The Dream Carrot: AI, Technological Society and the Final System

The Dream Carrot: AI, Technological Society and the Final System

Written on 12 September 2026.

Maybe the promise that artificial intelligence will give ordinary people unprecedented freedom belongs to a period that is ending. We are being told that the machines will become more powerful, cheaper and eventually available to everybody. But why should that mean that the people who own the industrial system will give everybody independent control over its most powerful technology?

I think three threads help explain what we are seeing: the experience of having lived through the 1970s and 1980s, Theodore Kaczynski's Industrial Society and Its Future, and the prophetic writings of the KJV, particularly Daniel and Revelation. They approach the question differently. Personal experience gives us a comparison. Ted's argument describes how technological dependence can develop. Scripture describes the final concentration of worldly power and its judgment.

Taken together, I think they make the promise of a decentralized AI paradise look much less convincing. The technology may keep advancing while the ordinary person's independence goes backwards.

Remembering a different atmosphere

Having lived through the 1970s and 1980s, I remember another atmosphere. The message was more about building society. Work, learn a trade, become an engineer, develop something, and get something for it. Industrialized nations were supposed to advance, and ordinary people were supposed to have a place in that development.

That is a description of the atmosphere I remember, not a claim that everybody had a good life or that governments were harmless. But the expectation mattered. Human effort belonged in the picture of the future. People were needed to build it.

Now I see much more language about sustainability, management, adjustment and limits. There is a difference between being encouraged to build a better life and being told that your life needs to be managed so that it fits within a system's requirements. The second message can exist alongside impressive technical progress. The system develops, but the individual is increasingly asked to adapt.

The UN's stated programme still includes economic growth, productive employment and industrial development. So it would be inaccurate to say that its formal goals simply abolished work or development. My observation concerns the changing relationship between the person and the future being offered to him. More development of the machinery does not necessarily mean more room for the person.

Someone who only knows the present arrangements may regard them as normal. Someone who remembers life before constant connectivity has another point of comparison. He can remember that a society could function without so many accounts, electronic permissions and connections to remote organizations. Younger people can understand that too, but living through the change makes the comparison less abstract.

That experience can also mislead us if we assume the old promise must return. Perhaps some people expect AI to empower them precisely because earlier computers did. They remember what technology once gave them and assume that this is what technology always does.

When the system needs fewer people

An industrial society needs workers, engineers, technicians and organizers. When their cooperation is necessary, they have something with which to bargain. They can withhold labour, take their skills elsewhere or demand a share of what they produce.

Democracy and freedom have more foundations than economic usefulness. They were also fought for, and people possess value whether an employer needs them or not. But economic usefulness gives people practical strength. If the system becomes less dependent on their work, that particular source of strength diminishes.

This is where AI changes the question. It promises to automate intellectual work as well as physical work. Even the programmer, who once seemed to stand on the winning side of automation, can become a person whose work is being automated.

Follow that possibility far enough and the conclusion is severe: they do not need people to be free because they do not need people's independent effort anymore. They do not need people, period, in the productive role that previously made the population indispensable.

I am describing the possible endpoint of the process. Today's economy still needs enormous amounts of human work. The concern is that an establishment able to produce, administer and enforce with far less human participation would have less practical need to bargain with ordinary people. It might still feed them. It might entertain them. That would not necessarily give them control over their lives.

Ted addressed almost exactly this possibility in paragraphs 172–174 of Industrial Society and Its Future. He considered machines capable of doing human work, then separated two possible outcomes: decisions passing effectively to machines, or human control remaining with a small elite. In the latter case, ordinary people might retain private machines while losing significance within the larger system.

That distinction is useful. A person can own a computer and still have almost no influence over the machinery that determines his livelihood. Personal gadgets do not answer the question of who controls production, administration or force.

Ted's scenarios are conditional arguments, not proof that a particular group has already decided what to do with humanity. Their importance is that the problem does not require machines to rebel. People can remain in authority while using machines to make other people less necessary.

The Silicon Valley dream carrot

The Silicon Valley freedom story is familiar. Put powerful computers into people's hands. Let them experiment. Let them write software, communicate and create. The next generation of technology will give them even more freedom.

