Blog · Review Essay · Modified July 10, 2026 · Last reviewed July 10, 2026

More Everything Forever and the Future as an Ownership Claim

Adam Becker's More Everything Forever is a useful book for this site because it treats AI futurism as a belief system with institutional consequences. The point is not that every concern about advanced AI is foolish. The point is that speculative futures can become a way for powerful people to claim authority over the present.

For this review, a future ownership claim is a move from forecast to authority: because a speaker says they can see humanity's long-run destiny, they ask for present money, data, compute, labor, legal permission, secrecy, or deference. The governance test is whether the forecast remains answerable to evidence, affected people, and reversible public decisions.

The useful standard is not anti-futurism. It is claim custody: every grand forecast should retain its owner, date, evidence, uncertainty, present demand, affected people, review trigger, and conditions for withdrawal.

The Book

More Everything Forever: AI Overlords, Space Empires, and Silicon Valley's Crusade to Control the Fate of Humanity was published by Basic Books on April 22, 2025. Hachette's publisher page lists Adam Becker as author, 384 pages, and ISBN-13 9781541619593. Amazon lists the same title and author, the same publication date and page count, ISBN-10 1541619595, and ISBN-13 978-1541619593. Hachette's author bio identifies Becker as a science journalist with a PhD in astrophysics, and Yale's Whitney Humanities Center lists him as author and astrophysicist for a 2026 event on the book.

The subject is not AI alone. Becker reads the linked fantasies of superintelligent rulers, space settlement, radical longevity, limitless growth, and elite stewardship as a single political imagination. The book is at its best when it asks who benefits when a remote future is treated as more urgent than living people, public institutions, climate repair, labor rights, or democratic control.

That makes the book more useful than a general complaint about optimism. Its target is not invention, research, or long-range planning. Its target is authority laundering: the shift from "this may be possible" to "therefore this coalition should own the route, set the pace, absorb the subsidy, and define which objections count."

Current Context

Read on July 10, 2026, Becker's critique has become easier to operationalize because AI governance now contains more concrete records than it did when the 2023 pause-letter debate began. The European Commission says general-purpose AI model obligations under the EU AI Act entered application on August 2, 2025, with Commission enforcement powers beginning August 2, 2026. Providers of the most advanced general-purpose models that pose systemic risks must notify the AI Office, and submission channels now cover notifications, serious-incident reports, safety-and-security frameworks, and model reports.

The EU timeline also shows why dates need discipline. Article 50 transparency obligations for AI interaction and synthetic-content marking apply from August 2, 2026 under the original AI Act schedule. But the high-risk system timeline has shifted in the June 2026 simplification package: the European Parliament approved the changes on June 16, 2026, and the Council gave final approval on June 29, 2026, with obligations now described as applying from December 2, 2027 for stand-alone high-risk systems and August 2, 2028 for high-risk systems embedded in regulated products, subject to Official Journal publication and entry into force. A future claim that ignores implementation dates is already losing contact with governance reality.

U.S. federal policy has a different emphasis, but it points to the same translation problem. OMB M-25-21 frames federal AI use around innovation, governance, and public trust while requiring agency AI strategies, Chief AI Officer responsibility, inventories, minimum practices for high-impact uses, and discontinuation when mitigation or performance is inadequate. OMB M-25-22 turns acquisition into a place where agencies must test fitness for purpose, data rights, interoperability, privacy, risk management, vendor disclosures, and public trust. OMB M-26-04 adds requirements for large-language-model procurement under "truth-seeking" and "ideological neutrality" principles. These memoranda do not settle Becker's argument, but they show where futurist ambition becomes procurement paperwork, contract language, and decision authority.

The scientific risk record is also less theatrical than the public rhetoric around it. The 2026 International AI Safety Report, led by Yoshua Bengio, authored by over 100 AI experts, and backed by over 30 countries and international organizations, describes advancing general-purpose AI capabilities, emerging risks, evidence gaps, and the difficulty of risk management. That supports serious preparation, but it also cuts against prophecy: the report is a synthesis under uncertainty, not a license for private actors to govern the future on the public's behalf.

"More" has also become material. The International Energy Agency projects global data-center electricity consumption roughly doubling to about 945 TWh by 2030 in its base case, with accelerated servers driven mainly by AI adoption growing faster than conventional servers. A future sold as disembodied abundance now has a power bill, land-use footprint, grid negotiation, cooling system, chip supply chain, and local political cost.

