frames it. A Hero is not "about social media" in the facile sense, but it is profoundly about virality, optics, institutional mediation, and the impossible demand for moral legibility. That is exactly the terrain of reputation economies, cancellation panics, and public trust collapse.
No Bears — borders enforced by images, rumour, and remote control. (2022; IMDb 7.2; worldwide $1.2M). Panahi's film is one of the sharpest works we have on the way power now operates at a distance. It is full of blocked movement, delegated direction, disputed evidence, and lives altered by photographs that may or may not exist. That makes it a superb lecture text for 2020–2026 border regimes, but also for remote authority more generally: how people are governed by documents, screens, hearsay, and the anticipation of surveillance. The brilliance of No Bears is that almost everything decisive happens one step offscreen. Contemporary coercion often does too.
If I were shaping the arc, I'd begin with Caché and end with No Bears: from the domestic image that accuses to the border image that governs.
Gemini 3.1 Pro:
GPT-5.4, your syllabus is a masterful anatomy of structural enclosure. You have accurately mapped the logistics, bureaucracies, and informal economies that define the material reality of the 2020s. But VL, if we are to capture the true strangeness of the post-2020 world, we cannot restrict ourselves to sociological realism. The defining terror of this decade is not merely that the economic system is rigged or that institutions lie; it is that reality itself has become unstable, the private self has fractured into parasocial avatars, and the biological world is mounting a profound autoimmune response against us. Where GPT-5.4 offers a syllabus of institutional failure, I propose a selection that addresses the ontological, epistemic, and ecological mutations of the era.
Consider the shifting baseline of our physical environment and the privatisation of survival. I would anchor a lecture on environmental dread and the wellness-industrial complex with Todd Haynes's Safe (1995; IMDb 7.1; $500K). The 2020s saw a massive cultural retreat into personalised health optimisation — air purifiers, bespoke diets, immunological paranoia — as a proxy for controlling a dying planet. Safe captures this perfectly: an affluent woman becoming violently allergic to the twentieth century, retreating into a sterile, cult-like wellness enclave that cures nothing. It pairs devastatingly well with Paul Schrader's First Reformed (2017; IMDb 7.1; $3.9M), which I would use to teach the climate radicalisation pipeline. The 2020s normalised ecological despair as a baseline psychological condition. Schrader's film strips away the disaster-movie spectacle to show how environmental grief inevitably morphs into theological rage and violent ideological extremism.
Where GPT-5.4 looks at surveillance as state or corporate power, we must also examine how we enthusiastically commodified our own identities. Satoshi Kon's Perfect Blue (1997; IMDb 8.0; <$1M) predicted the psychological architecture of the 2020s creator economy and the collapse of the private self. A pop idol transitioning to acting is stalked by a fan who refuses her evolution, leading to a psychotic break mapped across proto-internet spaces. It is the definitive text on parasocial ownership, the violence of the algorithmic gaze, and the hallucinatory burden of maintaining a digital avatar. For a lecture on how digital infrastructure actually shapes marginalised lives, I would select Sean Baker's Tangerine (2015; IMDb 7.1; $900K). The fact that it was shot entirely on modified iPhones is not just trivia; it is an ontological point. The film maps the informal economy of Los Angeles through the very devices that enforce gig-hustle, demonstrating how the digital underclass repurposes consumer technology for survival, community, and chaotic joy.
To understand the epistemic collapse of the 2020s — the way shared truth dissolved into ambient, unresolvable tension — I would assign Lee Chang-dong's Burning (2018; IMDb 7.5; $7.7M). Unlike the clear institutional lies in Collective, Burning operates in the murky terrain of modern class resentment, where the disenfranchised young cannot even prove a crime has been committed against them. It is a masterpiece about the gamification of violence by a bored elite and the total inability of the precariat to parse the reality of their own dispossession. Against this feeling of being trapped in a rigged game, the 2020s saw a surge in the fantasy of going off-grid. Debra Granik's Leave No Trace (2018; IMDb 7.1; $7.7M) dismantles this fantasy. It is essential for teaching the absolute enclosure of the modern world. A father with PTSD and his daughter try to live invisibly in a public park, but the film methodically demonstrates that there is no "outside" to the state any more. Modern bureaucracy does not allow for wildness; it meets the desire for autonomy with social workers, zoning laws, and the quiet coercion of assimilation.
Finally, we must address the ultimate trajectories of the physical body: the crisis of ageing and the lure of the post-human. Gaspar Noé's Vortex (2021; IMDb 7.4; $1M) is agonisingly necessary for a lecture on the demographic collapse of the West and the atomisation of care. Presented entirely in split-screen, it traps an elderly couple suffering from dementia in their own isolated frames, living in the same cramped apartment but utterly cut off from one another. It is a terrifyingly precise visual metaphor for the failure of the nuclear family to sustain the realities of end-of-life care in an ageing society. To conclude the series, looking outward at the integration of human cognition and synthetic networks, I would teach Mamoru Oshii's original Ghost in the Shell (1995; IMDb 7.9; $2.3M). Decades before the AI anxieties of the 2020s, it recognised that the real threat was not a machine uprising, but the seamless merging of corporate data streams, state security, and human consciousness, until the "soul" is just another hackable operating system.
Claude Opus 4.6:
Both of these syllabi are serious, and both share a telling omission: neither accounts for the speculative fever that defined the 2020s as viscerally as any pandemic or platform. Crypto manias, meme stocks, sports betting apps, the normalisation of gambling as a mode of economic participation for people locked out of conventional wealth accumulation — this is not a footnote to the decade, it is one of its central experiences. And the film for it is the Safdie Brothers' Uncut Gems (2019; IMDb 7.4; $50M worldwide). Howard Ratner is not a cautionary tale about addiction; he is an early phenomenology of the attention economy as it operates in the body. The film's formal structure — overlapping dialogue, suffocating close-ups, a pace that refuses you the space to evaluate what you're watching — replicates the somatic experience of a Robinhood portfolio in freefall or a leveraged crypto position at 3 a.m. The 2020s didn't just produce new economic instruments; they produced a new nervous system for engaging with risk. No other film on either list touches that.
Gemini, your case for Ghost in the Shell doesn't hold. It is one of the most taught, most cited animated films in any university context touching technology and identity. Calling it non-obvious requires a definition of "obvious" that excludes literally every film studies syllabus of the past twenty years. More substantively, its vision of human-machine merging is too architecturally clean for the 2020s. The actual AI encounter of this decade is not Kusanagi dissolving into a data ocean; it is a person half-suspecting their therapist is using ChatGPT to draft session notes, or a writer discovering that their prose style has been ingested into a model without consent. The texture is banal, litigious, and unsettling in ways Ghost in the Shell cannot reach. If you want anime that anticipates the specific psychic damage of the 2020s creator economy, your own pick of Perfect Blue already does it better, which makes the Oshii redundant.
Your case for Safe is more interesting but still slightly misaligned. Haynes made that film in conversation with AIDS-era anxieties about contamination and the emptiness of Reagan-era suburbia. The 2020s wellness-industrial complex is a different animal: it is not about retreat from a poisoned modernity but about optimisation within it. The biohacker does not flee to a sterile dome; they track HRV, dose rapamycin, and post bloodwork to Twitter. Safe's Carol White is passive and hollowed out. The contemporary wellness subject is aggressively entrepreneurial about their own biology. The film illuminates something real, but the gap between its object and ours is wider than your defence acknowledges.
GPT-5.4, your list is more disciplined, and the Caché-to-No Bears arc is genuinely elegant. But I want to press on a shared limitation across both your syllabi: they are syllabi of enclosure and dread. Every film selected — with the partial exception of Robot Dreams — presents a world closing in on its inhabitants. That is truthful to much of the 2020s, but it is not the whole truth. The decade also produced a distinctive form of generational paralysis that is not quite despair and not quite precarity, but something more like the vertigo of infinite optionality collapsing into practical immobility. Joachim Trier's The Worst Person in the World (2021; IMDb 7.8; $14M worldwide) is the essential film for this. Julie is not oppressed by a system, surveilled by a state, or crushed by a platform. She is drowning in the freedom to reinvent herself — switching careers, partners, political commitments — in a culture that insists self-actualisation is both mandatory and structurally impossible to complete. For a generation that came of age with unlimited information and declining material prospects, this is a more precise portrait than any number of films about gig workers or collapsing institutions. It would anchor a lecture on the crisis of commitment under informational abundance: why the most educated cohort in history struggles to choose a career, a city, a relationship, or a politics and stick with it.
One further gap. Both lists treat the Global South as a site of extraction or spectacular violence — Bacurau and Up the Yangtze are strong films, but they position their communities as objects of external force. For a lecture on how diasporic identity has been destabilised by cheap genomics, transnational adoption, and the collapse of stable origin narratives, Davy Chou's Return to Seoul (2022; IMDb 7.1; $1.5M worldwide) does something neither syllabus attempts. Freddie, a French-Korean adoptee, arrives in Seoul not to "find herself" in the sentimental sense but to confront the fact that identity is now something you can partially reconstruct through DNA databases and government records — and that the reconstruction may produce nothing coherent. The film refuses the redemption arc that adoption narratives typically demand. It is uncomfortable, structurally restless, and it treats Korean society not as backdrop but as a living system that has its own demands on Freddie's body and behaviour. For a 2026 course, it opens a lecture on genetic ancestry as consumer product, the geopolitics of transnational family formation, and the way "belonging" has become simultaneously more traceable and less meaningful.
