Español

Protection against the storm

Written by Pablo González and Pedro Nonay, trying to find what we can do in our adaptation to changes in world order.

Entry 8 – Protection against the “storm”.

Commandos vs. armies. Consequences of AI

February 22, 2025

There continues to be a lot of news about the actions of the new Trump´s administration. Of course, his attitude is almost frantic and leaves no one indifferent. It is clear that he is going to negotiate hard on almost everything. Therefore, there is a lot of uncertainty at all levels.

However, Trump has yet to act on the most fundamental issue of all. It is the one that will make clear the evolution of peace alternatives, economic risks, and geopolitical stability. It is about his relationship with China, creating coexistence agreements, or going to war (trade, military, or otherwise).

Nor has he defined himself completely in his relationship with Europe, except for some pre-negotiation threats, which are already beginning to be clear. He is saying that Europe will not play a role in the world if it does not assume its costs (military expenses, and many others), and that it is up to Europe to choose whether to become a “subject” of the USA or of the BRICS+. Well, he does not expressly say this last point yet, but it is what one senses. It is even clear that he wants the disappearance of the concept of “Europe”, and that he believes that some countries will be his “subjects”, and others will be “subjects” of the Chinese bloc.

It seems that what he is doing, before deciding his stance towards China, is to know which countries he is counting on in his “alignment”. But he is not “recruiting” them with incentives, but with threats. Rather than partners or friends, he is looking for “subjects” in need.

Of course, when choosing countries for your team, you take into account the resources you need. I am referring to the energy, raw materials and food that he needs for a correct autarkic balance of his bloc of countries. This is something I have already dealt with in my serie of entries called “Building the new order”, in which a summary of raw materials is given here, and another for agricultura here. It is an example of this that he speaks of Greenland for its raw materials; of the Gulf of Mexico, which, with the excuse of changing its name, he wants to turn into his territorial waters, and exploit the oil that is known to be there; and of Ukraine paying him aid with the rare earth mines.

Until Trump’s decision on his relationship with China occurs, I cannot continue my study of what it is advisable to do about it. For now, it is still valid to try to seek as much room for maneuver as possible, and the greatest speed of response.

*****

For that reason, I am devoting this entry to talk about something that does not depend so much on Trump’s actions. I will address the consequences of AI developments. They are all the rage these days, because of the Chinese competition from DeepSeek, which has generated turbulence because of the unexpected, and because of the quality of the tool

I wrote about this in May 2023 here Much progress has been made in development since that time, but the central ideas stated there remain, so I am not repeating them here. 

Perhaps the novelty is that everything is going faster than expected. Tomás Pueyo has written a very good article summarizing the state of the art. He concludes that the AGI moment can be achieved in less than 5 years (you can read it here). 

I make the precision that this AGI moment occurs when the AI is able to do anything better than the best of humans. That moment being very important, that is not where the AI is better than the whole of humanity, where it knows absolutely everything, nor where the AI can reprogram itself, without human intervention. That would be the Singularity, and it will take longer.

As is the purpose of this serie of entries, I will try to see what we can do to adapt as best as possible in these times of transition.

First IAs. Intellectual works at risk.

The first AIs (which are the ones that already exist) are not yet at the AGI moment (where the AI is better than any human being), but they are already starting to be useful to replace the average human being in many tasks.

As a result, AI is already replacing jobs that not long ago we thought had to be done by a highly trained human being.

This is what is often referred to as the “blue collars and white collars”. In that image, the “blue collars” were the “factory workers”, and they have long since been replaced, for the most part, by robots and technology. In other words, in terms of mass labor, they disappeared, just as manual farmers did long ago after the invention of tractors and combine harvesters

Now comes the replacement of jobs for the “white collars”. That means that many office jobs (even high-level ones) will be replaced by AI. That is, almost all routine, procedure-based work will be done better by AI than by humans.

This is already happening. We don’t realize it, but major companies are already applying AI for it. From Google to Spotify they use AI to give us the best response.

I mean that the first uses of AI will not be “disruptive” to the normal human being. They will be a way for existing companies to provide a better service in “transforming materials and energy into things and services“. Which is still the really important thing, because everything we humans need is based on transforming those materials and energy into what we want.

Institutions will be diluted. 

Precisely because of all this new offer of services that I have mentioned above, there will be a great problem of loss of usefulness of all the institutions that we have been accepting as “essential”. I am referring to institutions that may include nothing less than the concept of “nation”, even that of democracy. Of course, this will also happen with institutions of a lower rank, such as the UN, WHO, …

This will happen gradually, although it is inevitable. The fact is that all institutions (important or basic) are nothing more than the result of finding a way to adapt ourselves to the necessary management of resources to achieve the best and most pleasant coexistence of human beings. 

