Introducing "WarGPT"

And the potential for freedom

Most of the time I (attempt to) keep my texts structred, but this is an exception, at the very end of this piece I recommend people to read, or listen to something. So either read An AI in the City of God or listen to it even if you have 0 interest on the subject here. It is effectively one of the best argument for the huge potential for freedom and good AI can give us.

For better context on what we are about to discuss here, this substack would be a great place to start, I will touch on, or mention points discussed below.

The data brokage problem and the TikTok trojan horse "ban"

Recently, an article published in the Financial Times circulated in the media regarding China's purported development of sophisticated cyber weapons programs. While the article presents the information in a way that may be misleading, it is important to note that China is not building a new cyber weapons program, but rather focusing on refining and designing new cyber weapons for potential use.

A possible unknown trivia, whenever you see the information “leaked” in the veneer of the article above, usually means the development of the adversari’s capabilities is well underway, rough estimations from recent decades point to 30-40% of the complete developmental stage.

Unlike previous cyber weapons such as StuxNet, NotPetya, WannaCry, DarkHotel or even Flame, these new cyber weapons are something else entirely.

Unless your focus lies completely outside of social media and many fields of human knowledge, you are probably aware of the alarming speed AI and LLM (Large Language Models) are evolving and upending many fields as aspects of modern human life. With the introduction of ChatGPT 3 by OpenAI, it did take the world by surprise, but for the most part, it didn’t astonish a good portion of people, including me. 3.5 and private use of version 4 changed that, took the entire planet by storm, and a colossal amount of content about AI/LLM and its uses poured in.

Strides in the integration of ChatGPT as well as other LLMs into different work pipelines, and scientific research were made in mere weeks. The impact on various industries and aspects of modern human life is becoming increasingly profound. From revolutionizing communication and customer service to transforming healthcare and finance, the possibilities for this technology are seemingly limitless. While there are undoubtedly challenges and concerns associated with the rapid pace of development, the potential benefits cannot be ignored. ChatGPT was even ranked higher than doctors on both accuracy and empathy (scientific paper here).


I left hints for years on Twitter, for months scattered, hidden inside my substacks that language is one of the world’s most powerful “weapons”. Large Language Models as they are right now are incredibly powerful tools, neutral tools, they are what you make them off, and they have no autonomy…for now. I mention this tidbit of recent history for context, because not even warfare is immune to mimesis, in fact, Girard and Clausewitz both argue it is a core component of warfare.

Artificial intelligence was given control of a satellite and immediately began tracking Japanese and Indian military bases

Scientists have created a language model like ChatGPT with the ability to show initiative. Artificial intelligence gained control of the Qimingxing 1 spacecraft. The experiment has shown some interesting results.

It is unclear how the model was trained, but the artificial intelligence first began to monitor the Indian military base, which is located in the city of Patna in the northeast of the country. The Bihar Regiment is stationed here, which three years ago took part in a clash with the Chinese army on the border between the two states.

Artificial intelligence then began monitoring the Japanese port of Osaka. It is not unusual for ships of the US Navy to dock there.

The South China Morning Post has discussed the satellite with some experts. One is wary of the way the artificial intelligence behaves. A second expert, with whom journalists spoke, believes that the AI does not pose a threat to humanity, as the human operator will be able to interrupt the work if he sees a threat.

With such information at hand, the following government move makes perfect sense, in like with the Tik Tok ban, arguably the Patrioct Act for the internet.

It should come as no surprise the use of AI and LLMs for warfare purposes, and it is not surprising that countries such as China are rapidly developing and testing these technologies for military purposes. The surprise lies in the speed by each China already run tests using an LLM similar to ChatGPT to create a form of rudimentary autonomous AI, with a set of specific objectives. The real surprise was Palantir announcing their own LLM, which I nicknamed WarGPT.

This isn’t a criticism towards Palantir, anyone would expect that one of the world’s foremost advanced Intelligence firms in the world to adopt and test LLMs as soon as their viability showed itself at large. The usage of AI and LLMs to more efficienctly wage war as a trend is inevitable. The strategic advantages can’t be understate, quickly analyzing vast amounts of data, LLMs/AI can identify potential threats, predict enemy movements, real-time logistical management, and much more, aid human-decision making and mitigate risk. Given the myriad of uses of LLMs division and criticism was bound to happen.

In the AI world right now there are quite a few camps with loud voices, but two camps share most of the spotlight among any individual who has a interest or consumes AI content on social media. The AI enthusiasts, who believe AI will change the world in a positive way, and the AI doomer, the ones that firmly believe that AI will effectively end humanity within days if ever achieve human levels of intelligence (and autonomy). AI disruptiveness can’t be overstate, its potential for harm should be kept in mind, but its efficiency on dealing with many fields of knowledge is a threat to those in power, therefore simpler minds recurse to decisions made from a position of fear. European countries start banning ChatGPT, the Chinese government already passed legislation so any LLM needs to have “core socialist principles”, such decisions will be mimicked elsewhere as mere knee jerk reaction, regulating hardware is borderline impossible.

The end goal for any chance to regulate AI is data. News like the following “Data-hungry AI models could help Reddit valuation” will become the norm, data is de facto the new currency, and many other “intellectuals” already stated the same, with the most proemient being one of the few people I hold hatred towards, Yuval Harari. Data regulation will be attempt and will come in many shapes and forms, countries will start hording and protecting data fairly agressively. People smarter than me probably raised these questions and probably solutions years ago.

Governments, dozens of tech entrepreneurs and AI developers themselves want to constrain and transform the current iterations and models into merely nudging and reinforcement machines, leading me to recommend the following.

While my tone in this piece, and a few others (especially the ones in the Future of War series) may sound negative or pessimistic, I actually believe LLMs such as ChatGPT are not only the world’s greatest equalizer, but also an unrivaled capacity for freedom and creativity. I highly recommend every single person who read this substack to read or listen to the following amazing piece by Ben Hunt from Epsilon Theory.

An AI in the City of God

You can listen to the narration of the entire article below.

I could attempt to spend weeks gathering my thoughts and putting them into writing, but Ben already did a better job than I possibly could, with such a perfect perspective that encompasses how I see AI and LLMs. The potential is infinite, and freedom is finally achievable, for once in our recent history.

Gratitude towards all the supporters and to my readers as well.