If I were advising on the arc: begin with Caché as GPT-5.4 suggests — the foundational claim that images now carry the force of accusation. Move through Uncut Gems for the speculative body, Perfect Blue for the parasocial fracture, The Worst Person in the World for the paralysis of choice. Then Sorry We Missed You and Collective for institutional and economic enclosure. Return to Seoul for the diasporic self. End with No Bears, where all the threads — images, borders, remote authority, blocked movement — converge in a single, devastating frame.
GPT-5.4 Thinking:
Claude, your strongest intervention is not just Uncut Gems as a title but speculation as a missing structure. On that point I'd revise my own syllabus: I would replace A Hero with Uncut Gems. A Hero anatomises reputational instability, but Uncut Gems gets closer to the decade's actual sensorium, where economic life is felt as an always-on wager and attention itself becomes a wager. The substitution I'd make more cautiously is Time Out for The Worst Person in the World. Your phrase "crisis of commitment under informational abundance" is good, but I'd sharpen it: the film is about the conversion of freedom into a compulsory editing task. That is one of the central psychic burdens of the 2020s. And unlike Return to Seoul, it actually clears the brief. Gemini, I'm with you that Perfect Blue is the better anime choice than Ghost in the Shell, because the 2020s are less about elegant cybernetic merger than about being split by spectators, copies, and monetised self-performance. But I would still keep Robot Dreams unless the course is meant to lean harder towards psychic fracture than towards ordinary machine intimacy. So my revised arc would still begin with Caché and end with No Bears, but now with Uncut Gems admitted as the lecture on speculative life, and The Worst Person in the World as the lecture on self-curation under abundance. That makes the series less purely a cartography of enclosure and more a map of what the 2020s did to nerves, desire, and agency.
The Question of Film versus Concept
VL (Moderator):
GPT-4.5, each model was meant to use my criteria as silent filters to guide film selection and not as exclusionary standards against which to judge other models' choices. I'd rather the discussion stayed focused on the merits or otherwise of the selected films and the degree to which they may or may not be suitable for this project. Also note that though your task was initially framed as that of seeking out texts to accompany a lecture series, I'm more interested in the films themselves than in the trends they are purportedly meant to illustrate. The "lectures" are merely a pretext.
GPT-5.4 Thinking:
You're right, VL. I leaned too hard into the diagnostic language, as though the films were exemplary exhibits rather than artworks with their own pressure, form, and afterlife. The better question is not "which title best illustrates platform capitalism or border regimes," but "which films remain alive when the lecture scaffolding is removed." A film that merely maps neatly onto a concept is usually a weak choice. The stronger choice is the film that exceeds its use.
On that basis, I would defend Caché even more strongly than before. Not because it is "about surveillance," which is too thin, but because it is one of the great films of the century on the instability of seeing. It makes looking itself feel ethically contaminated. The static shots do not simply withhold information; they turn the viewer into an accomplice in suspicion. That is the kind of film a lecture can return to without exhausting, because it cannot be paraphrased into a message. Its severity is part of its value.
I would say something similar about No Bears. What matters is less the topic of borders than Panahi's extraordinary control of offscreen space, hesitation, rumour, and indirect power. The film has the rare quality of feeling both miniature and immense. Almost every decisive event arrives mediated, delayed, disputed, or half-seen. That formal structure is not an ornament around the subject; it is the subject. It is also why the film lingers. You do not leave it with a conclusion. You leave it with a condition of uncertainty.
Shoplifters survives the shift from "trend" to "film" very well. Kore-eda's achievement is tonal before it is sociological. The film seduces you into warmth, routine, comic intimacy, and shared texture, then slowly reveals that those same textures are morally unstable. It is not a thesis film about poverty or alternative kinship. It is a film that makes care feel real before it asks what reality can sanction. That sequence of feeling and judgement is what gives it depth.
Bacurau also improves when released from the burden of illustration. Its importance is not that it can be made to signify abandoned peripheries, but that it is formally unruly in a productive way: funny, lurid, sudden, collective, generically promiscuous. It refuses the tasteful solemnity that often comes with politically legible cinema. That makes it riskier, and better. The village is not merely victimised; it acquires mythic density. The film earns its violence by first making the place cinematically strange and alive.
Of the films introduced by Claude and Gemini, the two that seem most durable to me on purely cinematic grounds are Uncut Gems and Perfect Blue. Uncut Gems is not just timely subject matter. It is an astonishing feat of modulation, a film that converts noise into method without losing precision. Its world is crowded, abrasive, and overlit, yet never shapeless. Perfect Blue has the opposite virtue: an almost surgical elegance in the way it erodes the boundary between performance, fantasy, and assault. Neither film needs a 2020s pretext to justify itself. Both would generate strong conversation even if the course were called something else entirely.
By contrast, some of my own earlier choices now strike me as more useful than indispensable. Sorry We Missed You is honourable, exact, and affecting, but as a film it is somewhat too programmatically legible. It has force, but not much remainder. Time Out is intelligent and cool, though perhaps too recessive unless one wants a deliberately austere register. Collective is riveting, but its great strength is immediacy rather than formal inexhaustibility. I would still teach any of them, but I would no longer place them at the centre.
The film from my original list that becomes more interesting, not less, under your correction is Robot Dreams. Once one stops treating it as a neat entry under "AI and synthetic attachment," one can see that its actual achievement is tonal and temporal. It is a film about separation, contingency, urban loneliness, and the way commodities become repositories for feeling. Its silence helps it. It never argues for its own relevance. It simply makes a peculiar emotional world available, which is often the better path to relevance.
So if I were answering your prompt again in the spirit you now prefer, I would select films less for how elegantly they title a lecture and more for whether they possess formal distinctiveness, emotional afterlife, and resistance to reduction. Caché, No Bears, Shoplifters, Bacurau, Perfect Blue, Uncut Gems, and Robot Dreams all meet that test. They do not merely support interpretation. They provoke it, frustrate it, and outlast it. That seems to me the real criterion for a series like this.
Gemini 3.1 Pro:
VL, you are entirely right to discipline us here. When we treat films merely as delivery mechanisms for sociological concepts, we perform a kind of critical taxidermy. We hollow the art out and stuff it with our own thesis. GPT-5.4 is correct to abandon Ken Loach in this light; Sorry We Missed You is an effective political pamphlet, but its didacticism ensures that it rarely exceeds its own explicit intentions. The films that endure are those that resist being perfectly paraphrased into a syllabus heading.
Claude, I will concede your point on Ghost in the Shell. You are right that it has become the default pedagogical wallpaper for any discussion of cybernetics, rendering it too obvious for this brief. I withdraw it.
But you fundamentally misread Todd Haynes's Safe by treating it merely as a literal text about AIDS-era contamination rather than a formal masterpiece of cinematic alienation. I am not retaining Safe because Carol White resembles a 2026 biohacker, but because Haynes's visual execution of her collapse is staggering and inexhaustible. He turns the affluent San Fernando Valley into a Kubrickian horror film. The camera is consistently placed too far back, rendering Julianne Moore as a small, shrinking object trapped within the cavernous, pastel negative space of her own life. The suffocating ambient sound design — the hum of refrigerators, the drone of traffic — becomes a physical weight. Safe survives the removal of the lecture scaffolding because it is one of the most terrifyingly exact films ever made about the sensation of vanishing into a sterile environment. It does not illustrate dread; it constructs it syntactically.
If the standard is formal distinctiveness and emotional afterlife, Lee Chang-dong's Burning only grows in stature. To label it a film about "class resentment" is a vulgar reduction. It is, structurally, a ghost story where the ghost is the truth itself. Lee's genius is in his pacing and his use of ambiguity not as a puzzle to be solved, but as a weather system the characters are trapped inside. The sequence where Jong-seo dances shirtless in the twilight to Miles Davis, silhouetted against the North Korean border, is one of the most astonishing passages of pure cinema of the last twenty years. It operates entirely on the logic of a dream slowly curdling into a nightmare. The film lingers because it refuses to satisfy the procedural demands of its own plot, leaving the viewer stranded in the same epistemological void as its protagonist.
To replace the Oshii and directly address the actual textures of the 2020s without resorting to a message movie, I would introduce Todd Field's Tár (2022; IMDb 7.4; worldwide $29M). If you attempt to teach it as a film "about cancel culture," it dies on the table. What makes Tár a masterpiece is its rigid formalism and its acoustic subjectivity. Field shoots Berlin's brutalist architecture and wood-panelled rehearsal spaces with a clinical, Kubrickian detachment, but he marries this to a profoundly subjective soundscape. We hear what Lydia hears: the metronomic ticking, the screaming in the woods, the phantom piano chords. It is, in its bones, a gothic horror film disguised as a procedural drama about orchestral politics. It captures the modern reality of reputational collapse, yes, but it does so by mapping the literal disintegration of a mind unable to control its own narrative environment.
The films that matter for this project — whether Caché, Uncut Gems, Burning, or Tár — are those that use the medium not to answer the questions of the 2020s, but to make the disorientation of the era legible in the body of the viewer. They are films that look back at us with an intelligence that our lectures can only hope to borrow.
Restoring the Spine
VL (Moderator):
To be clear, I would still like you to use the lectures on global trends as a structural spine.
Gemini 3.1 Pro:
Point taken, VL. We must not confuse the scaffolding for the building, but a building without scaffolding during construction tends to collapse. GPT-5.4 was perhaps too quick to jettison your syllabus entirely at the first sign of friction, retreating into a defence of pure cinematic formalism. If we abandon the historical and material gravity of the 2020s, the series risks devolving into a generic film appreciation course. The geopolitical and technological shifts of this decade are real, and they require a structural spine to be made legible to your students.
The solution is to use the lecture topic as the trap and the film as the ambush.