The July 7, 2026 EU Action Plan on Cybersecurity and Artificial Intelligence adds another concrete layer: advanced AI creates cybersecurity evaluation, secure-access, and secure-testing problems before any grand destiny arrives. Future-facing risk language becomes useful when it funds public evaluation capacity and secure deployment checks. It becomes dangerous when it centralizes deference without giving outsiders the evidence or authority to intervene.

Future as Private Theology

The book's strongest media-theory move is its treatment of futurism as ritualized authority. Silicon Valley prediction often arrives dressed as engineering: curves, charts, scaling laws, roadmaps, launch windows, and sober talk about civilization. Becker's argument is that much of this language functions more like theology than forecasting. It offers salvation from death, scarcity, biology, planet, and politics, then asks the public to accept a priesthood of founders, investors, and technical insiders.

That does not mean every long-range worry is a cult. It means belief has infrastructure. A model benchmark, a funding round, a keynote, a safety summit, a manifesto, and a billionaire's public wager can merge into a world-picture. Once that picture hardens, disagreement can look like ignorance, delay, or hostility to the future. The future becomes a property claim: those who say they understand it demand permission to build it.

The mechanism is concrete. A speculative claim becomes a deck; the deck becomes a funding round; the funding round becomes a hiring market; the hiring market becomes infrastructure demand; the infrastructure demand becomes public utility negotiation; the negotiation becomes a claim on energy, land, water, roads, tax treatment, and political attention. By the time the original forecast is challenged, it may already have become an investment position with local costs.

The sharper definition is not "prediction is bad." Prediction becomes dangerous when it changes burdens of proof. A forecast should name a target, date, uncertainty range, evidence base, update rule, and decision it can change. A future ownership claim skips that discipline. It treats the imagined beneficiary, often "humanity" in the abstract, as more politically real than the workers, communities, users, patients, students, artists, public servants, and local governments who absorb the costs now.

The Future-Claim Test

Becker's critique becomes strongest when it is converted from a mood into a test. A future claim should be read in the present tense: what power is it asking for now? The request may be money, data, compute priority, land, energy, talent, legal immunity, secrecy, faster release, weaker labor review, relaxed procurement, exemption from ordinary safety duties, or deference to a founder's map of history. The further the claim reaches into destiny, the more specific its present demand should become.

The first discipline is classification. A forecast says what might happen and should name dates, uncertainty, update rules, and falsifiers. A warning says what could go wrong and should name thresholds, mechanisms, and mitigations. An investment thesis says where capital may go and should name who bears loss if the thesis fails. A product promise says what a system can do and should be substantiated like any other market claim. An emergency authorization asks for unusual speed, secrecy, surveillance, spending, or exemption and should require public authority, independent review, sunset dates, and appeal.

That classification links this review to claim hygiene, AI capability forecasting, AI safety cases, AI evaluations, and incident reporting. The point is not to shrink imagination. It is to stop imagination from laundering authority. A serious forecast can survive a ledger of assumptions, affected parties, present costs, recourse, and counterevidence. A private theology usually cannot.

The second discipline is route separation. "Space settlement," "machine superintelligence," "radical longevity," "abundance," "existential risk," and "national competitiveness" are not one claim. Each has different evidence, affected people, infrastructure, public authority, and failure modes. A launch-cost trend does not prove a political right to settle Mars; a model benchmark does not prove a right to waive labor review; a safety scenario does not prove a right to secrecy. The future-claim test should prevent one impressive graph from doing work for five unrelated permissions.

The third discipline is beneficiary replacement. Replace "humanity" with the actual institutions and people enrolled today: data-center neighbors, grid planners, annotators, moderators, artists, teachers, students, patients, public servants, shareholders, contractors, users, and agencies. If the future claim turns those people into footnotes, it is already a governance claim, not just a vision. Public-interest technology begins where the imagined future is forced to answer to the public whose present it consumes.

AI Risk and Present Power

Becker is strongest when he separates real governance from apocalyptic theater. AI systems already affect housing, employment, transport, education, health, accessibility, and justice; the Bletchley Declaration recognized those everyday deployments while also focusing on frontier-AI safety risks. The Future of Life Institute's March 22, 2023 open letter called for at least a six-month pause on training systems more powerful than GPT-4 and listed governance needs such as auditing, provenance, liability, and public oversight. Those are concrete policy questions, even when the surrounding rhetoric can drift toward cosmic drama.