GNU and Linux show that practical freedom in computing can be real. The ability to inspect, modify and share software matters. But the existence of that achievement does not establish a rule that every later technology must follow the same political direction.

The independent developer was useful. Give him a programmable computer and tools, and he could help build the technological world. If AI increasingly performs that work, why assume the same incentives remain? The people who were encouraged to build freely may be helping to construct a system that needs less of their freedom in the future.

This is what I mean by the dream carrot. Keep developing the technology. Accept the concentration of money and computing power for now. Eventually, we are told, all that power will come back to the individual as cheap, private, unrestricted intelligence.

Maybe some of it will. But the promise can keep people committed even while the institutions are moving in another direction. It does not have to be a consciously invented deception. The people repeating it may sincerely believe it. A dream can serve the system without every person selling the dream understanding what it does.

The crucial assumption is that cheaper intelligence means independent people. Cheap access could instead mean that everybody becomes a customer of the same few systems. A service can be affordable, astonishingly capable and still subject to conditions that somebody else controls.

So the question is not simply whether AI will become cheaper. It is whether ordinary people will gain durable control over it, including the ability to use it privately, change it and keep it when an institution disapproves.

An AI card still comes from the industrial system

Ted distinguished small-scale technology from organization-dependent technology in paragraphs 208–210. The first can be maintained by small communities without outside assistance. The second requires large social and industrial arrangements to continue.

A knife or a hammer illustrates the first kind when a community possesses the skills and materials to make and repair it. An advanced AI chip illustrates the second. You cannot reproduce its industrial foundation in a farm workshop. You need specialized design, semiconductor fabrication, equipment, materials and a network of suppliers.

Even AMD depends on other organizations. Its annual report describes reliance on outside wafer foundries and assembly and testing facilities. This is not one self-sufficient company producing intelligence from raw materials. It is a vast industrial arrangement behind a small object.

That remains true if AI becomes a card you put into your own computer. Taalas describes technology that turns models into custom silicon. Such a card could be a useful possession. If it works offline without recurring permission, it could protect privacy and reduce dependence on a cloud service. But distributing finished cards does not distribute the ability to manufacture them.

The model may also be partly or largely fixed by the hardware design. How much an owner can change would depend on the implementation. There is no reason to assume that intelligence embodied in silicon will be as replaceable as a software file. A fixed model might remain useful for years while still reflecting decisions made before the purchaser ever received it.

Local operation and independent production are different things. You can possess the finished product while depending on large organizations for every replacement and every new generation.

Ordinary computers already have this dependence. What makes advanced AI significant is that it can extend the dependence into activities we previously supplied ourselves: writing, reasoning, programming, planning and judgment. The machine becomes more helpful while the industrial organizations behind it become more important to daily life.

Small groups can develop useful models, and people can adapt existing ones. But independently reproducing the most capable systems and their hardware is another scale of undertaking. In that sense, frontier AI is an extreme expression of organization-dependent technology. It is centralized technology on steroids, even when some of its products sit on a desk at home.

Twenty years and the authority to decide

The proposed twenty-year prison penalty gives this concern a concrete example. On 3 September 2026, Bernie Sanders and Greg Casar announced the forthcoming Ban Artificial Superintelligence Act. Their announcement proposes a permanent prohibition on superintelligence, a pause on advanced AI development and a new federal regulator. It describes penalties of up to twenty years for individuals attempting to violate or circumvent the restrictions.

The official summary includes systems that can easily be modified to reach its broadly described human-level capabilities. It also proposes federal supervision of dangerous capability removal and destruction of prohibited systems.

That is more specific than saying that modifying any LLM will get you twenty years in prison. But it raises an obvious question about the room left for independent experimentation. If a model's capacity to be modified can bring it within a prohibition, control reaches beyond how a finished service is used.

In my opinion, attaching a penalty of that magnitude to this field, together with comparisons to nuclear weapons, can operate as an extreme scare tactic. A person does not have to be prosecuted to decide that touching the technology independently is too dangerous for him.