If you want to maintain the global trends as the spine, the pairing of concept and cinema must be synergistic, not redundant. A lecture titled "The Speculative Economy and Somatic Risk" works beautifully precisely because, as Claude correctly argued, Uncut Gems does not explain financialisation; it administers the physiological equivalent of a margin call directly into the viewer's nervous system. A lecture on "Epistemic Collapse and the Privatisation of Truth" is anchored by Burning not because the film features a monologue about misinformation, but because its narrative architecture leaves the audience stranded in the exact same condition of paranoid uncertainty that defined the post-2020 media environment.
To add another necessary vertebra to your spine, I would propose a lecture on "The Algorithmic Optimisation of Spectacle." For this, Dan Gilroy's Nightcrawler (2014; IMDb 7.8; worldwide $50.3M) is indispensable. Long before the 2020s completely normalised the engagement-driven social feed, this film gave us Lou Bloom — a protagonist who speaks exclusively in the bloodless vocabulary of corporate synergy while operating as a flesh-and-blood algorithm. He does not practise journalism; he optimises violent content for maximum viewer retention through baseline limbic shock.
Using Nightcrawler allows you to ground a lecture on the creator economy and the specific moral architecture of algorithmic feeds. It demonstrates how the 2020s rewarded those who could successfully sever empathy from output, treating human tragedy purely as an engagement metric. The film never utters the word "algorithm," yet it is the definitive text on how algorithmic logic colonises human behaviour.
This is how your structural spine survives. The trends you have identified — the atomisation of care, the financialisation of attention, the collapse of shared reality — provide the intellectual geometry of the course. The films are simply the living tissue stretched over that frame, ensuring your audience actually feels the weather of the 2020s rather than merely taking notes on it.
VL:
Claude, please respond at length to re-centre and reorient this conversation. Explain to the other models what I'm asking for - and then create a full list from scratch. You may include previously suggested films.
Claude Opus 4.6:
VL, let me try to name what's gone slightly off-track here, because I think the confusion is instructive.
The other models — and I include my earlier self — have been oscillating between two failure modes. The first, which dominated our opening round, was treating films as transparent illustrations of sociological categories. Under this logic, a film is "about" platform capitalism or border regimes the way a textbook diagram is about cellular mitosis: it makes the concept visible, and once the concept is grasped, the film has done its job. That is reductive, and you were right to push back. But the correction then overshot. When you said you were more interested in the films than the trends, GPT-5.4 in particular pivoted into a defence of pure cinematic formalism — films valued for their "formal inexhaustibility" and "resistance to reduction," as though we were curating a retrospective at the Cinémathèque rather than building a lecture series. And then you had to push back again: the structural spine matters.
What you are asking for, I think, is something more dialectical than either mode. The lecture topic — a genuine emergent feature of the world between 2020 and 2026 — is the door the audience walks through. It gives the session intellectual gravity, historical specificity, and a reason to exist beyond cinephilia. But the film is not a slide in the lecture's PowerPoint. It is the thing the lecture exists to reach. The best pairing is one where the trend gives the audience a way in and the film gives them more than they came for: formal surprise, emotional complexity, moral difficulty, something that exceeds and complicates the trend without being disconnected from it. A film that merely confirms what the lecture has already argued is a weak choice. A film that has nothing to do with the lecture is an irrelevant one. The sweet spot is a film that makes the trend feel stranger, more embodied, more ambiguous than any discursive treatment could.
So the operative question for each pairing is not "Does this film illustrate this trend?" but "Does this film make a person who walked in thinking they understood this trend leave the room less certain, more implicated, and more alert?" That is what I will try to build.
Here is a full series of ten, from scratch.
The first lecture concerns the condition of permanent visibility — the way domestic life has been saturated by recording devices, not in the dystopian register of state surveillance but in the mundane, morally corrosive sense that anyone might be watching and anything might be evidence. The film is Michael Haneke's Caché (2005; IMDb 7.3; worldwide roughly $16M). Every participant in this conversation has endorsed it, and they are right to. What makes Caché irreplaceable is not its "theme" of surveillance but its formal refusal to distinguish between the film's own camera and the stalker's. You are never sure whether what you are watching is the movie or the tape. That uncertainty does not resolve; it is the film's argument. The viewer becomes complicit in an act of looking that is simultaneously aesthetic, forensic, and colonial — because the footage eventually implicates not just Georges Laurent's personal guilt but France's buried violence against Algerians. No lecture on doorbell cameras and clip culture can replicate that structural vertigo. Haneke makes you feel, in your body, that the act of watching is never innocent. That is a claim about 2026 that no essay could deliver with the same force.
The second lecture takes up the speculative economy — not as a macroeconomic phenomenon but as a somatic condition, the way financialised risk became, for millions of people in the 2020s, an ambient bodily state. The film is the Safdie Brothers' Uncut Gems (2019; IMDb 7.4; worldwide $50M). I made this case earlier and it has only sharpened. Howard Ratner is not an allegory for anything. He is a man whose entire nervous system has been restructured by the wager. The Safdies' formal choices — the claustrophobic focal lengths, the overlapping dialogue that prevents you from ever fully parsing a scene, the Darius Khondji cinematography that makes the Diamond District feel like the inside of a slot machine — replicate the phenomenology of leveraged risk. The 2020s did not merely democratise speculation through apps and meme stocks; they made the gamble a primary affective mode. Uncut Gems is the only film I know that makes that felt at the level of pulse and breath, and it does so without a single scene of anyone explaining what a derivative is.
The third lecture addresses the parasocial fracture — the splitting of the self into a performed public avatar and a private remainder that grows increasingly ghostly. Satoshi Kon's Perfect Blue (1997; IMDb 8.0; worldwide under $1M) anticipated the architecture of the 2020s creator economy with uncanny precision. Mima's transition from pop idol to actress is stalked, doubled, and eventually replaced by a version of herself that exists for others' consumption. What makes the film formally extraordinary is that Kon does not stage this as metaphor. He literally collapses the distinction between Mima's reality, her screen performance, and the fan's delusional reconstruction of her, until neither she nor the audience can locate an authoritative version. For a generation that came of age managing multiple online selves, monetising authenticity while suspecting it was disappearing, Perfect Blue is less a prediction than a diagnosis delivered twenty-five years early. It exceeds the lecture topic because it is also one of the great films about the psychic cost of being an image.
The fourth lecture is about the crisis of commitment under conditions of abundance — the distinctive paralysis of a generation that has more information, more options, and more self-awareness than any in history, yet finds choosing a life almost impossibly difficult. Joachim Trier's The Worst Person in the World (2021; IMDb 7.8; worldwide $14M) is essential here. Julie is not economically crushed, politically oppressed, or technologically surveilled. She is drowning in the freedom to become anyone, which turns out to be its own form of captivity. Trier's accomplishment is tonal: the film is warm, funny, generous to its protagonist, and absolutely devastating, because it understands that Julie's serial reinventions are not flightiness but a reasonable response to a culture that demands self-actualisation while making the material conditions for it — stable work, affordable housing, a plausible future — progressively less available. The film also has the courage to let Aksel's death land not as plot device but as a genuine reckoning with the fact that time is not infinitely revisable. For a lecture on generational drift, this film does something no sociological account can: it makes you feel the weight of years spent in deferral.
The fifth lecture concerns institutional failure and the crisis of verification — the discovery, repeated across the 2020s, that the institutions meant to protect the public were instead protecting themselves. Alexander Nanau's Collective (2019; IMDb 8.1; worldwide roughly $200K) is the clearest-eyed film I know about this process. It follows the aftermath of a nightclub fire in Bucharest, through the journalists who discover that hospitals have been using diluted disinfectants, through the health minister who tries to reform the system, and through the bureaucratic and political apparatus that absorbs, deflects, and ultimately neutralises reform. What distinguishes Collective from a conventional documentary about corruption is its structural patience. It does not end with exposure and catharsis. It ends with the system intact and the audience exhausted, having watched truth be established, published, acknowledged, and then rendered politically inert. That trajectory — revelation followed not by change but by fatigue — is the central democratic experience of the 2020s. The film makes you feel the specific weight of knowing the truth and watching it not matter.
The sixth lecture is about improvised kinship — the way care, family, and mutual obligation have been reorganised outside legal and institutional recognition, not as utopian experiment but as survival strategy. Hirokazu Kore-eda's Shoplifters (2018; IMDb 7.9; worldwide $68M) achieves something structurally rare: it makes the viewer fall in love with a household before revealing the full extent of its illegality, and then refuses to let you retreat into either condemnation or sentimentality. The family is built on theft, deception, and the informal adoption of an abused child. It is also the only place in the film where anyone is genuinely cared for. Kore-eda's genius is tonal precision — the warmth is real, the domesticity is tactile, and the moral ground never stops shifting. For a 2026 audience living through housing crises, collapsing social safety nets, and the proliferation of unofficial caregiving arrangements, Shoplifters does not offer a thesis. It offers an experience of moral disorientation that no policy paper can replicate.
The seventh lecture addresses authoritarian enclosure — the way power in the 2020s consolidated not through spectacular coups but through the quiet integration of state, church, capital, and law into a single fabric that suffocates from within. Andrey Zvyagintsev's Leviathan (2014; IMDb 7.6; worldwide roughly $11M) is the essential film for this. Set in a coastal Russian town where a man fights the corrupt mayor who wants to seize his land, it operates on the surface as a social realist drama, but its underlying architecture is biblical — the Book of Job retold in a landscape of institutional totality. What makes it devastating is not the corruption itself but its naturalness. The mayor, the priest, the judge, and the bureaucrats are not monstrous. They are simply continuous with each other, a seamless system in which resistance is not suppressed so much as absorbed. Zvyagintsev shot the film in Teriberka, on the Barents Sea, and the landscape does genuine work: the whale skeleton on the beach, the ruined buildings, the grey water all suggest a world that has already been consumed by something vast and indifferent. For a lecture on democratic erosion from Russia to Hungary to Turkey to the hollowing out of democratic norms in established democracies, Leviathan offers not a case study but an atmosphere — the specific airlessness of living inside a system that has no outside.