If it happens that the tools available to us to do so change radically (and AI is a radical change), the result is that we will have to organize new institutions that allow us to better manage resources in pursuit of the same objectives.

Individuals, governments and companies will have to adapt to this new scenario. We will have to focus on the skills that we will still be able to do better than AI.

Current AI knows how to do (almost) well what depends on existing knowledge. When the AGI time comes, that itself will undoubtedly do it better than any human being. 

What today’s AI does not do (not even that of the AGI moment) is to create knowledge, nor what is usually called “out of the box” thinking. That is the gap we humans can still work in

And, that AI can’t create knowledge is not me saying it. The AI itself says so. It says it in a long conversation I had with ChatGPT last February 9, precisely as one of the previous tasks to prepare this entry. In one part of the conversation the AI says:

• Limitations in the “Creation” of New Knowledge
  
  
  • I do not have mathematical intuition or creativity in the same sense as a human. Great mathematicians like Euler or Ramanujan made advances based on intuition, experimentation and abstract thinking.
    
    
  • Although I can generate novel ideas, my knowledge is based on what already exists. I do not “think outside the box” in the same way as a human researcher working at the frontier of knowledge.
    
    
  • I cannot do physical experiments or generate new empirical data, which limits the validation of theories in applied science.
    
    

You can read my complete conversation with the AI in this link that I have posted on my website. However, if you are interested, I recommend you to do it after having finished reading this entry. It is probably more useful to do it this way (I could not).

For this reason, we humans must strengthen the skills that enable us to create knowledge. In this sense, what Harari said in his book “21 Lessons for the 21st Century” is very applicable. In chapter 18 of that book, Harari addresses a crucial issue: education and the skills taught in schools. He says that what is taught today may be completely useless in the future, because it will be done automatically by technology. He argues that the skills that need to be reinforced in teaching are the “four Cs”: critical thinking, communication, collaboration and creativity. The rest of life will be a matter of constant learning and adaptation.

I mention that Pedro tells me that, above the four Cs that Harari talks about, we should always think about “the senses”: seeing, smelling, hearing, feeling, …

Commands. We must support them

From what has been said above, I think that, using the military simile, we humans have to abandon the technique of armies, and we have to focus on that of commandos.

I say that because the technique of armies is based on a mass of soldiers and middle managers correctly following the instructions of the generals. And that will be done better by AI, with its drones and robots, than by humans. Not only in the military world, but in any activity. In other words, tasks with instruction manuals are for AI.

On the contrary, the tasks that require improvisation of multidisciplinary teams, formed ad-hoc for a very specific and conjunctural objective, are the ones that humans will still do best. These are the tasks performed by commandos. It is a team of a few people, with different skills, with a very specific mission, without an instruction manual, and with a lot of freedom to act.

There will be many possibilities of professional success for those who dedicate themselves to it. They will be people outside of conventional structures, who will work every so often on different missions and with different partners (although this is not the case for the team chosen for the photograph above).

The existence of these “commandos” should be encouraged. They should be given more social prestige and more remuneration than people who always work in the same place and always do the same job (who are the ones who will be replaced by the AI).

Unemployment will increase. Solutions are needed.

As I said, “army” type jobs will gradually be replaced by AI. And not many people will have the capacity to adapt to “commando” type jobs.

This means that the number of people out of work will increase significantly, although this will happen in the long term, as the substitution will be gradual.

In the short term, as substitution will be gradual, it may happen that the drop in demographics of a country will allow AI to do the work of people who are no longer available to do it. This would not increase unemployment, nor would it decrease the country’s total productivity.

What is clear is that at some point unemployment is going to rise sharply. And it will be endemic. That is to say, it will be normal not to work. Therefore, it is necessary to organize the mechanisms so that those people who do not work can live. And it is not only a question of them having an income to organize their lives, but also of them having ways of using their time without going into depression because they feel useless or displaced.

Thus, a whole problem of social organization appears. 

Regarding the income of these people, one has to ask: who pays them? It could be, as up to now, that it is done by the state using taxes levied on AI productivity. It could also be that they are paid directly by the companies that use these AIs.

As for the way to give meaning to the lives of these people and avoid their depression, we will have to encourage their artistic activities, social cooperation, creativity, … And give them social respect for these activities. In addition, it will be necessary to deepen (even more) in that they can occupy their time in entertainment, sport, … It will be something like the “Leisure Society”. It sounds nice, but it is more delicate than it seems.

Singularity: Techno-communism? 