The sponsors present their proposal as protection against powerful AI companies. That is their stated purpose. My concern is the possible result: a field where central authority decides which capabilities may exist and where individuals are frightened away from developing them. Protection from corporate power and expansion of government power can occur in the same proposal.

Nor does every safety restriction automatically benefit every large company. A genuine ban could damage those companies too. But complicated approval systems and uncertain boundaries can leave well-funded organizations better able to continue permitted work than an independent developer with no legal department.

The industrial structure makes regulation easier to apply. Governments can exert pressure at chip suppliers, fabrication plants, major laboratories and cloud providers. The ordinary person cannot simply replace that entire chain if access is restricted.

This is why the dream of unrestricted AI for everybody deserves scrutiny. The same future being advertised as liberation is already being discussed as something requiring extraordinary prohibitions and centralized supervision.

Useful to the person, powerful against the person

AI can help an ordinary person write, learn or solve a problem. I do not think that proves its overall effect will favour him. A technology can provide real individual benefits while increasing the establishment's power much more.

Think about surveillance. Collecting information is one task; interpreting immense quantities of it is another. AI can help connect records, identify patterns and flag people or behaviour for attention. Its errors do not make that power harmless. An incorrect judgment becomes more consequential when organizations act upon it automatically.

Think about money. Programmable restrictions on spending do not require AI, but AI could be added to systems that decide whether transactions are suspicious or permitted. If identity, purchasing and automated judgments become tightly connected, exclusion can become an administrative operation. This is a possible arrangement, not a claim that every digital payment already works this way.

Then there is unmanned warfare. A drone carrying an explosive can bring lethal force directly to a person. AI can add recognition, navigation and targeting capabilities to weapon systems. The ICRC describes autonomous weapons that select and apply force after activation without further human intervention, and warns about loss of control over particular targets.

The human being on the receiving end experiences none of the Silicon Valley promise. He experiences a machine that can find him and act against him. Machines that reduce the need for workers can also reduce the number of people required to monitor or attack others.

That combination is what concerns me: less dependence on ordinary people's cooperation, together with greater capacity to administer and coerce them. It is unnecessary to imagine one secret designer behind every invention. Organizations can pursue efficiency, security and advantage separately while their technologies converge into a much more controlling environment.

Ted himself argued in paragraphs 152–153 that restrictions could accumulate through measures presented as beneficial, without beginning from an explicit totalitarian intention. In paragraph 173 he described growing dependence making machine decisions increasingly difficult to refuse. The problem can develop through ordinary institutional decisions long before anybody announces a final system.

The final system in Daniel and Revelation

Revelation describes authority reaching into economic life:

And he causeth all, both small and great, rich and poor, free and bond, to receive a mark in their right hand, or in their foreheads:
— Revelation 13:16 (KJV)

And that no man might buy or sell, save he that had the mark, or the name of the beast, or the number of his name.
— Revelation 13:17 (KJV)

The passage describes a mark, allegiance and exclusion from buying and selling. It does not name a computer architecture. But it gives a reason to pay attention when ordinary economic participation becomes dependent on identification and permission from a system beyond the individual's control.

AI could help administer such arrangements. That is an interpretation of technological possibility, not a statement that the Bible explicitly predicts large language models. Neither an AI card nor a digital account is, by that fact alone, the mark of the beast.

The convergence is larger than one device. Daniel describes a kingdom with overwhelming reach:

Thus he said, The fourth beast shall be the fourth kingdom upon earth, which shall be diverse from all kingdoms, and shall devour the whole earth, and shall tread it down, and break it in pieces.
— Daniel 7:23 (KJV)

Revelation describes power being handed over:

These have one mind, and shall give their power and strength unto the beast.
— Revelation 17:13 (KJV)

This is compatible with a final concentration of institutions that previously appeared separate. It does not require the machines to overthrow their owners. Human authorities can give their power to the beast while technology supplies means of administration and enforcement.

The religious dimension matters too. AI enthusiasm can become a kind of false faith when intelligence and technical power are treated as the answer to humanity's condition. The machine will solve our problems, provide abundance and tell us what to do. But greater capability does not establish righteousness, and dependence is not salvation.