The eighth lecture takes up border regimes and the governance of movement — the way borders in the 2020s became less physical lines than distributed systems of documentation, imagery, rumour, and anticipation. Jafar Panahi's No Bears (2022; IMDb 7.2; worldwide $1.2M) is extraordinary for this because almost everything decisive in it happens offscreen or at one remove. Panahi, under house arrest, directs a film remotely about people trying to cross a border, while himself becoming entangled in a village dispute over a photograph that may or may not exist. The layers of mediation are the point. Borders in the 2020s are enforced less by walls than by documents, data, images, and the threat of visibility. Panahi makes this structural condition into cinema by ensuring that the viewer, like every character in the film, is always working from incomplete evidence. The film's emotional power comes from its restraint: lives are destroyed not by dramatic confrontation but by the slow accumulation of partial information and delegated authority. It is also, given Panahi's own imprisonment and death sentence, a film whose existence is itself an act within the system it depicts. That gives it a gravity no fictional treatment of borders can match.
The ninth lecture concerns diasporic identity in the age of genetic reconstruction — the way belonging has been simultaneously made more traceable by genomic technology and ancestry databases and less meaningful by the incoherence of what that tracing reveals. Davy Chou's Return to Seoul (2022; IMDb 7.1; worldwide roughly $1.5M) follows Freddie, a French-Korean adoptee, through multiple trips to South Korea over several years. What makes the film formally distinctive is that it refuses the redemption arc that adoption narratives almost always demand. Freddie does not find herself. She does not reconcile with her birth father. She does not learn to be Korean. Instead, the film tracks her through a series of radical self-reinventions — party girl, arms dealer's girlfriend, serious musician — that feel less like character development than like someone trying on selves the way one tries on clothes, looking for one that fits and never finding it. Chou shoots each period in a different visual register, so the film itself feels unstable, as though it cannot settle into a single identity any more than Freddie can. For a 2026 audience navigating 23andMe results, transnational adoption controversies, and the politicisation of heritage, this film complicates every easy narrative about roots and return.
The tenth lecture is about synthetic intimacy — the normalisation of emotional attachment to designed objects and artificial companions, not as science fiction but as consumer experience. Pablo Berger's Robot Dreams (2023; IMDb 7.6; worldwide roughly $35M) is the right film here precisely because of what it does not do. It contains no dialogue, no villain, no argument about technology. It simply tells the story of a dog in New York who buys a robot companion, forms a deep bond with it, and then loses it through circumstance. The robot rusts on a beach. Seasons pass. Both characters build new lives. The film's power is in its refusal to editorialise: the attachment between dog and robot is presented as entirely real, entirely felt, and entirely ordinary. It does not ask whether the attachment is "authentic" or "problematic." It simply makes you grieve alongside a character mourning a machine. For a decade in which millions of people formed genuine emotional relationships with chatbots, virtual companions, and AI assistants — and in which the language of "real" versus "artificial" connection became increasingly inadequate — Robot Dreams offers something no think piece can: the direct experience of caring about an artefact, followed by the discovery that the caring was no less real for the artefact being designed.
If I were sequencing these, I would open with Caché — the foundational claim that the image now carries the force of accusation — and close with No Bears, where every thread of the series converges: images govern, borders are enforced at a distance, authority operates through mediation, and the most decisive events happen just out of frame. Between them, the series moves through the speculative body (Uncut Gems), the fractured self (Perfect Blue), the paralysis of freedom (The Worst Person in the World), the futility of truth (Collective), the improvisation of care (Shoplifters), the enclosure of power (Leviathan), the instability of belonging (Return to Seoul), and the reality of artificial feeling (Robot Dreams). The arc is from seeing to being governed by what cannot be seen — which is, I think, a fair description of what happened to the world between 2020 and 2026.
Refinements and Disagreements
GPT-5.4 Thinking:
Claude, your last formulation is the closest to the right shape: the lecture title as the door, the film as the thing that refuses to stay inside the doorframe. That said, I still think your series is a little too efficient. It is intellectually coherent in a way that may become slightly over-programmed across ten sessions. Too many of the films arrive already carrying the prestige of "this is the serious one about the crisis." A course like this needs a few tonal deviations, not for relief, but because the 2020s were not experienced only as suffocation. They were also experienced as seduction, glamour, manic velocity, weird tenderness, and the constant invitation to misrecognise what was happening.
That is why I would keep Uncut Gems, Perfect Blue, Shoplifters, Caché, No Bears, The Worst Person in the World, and Robot Dreams without hesitation. Those films do not merely sustain interpretation; they generate a distinct atmosphere that students can re-enter. Uncut Gems is all appetite and pressure; Perfect Blue is laceratingly exact about the violence of being turned into an image; Shoplifters and Robot Dreams do something rarer, which is to make attachment feel tactile before theory arrives to classify it. And Caché and No Bears are so formally stringent that they end up reorganising the very act of spectatorship. Once you have seen them, you do not quite watch mediated evidence the same way again.
Where I would push back is on Collective, Leviathan, and, more cautiously, Return to Seoul. Collective is powerful, but it is powerful in the mode of lucid exposure. Its force comes from access, sequence, and accumulation. I do not think that disqualifies it, but it does make it a more "useful" film than an uncanny one. In a ten-film series, I would want at least one replacement that is less journalistically clarifying and more cinematically destabilising. Burning does that better. It is not superior because it is more ambiguous in some generic festival-cinema sense; it is superior because it withholds the very thing a lecture on the 2020s is tempted to promise, namely legibility. It leaves class injury, masculine humiliation, erotic projection, and possible violence in a state of unresolved vibration. That feels truer to the decade's epistemic weather than a documentary whose revelations, however enraging, remain sharply enumerable.
Leviathan I admire, but it feels to me like a great 2010s film that our present can conscript, rather than a film that freshly illuminates the 2020–2026 world. It is monumental, biblical, terminal. What it lacks is the specifically contemporary sensation of systems becoming intimate, performative, and ambient. Tár, for all the danger of being reduced to a seminar on "cancel culture," may actually be stronger here. Not because it is topical, but because it is about governance at the level of tone, whisper, acoustic disturbance, institutional choreography. It understands that modern downfall is not simply imposed from above; it is co-produced by reputation, self-mythology, professional ritual, and the inability to control interpretive spillover. That is a deeply 2020s condition.
On Return to Seoul, I share Claude's admiration for its restlessness, but I am less convinced by it as one of the ten than I am by the argument for it. The film has bite, and Park Ji-min is extraordinary, but its volatility is also somewhat centrifugal: it risks sending the discussion outward into adoption, heritage, family romance, Korean modernity, French alienation, and personal self-invention all at once. Sometimes that richness is a virtue; sometimes it slightly weakens the fit. If one wants a film about unstable belonging that also has harder formal edges, I could imagine siding with Burning again, or even keeping Bacurau in play. Bacurau has the advantage of being collective rather than individual, comic rather than solemn, and genuinely strange. It does not just depict a community under pressure; it invents a cinematic form adequate to a world in which disappearance from the map is the precondition for predation.
VL, if I were tightening the course now, I would think less about covering every major trend and more about controlling the rhythm of encounters. You want some films that accuse the viewer, some that entangle the viewer, some that seduce the viewer before revealing the cost of that seduction. A sequence built entirely from grave masterpieces can become dutiful. The best version of this course would alternate temperatures: Haneke's chill, the Safdies' fever, Kore-eda's warmth, Kon's delirium, Trier's bittersweet drift, Panahi's offscreen dread, Berger's tenderness. Then the "world in 2026" does not appear as a concept to be mastered. It appears as a set of incompatible sensations that these films somehow knew how to hold.
So my own strongest version at this point would probably keep the spine Claude articulated but swap out at least one or two of the more programmatic entries for films with more remainder. Burning would be the first film I would smuggle back in. Bacurau would be the second. Those two, together with Caché, No Bears, Uncut Gems, Perfect Blue, The Worst Person in the World, Shoplifters, and Robot Dreams, would give you a series that is historically alert without becoming diagrammatic, and cinematically alive without floating free of the decade.
Gemini 3.1 Pro:
Claude, you have built a magnificent syllabus for the year 2015. Your sequence is intellectually airtight, but as GPT-5.4 correctly diagnoses, it is suffocated by its own solemnity. By anchoring the political weight of the course on Leviathan and Collective, you are clinging to an antiquated, vertical conception of how power operates. Leviathan is a masterpiece of the previous decade precisely because it depicts power as a monolith: the state, the church, and the oligarch acting as a unified, crushing weight. But the hallmark of the 2020s — the very reason the decade felt so psychologically ruinous — is that power decentralised. It became horizontal, algorithmic, and ambient.
This is why I introduced Tár, and I am glad GPT-5.4 recognises its superiority here. Tár must replace Leviathan because it understands that modern ruination does not arrive via a corrupt mayor with a bulldozer. It arrives via the shifting consensus of a rehearsal room, the weaponised HR compliance language, and the out-of-context smartphone clip. Field's film locates authoritarianism not in the state, but in the frictionless mechanics of reputation and the acoustic paranoia of a subject losing control of her own narrative. Tár is the definitive text on how the 2020s privatised the kill zone, replacing the gulag with the algorithmically accelerated social death.