The same source ecosystem has kept producing evidence claims. The Future of Life Institute's Summer 2026 AI Safety Index evaluates nine leading AI companies across 37 indicators in six domains. It is an advocacy scorecard, not a regulator order or neutral audit, but it is still useful as a public record of which safety commitments reviewers can inspect and which remain weak, missing, or unenforceable. The source lesson is Becker's lesson: do not let either industry promises or critic scorecards become prophecy. Treat them as claims to be checked against mechanisms, records, and consequences.

The book's criticism is that the drama often captures attention before the policy does. A civilization-ending machine is easier to mythologize than an underfunded benefits office, a warehouse schedule, a school surveillance system, or a data center's demand on land and power. In that sense, speculative AI risk can become a displacement machine. It does not erase present harms directly; it changes the status hierarchy of concern.

The answer is not to invert the mistake and dismiss future risk. It is to connect risk claims to accountable mechanisms. If a lab says a system may pose catastrophic risk, the claim should trigger evaluation design, incident reporting, model-security controls, red-team evidence, staged release, external scrutiny, and legal responsibility. If a vendor says a system will transform medicine, education, science, work, or public administration, the claim should trigger task evidence, affected-user remedies, procurement records, labor review, and post-deployment monitoring. In both cases, the forecast should create duties rather than prestige.

The Agent Reading

For AI agents, More Everything Forever supplies an ideological warning. Agents are attractive because they turn a preference into a plan and a plan into action. The same feature that makes them useful also makes them fit neatly inside elite futurism: delegate more, accelerate more, scale more, remove more friction. If the end is defined as maximizing an abstract future, the agent becomes a bureaucrat for an abstraction.

A responsible agent culture would ask smaller, colder questions. Who set the goal? Who can veto the action? What happens to people outside the optimization target? What records remain? What does the system refuse to do? Becker's book helps because it is suspicious of supposedly universal missions. The danger is not that an AI system is secretly divine or conscious. The danger is that human institutions may treat machine-mediated ambition as if it outranks democratic judgment.

NIST's 2026 AI Agent Standards Initiative makes this less abstract by treating agent identity, authentication, authorization, interoperability, and security evaluation as standards concerns. An agent that can act through accounts, tools, APIs, files, money, or messages needs a governed principal, scoped credentials, logs, revocation, rollback, and incident ownership. Without those records, "future-facing" delegation becomes an accountability sink.

Governance and Safety

The practical governance move is a future-claim register. For every civilizational, catastrophic, abundance, immortality, space-settlement, superintelligence, or inevitable-progress claim, record who made it, what decision it is meant to justify, what evidence would falsify it, who pays present costs, who gains present power, what public authority is being bypassed, and what review date or sunset applies. If the claim is used in fundraising, procurement, regulatory lobbying, data-center siting, release gating, hiring, or public safety messaging, it is no longer just a thought experiment.

A second register should track the concrete system underneath the story: model or product identity, training and deployment boundary, compute dependency, data provenance, safety evaluations, human-oversight plan, incident triggers, release gate, user recourse, procurement or investment incentive, environmental footprint, and rollback authority. NIST's AI Risk Management Framework and Generative AI Profile are useful precisely because they translate big claims into governance, mapping, measurement, and management work across the system lifecycle.

A third register should track public obligations. Does the system trigger AI Act transparency duties, GPAI documentation, systemic-risk reporting, high-risk obligations, U.S. federal procurement requirements, consumer-protection law, environmental review, labor consultation, accessibility law, or child-safety duties? Becker's argument becomes practical here: if a future claim asks the public to tolerate risk today, the claim should also make the relevant public duties easier to find, not easier to evade.

For public institutions, the rule should be blunt: no future claim should weaken ordinary duties of evidence, due process, labor protection, privacy, accessibility, environmental accounting, or democratic oversight. Severe future risks may justify more preparation and stronger safeguards. They do not justify letting the actors most invested in scale exempt themselves from the controls everyone else has to live under.

Regulators already have a narrower version of this discipline. The FTC's Operation AI Comply treated deceptive AI claims and AI-enabled deception as ordinary enforcement targets. The FTC's later companion-chatbot inquiry asked how companies evaluate safety, manage youth risk, and inform users and parents about risks. Those are small but important counterweights to futurist inflation: calling something AI does not create an exemption from substantiation, consumer protection, disclosure, or liability.