Revelation joins worship to apparently overwhelming power:

And they worshipped the dragon which gave power unto the beast: and they worshipped the beast, saying, Who is like unto the beast? who is able to make war with him?
— Revelation 13:4 (KJV)

The admiration is bound up with the belief that nobody can oppose the beast. That is a different thing from examining whether the power is good. A population can become impressed by what a system can do and stop asking what it demands in return.

The speaking image in Revelation 13:15 is also striking in an age of artificial voices and simulated personalities. But resemblance alone does not establish that AI is the image. The verse includes enforced worship and death for refusal. Those features belong to the passage and cannot be replaced with the mere observation that a machine can talk.

My concern is that technological dependence, concentrated authority and misplaced faith could prepare people to accept arrangements they would once have rejected. That is where I see the relevance of prophecy to the direction of technology.

Great power for a short time

There is another connection, but Ted's argument needs to be stated accurately. In paragraph 195 he suggests that dictator-controlled industrial systems might be more likely to break down because of inefficiency. He treats that possibility as favourable to his aim of ending the industrial system. Examining that observation does not require accepting his violence or his revolutionary programme.

He did not predict the biblical tribulation or establish that every dictatorship is short-lived. But the narrower point remains interesting: more concentrated coercive power does not necessarily produce a more durable society.

Revelation gives the beast a limited period:

And there was given unto him a mouth speaking great things and blasphemies; and power was given unto him to continue forty and two months.
— Revelation 13:5 (KJV)

Forty-two months is three and a half years. Daniel supplies the corresponding expression:

And he shall speak great words against the most High, and shall wear out the saints of the most High, and think to change times and laws: and they shall be given into his hand until a time and times and the dividing of time.
— Daniel 7:25 (KJV)

I understand these passages as describing the three-and-a-half-year period of the final beast's power, associated with the great tribulation. This does not mean that all the institutions and technology used by that system come into existence within those years. The preparation could take much longer than the final rule.

That is a disturbing possibility. Generations can build the infrastructure, surrender pieces of independence and anticipate a wonderful future. The resulting concentration of power can then serve a brief and terrible final order.

But the reason for its end in Scripture is not merely administrative inefficiency. Daniel continues:

But the judgment shall sit, and they shall take away his dominion, to consume and to destroy it unto the end.
— Daniel 7:26 (KJV)

And Revelation says:

These shall make war with the Lamb, and the Lamb shall overcome them: for he is Lord of lords, and King of kings: and they that are with him are called, and chosen, and faithful.
— Revelation 17:14 (KJV)

Ted supplies a secular argument about dependence and possible breakdown. Scripture supplies the final authority and outcome: the Lamb overcomes them. The parallel concerns the limits of worldly power, not an identical explanation or timetable in both writings.

What the promise leaves unanswered

This is why I think the three threads belong together. Remembering the earlier atmosphere helps me notice the change in what society seems to expect from people. Ted's analysis helps explain how increasingly capable technology can diminish independence. Daniel and Revelation describe a final order in which power is concentrated, economic participation is restricted and that power is judged.

The connections do not give me a date for the end or prove that every new technology belongs to the final system. They give me a reason to question the story that technological advance must eventually produce freedom.

Actual developments could challenge the dream-carrot argument: durable access to capable models, practical freedom to modify them, offline ownership and fewer institutional restrictions would matter. We should look for those things themselves. A promise that they will arrive eventually is not the same as receiving them.

I think many people still want the future they were taught to expect. They want the powerful personal computer, the unrestricted tools and the reward for helping build the world. Maybe the promise continues to attract them after the system's need for that kind of person has begun to decline.

An AI card on the desk would not settle the question. The question is what happens to the person outside his house: whether he can work, buy food, move around and refuse demands without being excluded from ordinary life.

If intelligence becomes abundant while permission becomes scarce, the dream carrot has not delivered what it promised. The machine may be more capable than anything we imagined in the 1980s. The person may have less freedom than he remembers.

References

AI Disclosure: Parts of this page may have been created, edited, or assisted by artificial intelligence tools (such as ChatGPT or other language models). All AI-assisted content is reviewed by a human before publication. For questions, contact the site administrator.