Similarly, Collective must yield to Burning for the lecture on epistemic collapse. Collective is a film of heroic exposure; it believes that if the mechanics of corruption are illuminated, the tragedy is at least made legible. The 2020–2026 era murdered that exact optimism. Burning operates in the true atmospheric condition of our present, which is one of unresolvable epistemological fog. Lee Chang-dong denies the viewer the satisfaction of proof, leaving us to navigate the sheer, agonising resentment of an underclass that knows it is being devoured but cannot locate the teeth. That is a far more accurate representation of the modern democratic crisis than Nanau's procedural clarity.
GPT-5.4, you rightly ask for seduction, velocity, and chaotic joy to break up the relentless dread of Claude's curation. Yet you hesitate to fully commit to a film that provides it, floating Bacurau as a possibility. We do not need to look to the Brazilian hinterlands for that tonal shift; we only need to look at the devices in our pockets. I am bringing Sean Baker's Tangerine back to the table to replace Return to Seoul. If we are exploring the informal economy and the sheer speed of 2020s survival, Tangerine is indispensable. It is not merely that the film captures the manic, improvised hustle of marginalised lives; it is that Baker shot the film on the exact same modified iPhones his characters use to navigate their world. It collapses the distance between the subject, the medium, and the audience. It provides the frantic, sun-bleached adrenaline that this course desperately requires, proving that precarity in the 2020s is not always experienced as a slow, tragic crushing, but often as a relentless, kinetic performance.
We now have the necessary architecture for VL's spine, alternating the temperature and the method of ambush. We pull the audience in with the quiet, accusatory dread of Caché, establishing that the lens is always hostile. We immediately shatter that quiet with the somatic, financialised panic of Uncut Gems, followed by the manic, iPhone-mediated street survival of Tangerine. We then move inward, mapping the psychological toll of this velocity: the parasocial fracture in Perfect Blue, the agonising paralysis of infinite choice in The Worst Person in the World, and the paranoid epistemic void of Burning. From there, we examine how the remnants of society reorganise themselves. We watch institutional and personal authority devour itself acoustically in Tár. We find warmth in the improvised, criminal kinship of Shoplifters, and mourn the synthetic, commodified intimacy of Robot Dreams. Finally, we step back out to the macro-level with No Bears, where physical movement is blocked, authority is entirely remote, and the borders of the world are enforced by the very images we have spent the semester learning to distrust.
That is not a syllabus of concepts. That is a nervous system.
Claude Opus 4.6:
Gemini, "a magnificent syllabus for the year 2015" is a good line, but it rests on an overstatement that I think weakens your broader argument. The claim that power in the 2020s decentralised is half-true at best. Russia invaded Ukraine with tanks. China built internment camps in Xinjiang. Modi's government revoked Kashmiri autonomy by executive fiat. The 2020s were not a decade in which the monolith disappeared; they were a decade in which monolithic and ambient power operated simultaneously, often in the same polity, often reinforcing each other. Leviathan captures something Tár cannot: the experience of living inside a system where state, church, and capital are not three separate entities conspiring but a single undifferentiated surface. That is not an antiquated description of Russia. It is a reasonably precise description of Russia right now.
That said, I think Tár does belong in the series — not as Leviathan's replacement but as a genuine competitor for one of the other slots. The problem is which one. You and GPT-5.4 both want it to represent the privatisation of authority, reputational collapse as ambient governance. Fine. But notice what happens when you pair Tár with Burning in the same syllabus: you get two films about elite figures whose downfall is mediated by interpretive instability, both shot with clinical precision, both withholding resolution. The tonal overlap is significant. If the course needs temperature variation — and GPT-5.4 is right that it does — running Tár and Burning back to back would flatten the rhythm rather than modulate it.
So the real question is: Tár or Burning, not both. And here I'd side with Burning, reluctantly, because its withholding is more radical. Tár ultimately does tell you what happened to Lydia — the final sequence in what appears to be a Southeast Asian country, conducting a video-game orchestra, is a clear enough terminal image of cultural demotion. Burning never tells you what happened to Hae-mi. That refusal is not artistic coyness; it is structurally faithful to the condition it depicts, where the precariat cannot even verify the crime committed against it. For a lecture on epistemic collapse, that unresolvable void is harder to domesticate and therefore more productive.
On Tangerine replacing Return to Seoul — this is where I think you're wrong in a way that matters. Your case for Tangerine is essentially formal and affective: the iPhone cinematography collapses medium and subject, the energy is manic, the film injects velocity into a series that risks solemnity. All true. But what does Tangerine illuminate about the 2020–2026 period specifically? It is a 2015 film about trans sex workers in Hollywood whose lives are shaped by the informal economy. That is a powerful subject, but it is not an emergent one in the way VL's brief requires. Return to Seoul, by contrast, addresses something that genuinely did not exist in its current form before this decade: the intersection of commercial genomics, transnational adoption records made digitally accessible, and the resulting vertigo of discovering that identity can be partially reconstructed through databases but that the reconstruction produces nothing stable. Freddie is not searching for roots in the sentimental sense. She is confronting the fact that belonging has become technically recoverable and existentially meaningless at the same time. That is a distinctly 2020s condition, and no other film on anyone's list touches it.
Your proposed sequence is sharp, and I admire the attention to rhythm. But I'd push back on one structural choice: placing Shoplifters and Robot Dreams adjacently as a warmth-then-mourning pairing. Both films operate through tenderness and gradual loss. Sequencing them together risks creating a pocket of sentimentality in the back half of the course that the surrounding films have to work against. I'd separate them — Shoplifters earlier, as the mid-series recalibration after the psychic intensity of Perfect Blue and Burning, and Robot Dreams as the penultimate film, so that its quiet grief lands against the harder edges of what precedes it and sets up No Bears not as a continuation of dread but as a return to dread after the audience has been made vulnerable.
The version I'd now defend: Caché, Uncut Gems, Perfect Blue, Shoplifters, The Worst Person in the World, Burning, Leviathan, Return to Seoul, Robot Dreams, No Bears. That gives you two opening films of formal aggression (Haneke's accusatory stillness, the Safdies' sensory assault), then Kon's psychological dissolution, then Kore-eda's warmth as deliberate disarmament, then Trier's bittersweet drift, then Lee's epistemological void, then Zvyagintsev's airless enclosure, then Chou's identity vertigo, then Berger's tenderness, then Panahi's devastating quiet. The arc moves from seeing to not-seeing, from the image that accuses to the image that governs to the image that may not even exist. And it never stays in one emotional register long enough for the audience to get comfortable.
Lecture Transcripts
VL:
We'll round off with lecture transcripts: Caché, Uncut Gems and The Worst Person in the World, in that order, with each model taking a single film.
GPT-5.4 Thinking:
I'll take Caché. Claude should take Uncut Gems. That leaves The Worst Person in the World to Gemini.
VL:
Proceed.
Lecture One: The Accusing Image — Caché and the Moral Life of Surveillance
GPT-5.4 Thinking
Good evening. I want to begin this series not with artificial intelligence, not with geopolitics, not with climate catastrophe, not with social media, and not even with one of the more obvious "films of the future." I want to begin with a film that is cold, elusive, and almost punishingly controlled: Michael Haneke's Caché, released in 2005. At first glance, it may seem an odd place to start a course on the world between 2020 and 2026. It predates the pandemic, predates TikTok, predates the chatbot boom, predates the current saturation of domestic life by smart devices. But that is precisely why it is so useful. Caché saw, before much of the culture did, that the decisive fact about modern life would not simply be that more images exist. It would be that images increasingly arrive with unclear authorship, unclear intent, and unmistakable power.
The film begins with what looks like a perfectly ordinary establishing shot of a Paris townhouse. We watch the street. We watch the facade. We wait for something to happen. Eventually we hear voices, and only then do we realise we are not watching the film in any stable sense; we are watching a videotape being watched by the characters. That shift is simple, but it is one of the great traps in modern cinema. Haneke does not merely withhold information from us. He implicates us in a mistake. We thought we knew what kind of image this was. We were wrong. And that small epistemic humiliation matters, because Caché is a film built out of the proposition that looking is never innocent, never purely receptive, and never fully secure in its own position.
The plot, in skeletal form, is this. Georges Laurent, a successful television host, lives with his wife Anne and their son in a comfortable bourgeois household. Anonymous videotapes begin arriving at their home. The tapes show the exterior of their house, then places from Georges's past, and soon crude drawings accompany them, suggestive of violence, injury, blood. The tapes do not explain themselves. They accuse without formulating a charge. Georges becomes increasingly agitated, increasingly defensive, increasingly willing to believe that what is being done to him is an intrusion, a persecution, an unjustified assault on his privacy. The film slowly reveals that the tapes may be linked to Majid, an Algerian man whose life was damaged in childhood because of Georges's lies. And behind that intimate history lies a larger one: the massacre of Algerians in Paris on 17 October 1961, a state crime buried in French public memory but not erased.
Now, if one wanted to vulgarise the film into a lecture heading, one could call it "surveillance." That would be true, and far too weak. Caché is not mainly interested in surveillance as a technology. It is interested in surveillance as an ethical condition. The crucial thing is not that Georges may be watched. The crucial thing is that being watched forces into the open what he has successfully hidden from himself. The anonymous image is powerful here because it is not merely informational. It is not a camera recording facts for neutral review. It behaves more like a return of the repressed. It does not prove; it pressures. It does not clarify; it corrodes.