I make a comment for when the singularity is reached, which will be in later times. That will be the time when AI will know absolutely everything, and there will be no doubt that it will be able to do everything better than any human being.

At that point, the logical thing will be that no human will work (in the sense of producing), and that everything will be done by AI, because it will do it better. This even affects the work of running the country (or the world).

If at that point the AI still accepts that its role is to do everything to help humans (which is not guaranteed that it will decide to do that), what will have to be redefined is the type of social contract that will apply. That should be in the Constitution with which the AI is programmed.

It could be that it would be something very similar to communism. Centralized power with absolute knowledge of everything. Contrary to the experience of the USSR, the power (the AI) will receive all the information, with guarantees of veracity, and will have the capacity to process it correctly. That is to say, there will be no possibility of concealment due to corruption, nor of failures due to bureaucracy, nor of wrong decisions due to erroneous data processing. Under these conditions, it is clear that better decisions will be made from this highly centralized entity than from smaller entities (be they companies or governments), which would have more incomplete information

It could be called “Techno-communism”. And, since nobody would work (in the sense of producing wealth), the machine will have to decide how to distribute goods and services among humans. And it will have to choose whether to incentivize some kind of human behavior with more goods.

This sounds crazy, but we are likely to see it while we are still alive. However, since it refers to a longer time frame, it is not the type of situation targeted in these entries, where we look at what we need to do to prepare for the imminent.

Conclusions.

AI is advancing faster than many expected. In the short term, its impact will focus on replacing routine jobs and improving processes. As it evolves, its reach will go much further. This raises fundamental questions about employment, social structure and the role of institutions.

What is clear is that the days of jobs based on following instructions are numbered. Anything an AI can do better than a human will eventually be done by the AI. Therefore, the only viable option is to focus on what AI does not yet do well: creativity, improvisation and the ability to generate new knowledge. It is no coincidence that the key skills for the future are Harari’s “four Cs”: critical thinking, communication, collaboration and creativity.

From a practical point of view, the “commando” strategy is the most appropriate for those who want to stay in the game. These are small, multidisciplinary teams working on specific missions, without instruction manuals. Here, humans can still make a difference.

But, we have to be realistic. Not everyone will be able to adapt to this model. And as substitution will be gradual, there will be a period in which the demographic drop can compensate for the lack of employment. Sooner or later, however, the time will come when structural unemployment will skyrocket. There will be an inevitable debate: who pays those who do not work? How can we prevent part of society from falling into depression or chaos?

Further down the road, when the singularity arrives, the landscape will change even more. If AI continues to work in favor of humans, it will have to decide how resources are managed and what kind of society we want. We may be approaching a “Techno-communism” model, where AI has all the information and makes the most efficient decisions. It sounds radical, but it is a scenario we could see in our lifetime.

*****

Tips for our adaptation.

As I have been saying in previous entries, I do not write this because I consider myself to be a holder of the truth, nor to convince anyone. I do it in my internal process of creating my opinion, in order to make my future decisions.

Therefore, I don’t think it would be honest not to share my conclusions. This is what I do below, but remember that they may be wrong.

For individuals.

Depending on our age and abilities, each of us will have to decide whether to choose one of the following paths for the near future:

• Trying to live as usual, and crossing our fingers that changes do not occur before our life is over.
• Accept the changes that will come with the “advent of AI”, and prepare for them.
• To dedicate ourselves to live outside that world, and with our own means. A bit like the pioneers of the conquest of America.
• Go underground, and be part of a kind of “resistance” within the AI-ruled zones.

I choose the second option. In addition, and before the Singularity arrives, I will try to improve my skills in order to be useful in some “command”, as I said above. After that, I will try not to “displease” the AI and do what it tells me to do. That is, I will try to be useful while I can, and accept what happens later (I will be old enough to get into trouble). Besides, I will try to enjoy myself along the way.

For companies.

It is very important that they analyze whether their service and internal organization will be useful in the different stages of AI implementation.

At first, they should use it to improve their productivity. When they see that the time is approaching when the AI will be able to provide the service without their existence, they should organize their orderly closure, before the moment of bankruptcy due to inefficiency arrives. In other words, they must avoid repeating what happened to Kodak for not adapting to the digital world (and this time the effects will be much greater).

For governments.

If they really think about being useful to the governed (which is something I doubt in many cases), they should recognize that their service will cease to be useful very soon. And they should accept its gradual dissolution. Abandoning the attempt to control in different aspects, as the AI becomes able to do each thing better than those governments.

Unfortunately, though, I don’t think they will give up their lust for power easily, which will lead to many conflicts and inefficiencies (but that’s something we are used to from governments).

*****

As always, I welcome comments on my email: pgonzalez@ie3.org