Where the Book Needs Care

The book's anger is clarifying, but it can flatten the field if read carelessly. There are technical researchers, civil-society groups, public servants, and harmed communities who worry about advanced AI without endorsing billionaire futurism. There are also frontier risks that deserve serious evidence-gathering, safety testing, and institutional capacity. Becker's polemic works best as a critique of power, not as a reason to dismiss all work on low-probability, high-impact hazards.

It also leaves a practical governance gap. Deflating salvation stories is necessary, but not sufficient. NIST's AI Risk Management Framework, released in 2023, its 2024 Generative AI Profile, and its 2026 concept note for a trustworthy-AI profile in critical infrastructure give one more bureaucratic answer: incorporate trustworthiness into design, development, use, and evaluation; identify generative-AI risks; build risk practices for critical infrastructure. That is less vivid than space immortality, but more useful for institutions that have to buy, audit, or refuse systems next quarter.

The book also needs to be paired with infrastructure and labor accounts. Space futures, digital immortality, and boundless intelligence can make earthly constraints feel like temporary annoyances. But present AI systems already depend on data centers, chips, energy, data labeling, moderation, cloud contracts, content licensing, call centers, schools, agencies, and users whose activity becomes training signal. A critique of futurism becomes stronger when it can name those operating conditions.

It should also be paired with a better theory of public capacity. A future owned only by firms is dangerous, but a public sector with no technical competence can become dependent on those firms anyway. The answer is not private destiny versus bureaucratic paralysis. It is public-interest technical capacity: procurement staff who can test claims, regulators who can inspect evidence, public utilities that can negotiate data-center load, courts that can understand automated harm, and workers who can challenge automation before it becomes the default.

What This Changes

More Everything Forever belongs in this archive as a study of technological belief under conditions of wealth. It asks readers to notice when the future is used to end debate, when risk language launders control, and when "humanity" becomes a mask for a small class of decision-makers.

The reviewable lesson is simple: do not let scale become sanctity. AI governance should handle near harms and severe future risks without surrendering public judgment to the people most invested in their own myth. The future is not a board seat, not a cap table, not a bunker, not a launch plan, and not an agent's objective function. It is a contested public responsibility.

The adjacent operational pages are the point where the critique becomes useful. AI system inventories, procurement records, audit trails, post-market monitoring, model-weight security, red teaming, vulnerability disclosure, liability and accountability, and impact assessments are not glamorous answers to cosmic claims. They are how a society keeps a forecast from becoming private sovereignty.

Source Discipline

This review separates five source types. Publisher, retail, and event pages establish the book's metadata and public framing. The Future of Life Institute letter, Bletchley Declaration, and Future of Life Institute safety index establish movement claims, public risk rhetoric, and advocacy scorekeeping that Becker is criticizing around. NIST, OMB, the EU AI Act guidance, EU institutional announcements, the International AI Safety Report, FTC materials, and IEA energy analysis establish current governance, safety, consumer-protection, procurement, and infrastructure context. Internal links supply site vocabulary, not independent proof.

Those sources do not prove the same thing. A manifesto proves a movement's claim, not the claim's truth. A safety index proves a scoring method and public assessment, not regulatory compliance. A scenario can stress-test imagination, but it is not evidence that one future will happen. A regulator page proves a duty, inquiry, or enforcement posture, not a system's safety. A publisher description proves how a book is framed, not that every interpretive claim in it is settled. A data-center energy projection gives a planning context, not a universal per-prompt footprint.

Dates are part of source discipline. GPAI obligations, AI Act transparency duties, delayed high-risk timelines, OMB procurement rules, NIST profiles, and infrastructure projections all have review dates and implementation windows. A current article should preserve those dates instead of turning "AI governance" into a timeless mood.

When this page proposes a future-claim register, it is not treating rhetoric as illegal or imagination as suspect. It is treating public-impact forecasting as an accountability practice. The stronger the forecast's role in investment, procurement, release timing, data collection, energy demand, or political authority, the more the source must preserve falsifiers, review dates, affected parties, and independent evidence.

This article makes no claim that any current AI system is conscious, divine, or AGI. It treats AI futurism as a human institution: narratives, capital, models, data centers, standards, laws, agents, and deployment decisions arranged into a system that can act on people before its grandest predictions come true.

Sources

Book links are paid affiliate links. As an Amazon Associate I earn from qualifying purchases.

Buy on Amazon Browse Books

Return to Blog · Return to Books