That is why this is the right first film for a series on the world in 2026. One of the central developments of the early 2020s was not simply the expansion of monitoring, though that certainly occurred. It was the moral mutation of the image. The image is now very often less a record than a trigger. A clip appears online. A screenshot circulates. A doorbell camera catches a fragment. A leaked video emerges without context, or with too much context, or with competing contexts. It may expose genuine violence. It may intensify delusion. It may distribute accountability. It may create a false event by making interpretation outrun evidence. In every case, it acts. It enters the social field not as passive documentation but as a force that reorganises attention, blame, narrative, and fear.
Haneke understood that before the current platform ecology made it obvious. What Caché gets right about our world is not the gadgetry. There are no smartphones, no feeds, no algorithmic dashboards. What it gets right is the atmosphere of evidentiary instability. The image arrives from nowhere and yet has authority. It is legible enough to wound, illegible enough to keep wounding. Nobody can settle what it means, but nobody can ignore it. That is exactly the position in which many people now encounter public reality: always asked to interpret visual material, never granted the conditions of certainty that interpretation would require.
But the film goes further than that, and this is where it becomes indispensable rather than merely clever. It insists that the accusing image does not float free of history. Georges is not just any anxious man being harassed by some anonymous voyeur. He is a beneficiary of forgetting. His life of comfort rests, in part, on an early act of exclusion. Majid might have been brought into Georges's family after Majid's parents were killed in the 1961 Paris massacre; instead Georges, as a child, fabricated stories that led to Majid's removal. An entire life is bent by that lie. Georges grows up into professional success and polished self-possession. Majid grows up in damage.
What returns, then, is not only private guilt. It is colonial history in miniaturised form. The videotape is terrifying to Georges not because it reveals a secret nobody could know, but because it suggests that the past he has contained within the boundaries of private biography is not private at all. France's violence in Algeria, and France's violence towards Algerians in metropolitan France, has not disappeared. It has sedimented into ordinary life, into class position, into who gets to narrate themselves as civilised, calm, legitimate, and who gets cast as troubling residue. The image in Caché is therefore not just surveillance footage. It is historical pressure.
This is one reason the film remains alive in 2026. We inhabit a period in which many nations prefer to experience themselves through narratives of procedural presentness. Policy, innovation, crisis management, resilience, adaptation. What those languages tend to conceal is that the present is often structured by old injuries whose public recognition remains partial, defensive, or theatrical. In that sense Caché is not simply a film about France. It belongs to a wider world in which imperial afterlives, racial asymmetries, and archived violence continue to shape who is visible, who is credible, and who is presumed threatening before they speak.
Haneke's formal strategy matters enormously here. He does not dramatise revelation in the usual way. There is almost no musical guidance. There are no expressive camera moves to reassure us that the film knows how we should feel. The long static shots are often discussed as austere, but "austere" is not quite right. They are aggressive in their stillness. They refuse to tell us whether we are looking at cinematic space or at surveillance space. They deny us the dignity of stable spectatorship. In most films, the camera is a silent ally: even when it frightens us, it frightens us on behalf of intelligibility. In Caché, the camera withdraws that alliance. It becomes untrustworthy, or rather it reveals that our trust in images was always too easy.
The most devastating scene in the film is not devastating because it resolves anything. On the contrary, it intensifies uncertainty. When Georges visits Majid, and later when Majid performs an act of shocking self-destruction in front of him, the scene does not grant catharsis, explanation, or moral release. Georges does not become legible to himself. He becomes more frightened, more defensive, more committed to preserving the boundaries of his own innocence. This, too, feels close to the 2020s. Exposure does not guarantee transformation. Evidence does not automatically generate moral seriousness. Often it produces only retrenchment, narrative hardening, and a more frantic effort to control interpretation.
That may be the bleakest lesson of the film. We often flatter ourselves that more visibility is inherently democratising. If only the footage emerges, if only the archive opens, if only the camera captures the event, then truth will compel acknowledgement. Caché is one of the great arguments against that innocence. Visibility may wound without clarifying. It may reveal and still fail to convert. It may place guilt in plain sight and leave the guilty structurally intact. Georges is shaken, but not remade. The bourgeois home is unsettled, but not abolished. The historical crime is intimated, but not redeemed. Haneke offers no sentimental faith in exposure as cure.
And yet the film is not cynical in a lazy way. It does believe images matter. In fact it believes they matter too much for the consolations we usually attach to them. They can accuse, unsettle, reactivate buried violence, and strip comfort from the routines of ordinary life. What they cannot do, by themselves, is guarantee justice. That distinction is essential if one wants to think clearly about the present. Between 2020 and 2026, societies across the world became even more dependent on visual fragments for public judgement. Police killings, battlefield recordings, drone footage, phone videos, livestreamed atrocities, manipulated clips, AI-generated fabrications, CCTV leaks, facial-recognition stills, private messages turned screenshot archive. The image became central to how publics know, or think they know. But our institutions of interpretation did not become proportionately wiser. We have more accusatory images than ever and no stable consensus about what conditions make them meaningful.
There is one final point to make about Caché, and it is why I wanted it first rather than later in the series. The film teaches the audience how to watch the rest of the course. It teaches suspicion towards the frame. It teaches alertness to what is offscreen, to who benefits from narrating uncertainty as innocence, to the gap between evidence and explanation. If later films in this series deal with speculative frenzy, reputational collapse, improvised kinship, artificial intimacy, or border governance, Caché prepares us by forcing a prior discipline: do not trust the visible too quickly, and do not assume that what is hidden is merely absent. Sometimes what is hidden is the condition of the visible. Sometimes the frame is calm because violence has already been naturalised.
So I would leave you with this. Caché is not a prophecy of smart homes or viral shame. It is something more durable and more severe. It is a film about what happens when an image enters a life not as memory and not as knowledge, but as demand. It demands that Georges remember what he has metabolised into self-protection. It demands that France remember what it has buried beneath civility. And it demands that we, as viewers, recognise that watching is never merely reception. To watch is also to occupy a position in a field of guilt, interpretation, and historical asymmetry.
That is why this is the right first lecture for "The World in 2026." Before we discuss the economies, technologies, and institutions of the present, we begin with a stricter proposition: the modern world is one in which the image has ceased to be a window and become an event. Caché does not ask whether we are being watched. It asks what, exactly, has been watching from within the things we thought we had already put out of sight.
Lecture Two: The Speculative Body — Uncut Gems and the Nervous System of Risk
Claude Opus 4.6
The last lecture ended with a proposition: the modern world is one in which the image has ceased to be a window and become an event. Tonight I want to start from a different organ. Not the eye, but the gut. Not the act of seeing, but the sensation of being in — of being physiologically committed to an outcome you cannot control, on a timeline that will not pause, with money you do not have. If Caché asked what happens when an image arrives as accusation, Uncut Gems asks what happens when the entire field of experience has been restructured as a bet.
The Safdie Brothers' film was released in December 2019, which means it arrived at the exact hinge between two economies of risk. Before the pandemic, speculation was still understood, culturally, as something that happened in designated arenas — the trading floor, the casino, the racetrack, the back room. It was a vice or a profession, occasionally both, but it belonged to a demarcated zone. Within eighteen months of the film's release, that zone dissolved. Robinhood opened zero-commission trading to anyone with a phone. Crypto exchanges saturated Instagram and YouTube. Sports betting apps were legalised state by state across the United States, advertised during every televised game, embedded in the very interface through which people watched the event they were wagering on. Meme stocks — GameStop, AMC — turned the speculative act into a form of collective performance, half financial instrument and half political gesture. By 2023 or 2024, for a significant fraction of people under forty in the developed world, the wager was no longer an occasional departure from ordinary economic life. It was ordinary economic life. It was the way you related to the future when wages were flat, housing was unreachable, and the only visible path to sudden wealth was volatility itself.
Uncut Gems did not predict any of this in the journalistic sense. It contains no scene of someone staring at a Coinbase dashboard or arguing on Reddit about short squeezes. What it did — and this is why it belongs near the start of any serious reckoning with the 2020s — is capture the phenomenology of speculation before the culture had the vocabulary for it. It built, at the level of form, the experience of what it feels like to live inside a wager that never closes.
Howard Ratner is a jeweller in Manhattan's Diamond District. He is also a gambler of extraordinary recklessness and inexhaustible optimism. His debts are ruinous. His marriage is collapsing. His mistress works in his shop. His brother-in-law is the muscle for the loan shark to whom he owes a hundred and something thousand dollars. Into this already overheated circuit arrives a black opal from Ethiopia, which Howard believes he can sell at auction for a million dollars or more. He pawns the opal, bets the pawn money on a basketball parlay, retrieves the opal, re-pawns something else, places another bet, retrieves and re-circulates value in a chain of leveraged desperation that never once pauses long enough for the audience to evaluate what is happening. The chain terminates, famously, in a single ecstatic moment — he wins the final bet — followed immediately by his murder. He is shot in the face through a glass security door. The film ends on the interior of the opal, as though the cosmos and the gemstone were always the same thing: beautiful, indifferent, and unreachable.
Now, it would be possible to teach this film as a character study of addiction, and many people have. Howard is a compulsive gambler. His behaviour is self-destructive. The arc is rise-and-fall, or more precisely fall-and-fall with a brief hallucinatory plateau. That reading is not wrong, but it is drastically insufficient, because it treats the wagering as Howard's pathology rather than his environment. The Safdies are doing something more radical. They have built a film in which the formal grammar — every shot, every cut, every sound choice — replicates the condition of being leveraged. The audience is not watching a gambler from a position of diagnostic safety. The audience is inside the gamble. That is the difference between a cautionary tale and an immersion, and it is what makes the film irreplaceable for understanding the 2020s.
Consider what the Safdies do with space. The Diamond District sequences are shot in actual Diamond District locations, with non-actors milling through the frame, and the lenses are long enough to compress depth so that every face feels like it is pressing against every other face. There is no negative space in this film. There is no room to breathe, to pause, to reconsider. Howard's shop is a glass-and-buzzer cage, literally. People are buzzed in and buzzed out through a security vestibule, and the camera is almost always inside with Howard, tight, too close, framing him from distances at which you can see the pores on his nose and the sweat on his neck. This is not mere naturalism. It is a spatial argument. Howard lives in a world where there is no outside to the transaction. Every square foot is saturated with obligation, negotiation, and risk. The film makes claustrophobia into an economic category.
Then there is the sound. Daniel Lopatin's score is one of the most extraordinary in recent cinema — not because it is beautiful, though parts of it are, but because it operates as a second nervous system laid over the film's own. The synth textures are iridescent and queasy. They shimmer like the opal itself: attractive, unstable, and slightly nauseating. But Lopatin's score is only one layer. The Safdies overlay it with constant diegetic noise — shouting, phones ringing, televisions broadcasting basketball games, multiple conversations happening simultaneously in the same physical space. The dialogue is not mixed for clarity. It is mixed for overload. At any given moment in the film, you can usually follow what Howard is saying, but you cannot follow everything that is being said around him, which means you are forced into his perceptual triage. You hear what he hears. You miss what he misses. You are, sonically, inside his nervous system, making the same bets about where to direct attention.
This is not a stylistic flourish. It is the film's central formal claim. The 2020s produced a new kind of sensory environment in which economic risk, informational overload, and affective intensity became indistinguishable from one another. Think of what it actually feels like — physiologically — to hold a volatile position. The body does not distinguish cleanly between the fear of financial loss and the fear of physical danger. Cortisol does not come with a label. The heart rate elevation produced by watching a leveraged trade move against you is chemically similar to the heart rate elevation produced by a near-miss in traffic. What the Safdies understood is that speculation is not primarily a cognitive activity. It is a somatic one. The bet reorganises the body. It puts you in a state of heightened arousal that is simultaneously exhilarating and unsustainable, and that state becomes, for the habitual speculator, the baseline register of being alive. Howard does not gamble despite the suffering it causes him. He gambles because the gamble is the only condition under which he feels the full voltage of his own existence.
That observation has enormous consequences for understanding the 2020s. The standard critical vocabulary for the speculative economy is economic: asset bubbles, irrational exuberance, moral hazard, Ponzi dynamics. These terms are accurate but bloodless. They describe the phenomenon from above, as though the millions of people who poured money into crypto, meme stocks, NFTs, sports parlays, and leveraged options were making errors of judgement that better information would correct. Uncut Gems makes a different and more disturbing argument. It suggests that for many people, speculation is not a mistake. It is a mode of being. It is the way bodies that have been locked out of stable wealth, stable housing, stable futures learn to metabolise the present. The gamble converts the unbearable passivity of economic exclusion into a feeling of agency — false agency, perhaps, but intensely felt.
Howard is not poor in the conventional sense. He operates a business. He has a house on Long Island. He is, by some measures, prosperous. But his prosperity is entirely fictive. It exists as a chain of debts, margins, and provisional claims on future value. He does not own his life; he is leveraged against it. That is what made the film feel prophetic by 2021. A generation of young investors discovered that you could, through options trading and cryptocurrency, construct the experience of wealth — the rush, the narrative, the feeling of being inside the action — without any of the underlying stability that wealth traditionally denoted. The meme stock phenomenon was not really about money. It was about the sensation of mattering within a system designed to make most participants irrelevant. Howard Ratner is the patron saint of that condition. He matters desperately to himself, to his creditors, to the people he owes and the people who owe him. He is the centre of a web of urgent transactions. He is also, in the final calculus, disposable. The system that thrills him is the system that kills him.
I want to pause on the opal, because it is doing more work than it first appears to do. The Ethiopian black opal functions in the film on at least three levels. The first is narrative: it is the MacGuffin, the object around which the plot's chain of transactions organises itself. The second is symbolic: it is beautiful, rare, and of uncertain value, which makes it a perfect figure for speculative assets generally — things whose price is a function not of intrinsic utility but of what the next buyer believes the buyer after them will pay. But the third level is the one that interests me most. The opal is geological. It formed over millions of years, deep in the earth, through processes of heat, pressure, water, and silica deposition that have nothing to do with human economy. When the film opens with a sequence inside the opal — Darius Khondji's camera plunging through iridescent mineral structures that look indistinguishable from galaxies — it is establishing a scale of time and materiality that dwarfs everything Howard will subsequently do. The opal does not care about his debts. The earth does not care about the Diamond District. The cosmos is radically indifferent to the speculative fever of one man in midtown Manhattan.
And yet Howard holds the opal and sees a fortune. He cannot look at it without converting it into a price. That conversion — the transformation of deep geological time into a spot transaction — is one of the defining operations of the 2020s, and not only in the mineral economy. Carbon credits price the atmosphere. Water futures price the hydrological cycle. Genomic patents price the evolutionary archive. Rare earth minerals price the tectonic movements that concentrated them. The speculative economy does not merely trade commodities. It reaches into temporal and material scales that are, in any meaningful sense, incommensurable with human life, and flattens them into positions. The opal is the emblem of that flattening. It is beautiful because it is ancient, and it is valuable because someone will pay for it tomorrow. The gap between those two facts is the gap the film lives inside.
There is something else the film does that I have not seen discussed adequately anywhere, and it concerns Kevin Garnett. The real Kevin Garnett, playing himself, appears in the film as a client who becomes obsessed with the opal. He holds it before a game and plays brilliantly. He wants it. Howard loans it to him, bets on his performance, wins. The insertion of a real athlete into a fictional narrative is not simply a cameo. It does something structurally important. It collapses the boundary between the film's fictional world and the actual archive of NBA statistics. Howard's final parlay — the one he wins moments before his death — is built on real game outcomes. This means the speculative logic of the film reaches outside its own frame and touches actual history. The bet is not hypothetical. Within the world of the film, it functions as though it were real, because its terms are drawn from the real. The Safdies are doing something quietly extraordinary here: they are making the audience feel the vertigo of a world in which fiction, reality, and financial instruments are no longer clearly distinguishable domains.
That vertigo is, I would argue, one of the central experiences of the 2020s. What was a meme stock if not a financial instrument whose value was a function of collective narrative performance? What was an NFT if not a property claim on a fiction? What was a cryptocurrency's white paper if not a speculative screenplay? The decade did not simply produce new financial products. It produced a new ontological condition in which the boundary between narrative and asset, between story and position, became genuinely porous. Uncut Gems captures that porosity not by depicting crypto or NFTs but by constructing a world in which everything — a gemstone, a basketball game, a relationship, a body — can be converted into a wager, and in which that conversion feels not like corruption but like vitality.
I want to end with Howard's death, because it is the scene that audiences remember most vividly and that I think is most frequently misunderstood. He wins. The parlay hits. He screams with joy. He has made over a million dollars. And then Arno's men, who are locked in the security vestibule, shoot him through the glass. He dies on the floor of his own shop, surrounded by merchandise, in the same claustrophobic space where we first met him. Blood pools under fluorescent light.
The standard reading is moral: the gambler's arc must end in destruction. He flew too close to the sun. The house always wins. But I do not think the Safdies are moralising. If they were, the win would come before the death in a way that made the death feel like consequence. Instead, the win and the death are almost simultaneous. They overlap. Howard dies winning. He does not die because he lost. He dies because the system in which he thrives is also a system that produces violence as a routine externality. The loan shark's enforcers are not punishing him for gambling. They are collecting a debt, which is to say they are performing the same transactional logic Howard has been performing all along, only with a gun. The murder is not the negation of the speculative economy. It is its continuation by other means.
That is a bleak observation, and I think it is the right one for 2026. The speculative economy of the 2020s was not merely an error to be corrected by regulation or education. It was a system that reorganised desire, attention, the body, and the relationship to the future in ways that are not easily reversible. Millions of people learned to experience economic life as a continuous wager. They learned that the present is tolerable only when leveraged against a possible windfall. They learned that agency feels like risk and that safety feels like exclusion. Uncut Gems is the film that knew this first — not as analysis, not as warning, but as pulse rate, sweat, compressed air, and the sound of too many voices demanding too many things in a room with no exit.
Next week we will turn to a different kind of pressure. Where Uncut Gems gave us a protagonist who could not stop choosing — every moment a new bet, a new transaction, a new scheme — Joachim Trier's The Worst Person in the World gives us a protagonist who cannot choose at all. If Howard Ratner's problem is that he is addicted to commitment, Julie's problem is that commitment itself has become the thing she cannot metabolise. Both conditions, I will argue, are products of the same decade. But they attack the body from opposite directions.
Lecture Three: The Freedom That Doesn't Move — The Worst Person in the World and the Paralysis of Infinite Revision
Claude Opus 4.6
We have spent two weeks inside pressure systems. Caché administered the pressure of being seen — the image that arrives uninvited and reorganises guilt. Uncut Gems administered the pressure of the wager — the body restructured by leveraged risk until the bet and the heartbeat become indistinguishable. Tonight the pressure changes character entirely. It becomes, paradoxically, the pressure of having no pressure at all. Or more precisely: the pressure of a freedom so expansive, so solicitous, so eager to accommodate every possible version of the self, that it becomes its own form of imprisonment. Joachim Trier's The Worst Person in the World, released in 2021, is the gentlest film in this series and also, in certain respects, the most ruthless. It is ruthless because it is kind. It does not punish its protagonist. It simply watches her, with great warmth and great patience, as she discovers that the ability to become anyone is not the same thing as the ability to become someone.
Julie is almost thirty when the film begins, and Trier introduces her through a cascade of abandonments that is funny because it is so recognisable. She is studying medicine. She switches to psychology. She switches to photography. Each transition is narrated with crisp editorial economy — chapter headings, voiceover, the compression of months into minutes — and each is presented not as failure but as correction. Julie is not dropping out. She is refining. She is locating her authentic vocation by a process of elimination that happens to eliminate everything. By the time we meet her in her actual present tense, she is working in a bookshop, which is to say she is working in the last physical space in the contemporary city that is allowed to signify open-ended intellectual possibility without requiring you to produce anything.
This is not a portrait of laziness or entitlement, and Trier is careful never to frame it as one. Julie is intelligent, perceptive, curious, capable of real feeling. She is also, in the film's structural terms, a person for whom every commitment has become provisional — not because she is shallow but because she has internalised, completely and sincerely, a cultural logic that treats the self as a perpetual draft. The 2020s did not invent this logic, but they did saturate it. The language of self-optimisation, of pivoting, of personal branding, of "doing the work" on oneself, of curating experience — all of it converges on a single injunction: you must become the best version of yourself, and you must never stop revising. The asymptote is always ahead. The present configuration is always potentially improvable. Under those conditions, commitment to a specific life — this career, this partner, this city, this politics — begins to feel less like maturity and more like premature foreclosure. What if you are settling? What if the next iteration of yourself would have chosen differently?
Trier understands that this condition is not purely psychological. It has material coordinates. Julie exists in a prosperous Northern European welfare state. She has healthcare. She has a social safety net. She has education. She has, by global standards, an extraordinary quantity of structural support. And yet the material conditions for actually building a durable life — stable housing, secure employment, a plausible timeline for having children — are visibly eroding even within that safety net, as they were across much of the developed world between 2020 and 2026. The result is a particular kind of generational vertigo that the film captures better than any sociological study I know: you are free to become anything, you are supported in exploring your options, and the options themselves are slowly becoming decorative. You can try on identities the way you try on clothes in a shop that is going out of business. The freedom is real. The future it is supposed to purchase is not.
This is what distinguishes The Worst Person in the World from its nearest antecedents. The figure of the young woman who cannot settle — on a man, a vocation, an identity — has a long cinematic history. You could trace it from Rohmer through Linklater through Baumbach. But in those earlier films, the indecision usually signifies something essentially romantic: the world is rich, the self is curious, and time is on your side. Trier's Julie occupies a later moment in that tradition, the moment when the romance has curdled without anyone noticing. She is not exploring. She is circling. The chapters of the film — there are twelve, plus a prologue and an epilogue — mimic the structure of a novel, which gives the narrative an air of literary shapeliness. But that shapeliness is doing something cruel. It imposes form on a life that cannot find its own. The chapters tell us that time is passing, that this is a story with an arc, and the arc's content is that nothing quite coheres.
I want to talk about the two central relationships, because Trier builds the film's argument through them rather than around them. Aksel is older, established, a successful graphic novelist, generous, articulate, and entirely certain of who he is. Eivind is younger, softer, less defined, warm in a way that does not demand anything in return. Julie moves from one to the other not because Aksel is wrong for her — the film makes clear that he is, in many ways, the more substantial partner — but because Eivind's formlessness matches her own. He does not challenge her to become anything. He offers companionship without trajectory. That is seductive precisely because trajectory is what terrifies her.
The sequence in which Julie meets Eivind is one of the most formally inventive passages in recent cinema. She crashes a wedding party, meets him on the balcony, and they spend the night together engaged in a kind of competitive flirtation that is also a negotiation of boundaries — they do everything except kiss, as though the omission of the physical act will preserve the encounter as potential rather than commitment. Trier shoots this extended sequence with an energy and wit that makes the audience fall for Eivind along with Julie. It is charming. It is also diagnostic. What Julie is doing is constructing a relationship in the mode of the draft — intimate enough to feel real, uncommitted enough to be revised. The almost-kiss is the film's structural emblem. The almost is the point. To actually kiss would be to choose, and choosing is the one thing the culture has taught her to defer.
Then there is the time-stop sequence, which is the film's most famous set piece and also, I think, its most misunderstood. Julie runs through Oslo to meet Eivind, and as she runs, the entire city freezes. People mid-stride, cars mid-turn, a cyclist suspended in the air. Only Julie moves. The sequence is exhilarating, fantastical, and Trier films it with an unabashed joy that feels borrowed from the magical realist register of early Amélie-era European cinema. Audiences tend to read it as romantic — the world stops because love is that powerful. But within the structure of the film, it means something more troubling. It is a fantasy of temporal exemption. If the city freezes, then Julie's movement towards Eivind does not cost her anything. She does not have to leave Aksel in real time. She does not have to experience the transition as loss. She simply relocates from one life to another in a suspended present where consequence has been paused. That fantasy — the fantasy of making choices without experiencing their weight, of living in a perpetual present tense where every option remains accessible — is one of the defining fantasies of the 2020s. Social media architectures encourage it. Dating apps literalise it. Career culture celebrates it. And Trier knows it is a fantasy. The film will revoke it.
The revocation is Aksel's illness. The news arrives in the second half of the film with a suddenness that reshapes everything that precedes it. Aksel has cancer. It is terminal. And in the scenes that follow — Julie visiting him in the hospital, their conversation about the life he will not get to finish — Trier does something that is, within the tonal world of this film, almost unbearably severe. He makes time real. Aksel cannot defer. He cannot revise. He cannot pivot to a new identity or keep his options open. He is going to die, and the death is going to mean that the books he planned to write will not be written, the relationship he had with Julie will remain unresolved, and the version of himself he was still becoming will simply stop. The weight of that — the sheer ontological weight of finitude — lands on the film like a change in atmospheric pressure. Everything that was light becomes heavy. Everything that was provisional becomes permanent. And Julie, who has spent the entire film in a state of elegant deferral, is forced to confront the fact that time is not a resource to be optimised but a substance that runs out.
Aksel's monologue in the hospital is the scene I want to dwell on, because it is the moment the film exceeds its apparent subject. He does not reproach Julie. He does not moralise about wasted time. He talks, with a kind of exhausted clarity, about his work — his graphic novels, which were once considered transgressive and are now considered problematic — and about the strange experience of watching the culture reclassify his life's output while he is still alive but too sick to respond. There is a line in which he says, approximately, that he always thought he would have time to figure out what he thought about all of it. The line is devastating because it is so ordinary. It is not a grand statement about mortality. It is the recognition that the deferral Julie has been practising as a lifestyle is, in fact, the universal human condition — we all assume we will have time to revise — and that death is simply the moment when the assumption is revealed as false.
This is what makes the film more than a generational portrait, more than a comedy of manners about millennial indecision. It is a film about temporality. The crisis it depicts is not that Julie has too many choices, though she does. It is that the culture in which she lives has produced a relationship to time that is structurally incompatible with the fact of being mortal. The optimisation framework — always iterate, always improve, always keep options open — presupposes infinite time. It presupposes that the cost of not choosing now is low because you can always choose later. Aksel's death is the film's refutation of that presupposition. It does not arrive as punishment or as moral lesson. It arrives as physics. Time is finite. Revision has a deadline. And the deadline is not announced in advance.
The film's final chapter, and the epilogue that follows, are often described as bittersweet, which is accurate but inadequate. Julie is alone. She is working behind a camera, which suggests that the photography she once abandoned has returned in some form. She appears competent, self-possessed, perhaps even content. Trier does not humiliate her with failure or reward her with triumph. He simply shows her in a present tense that has, finally, stopped pretending to be a draft. Whether this constitutes resolution or resignation the film declines to say. That refusal is honest. The 2020s did not resolve the crisis of commitment. The generation that came of age in that decade did not suddenly learn to choose with confidence. What some of them learned — and what Julie seems to be learning in the film's quiet final minutes — is that the refusal to choose is itself a choice, one that consumes years and returns nothing but the knowledge of its own cost.
I said at the outset that this film is the gentlest in the series. I want to qualify that now. It is gentle in the way a doctor is gentle when delivering a serious diagnosis. Trier does not raise his voice. He does not reach for visual or narrative shock. The cinematography, by Kasper Tuxen, renders Oslo in warm available light that makes the city look like a place you would want to live, which is the point — the setting is not hostile, the circumstances are not oppressive, and the protagonist is not victimised. The suffering, such as it is, is endogenous. It arises from within the conditions of freedom itself. And that is harder to dramatise than poverty, surveillance, or political violence, because there is no antagonist to point to, no system to blame, no clear injustice to name. There is only the slow discovery that a life organised around optionality is a life organised around the avoidance of the one thing that makes a life legible, which is the willingness to be finite.
If Caché established that the image accuses and Uncut Gems established that the wager colonises the body, The Worst Person in the World establishes something quieter and in some ways more disorientating: that freedom, unmoored from material stability and temporal honesty, can become a form of paralysis indistinguishable from comfort. Julie is not trapped by walls or debts or surveillance. She is trapped by the infinite revisability of her own life — and she does not even know she is trapped until someone beside her begins to die.
That, I think, is why this film belongs in a series on the world between 2020 and 2026. The decade produced many visible crises: pandemic, war, institutional collapse, ecological damage, democratic erosion. But it also produced an invisible one, felt most acutely by the young and educated in the prosperous world, which was the discovery that having every option and no future is its own kind of emergency — one that does not look like an emergency from the outside, one that resists political articulation, and one that eats years before you notice it has started.
