[EDIT] - I decided to improve existing articles instead of writing new ones. This post was originally published a few years ago on a separate blog called “The Matrix of Deception”, which focused on information warfare and propagation. I realized my thoughts were not clear enough, so I read numerous books, unpublished the articles, and merged the blog with Kernel because they essentially discuss two sides of the same coin.
In the meantime, AI became a top online contributor, and information warfare worsened. Also, my PhD supervisor tortured me into decent English writing (just joking—I was bad, and now I’m slightly less bad). Edited in July 2025.
A personal story
Back in 2017, I was a quantitative analyst for a bank in New York City. My job was to provide traders with indicators on currencies and commodities, so they could make more money and get better bonuses. At some point, they gave me the following task: quantify, however possible, geopolitical risks and find potentially breaking news that could induce massive shifts in the assets of interest. This was before large-language models became a thing, and the NLP tools at my disposal were rudimentary. I also considered the assignment pointless and had little fondness for the bankers in suits (though that’s beside the point of this piece).
More importantly, I did not honestly believe my indicators, however sophisticated, were going to be of any use. I was essentially trying to predict the unpredictable, as any potentially impacting event belonged to the potential black-swan domain, or at least carried an intractable probability distribution, to use mathematical jargon. Matter of fact, traders were primarily interested in whether or not the dollar could lose its status as a safe-haven asset1.
One day, Mexico City was hit by a mild earthquake, which is not a big piece of news if you are familiar with the region. Yet, all the traders interested in sovereign debts and currencies started trading the information. A mild earthquake in Mexico City, not a deadly severe one, is pretty much as big an event as a bar fight in Belfast. Still, all those folks looked for a way to integrate this information into their models and predict, whether or not, it would have a significant effect on their portfolio. Essentially, a piece of information hit them, they processed it, and gave the market their output once they added their interpretation.
I won't lie, telling you that this day was a revelation in my life and that I started writing about information warfare. At this stage, I just thought markets were highly dysfunctional and found even better reasons to love Benoit Mandelbrot and Nassim Taleb2. Still, it was an exciting experience, and something I was not going to forget.
It was the first time I witnessed people professionally acting on a piece of information. More importantly, they were doing so in a way that could influence people’s lives on a large scale. Simply put, in most critical domains like war, intelligence, or the market, information is likely the core resource you want to have unfair access to. Otherwise, insider trading would not be illegal, and intelligence agencies would not exist. Now more than ever, people work hard for good pieces of information. Now more than ever, good information is harder to find than a needle in a haystack, because noise drowns it. Sometimes this noise is a collateral effect of everyday human activity, but it can also be a weapon malicious actors use. This is what these articles on information warfare are about.
In the end, what is that trader's story about? It's not about the unpredictability of geopolitical risks or the silliness of trading every single piece of news, as they are mostly noise. It's about a bank investing a considerable amount of money for its traders to receive every single news in real-time, on multiple screens, from various platforms, and to act on it as soon as it's happening. Their main goal here was to have privileged and unfair access to information.
Information propagation
Over the past three decades, the way information spreads has transformed more than in the previous twenty centuries. The last one hundred and fifty years have seen three major horizontal, structural changes: electricity, the Internet, and artificial intelligence. Each of these has enhanced our ability to send information and expanded the size of our communication networks. By enabling humans to send information at scale and continuously, it has turned the world into a giant data jacuzzi where each bubble popping can be of crucial importance.
Lastly, social media has moved humanity from top-down, centralized, state-controlled information channels to global, decentralized networks, where a king’s son enjoys the same features as any random iconoclast eager to spread anti-idol propaganda. This is an underestimated feature of social media—and of most other horizontal, global digital tools: they grant identical capabilities and access to new information whether you’re a billionaire or a broke troll. Jeff Bezos doesn’t see the latest Middle-Eastern news any faster than you do; in fact, he may catch it later because he’s busy fending off backlash over his lavish Venice wedding.
Not only has social media revolutionized the mechanisms of information propagation, it has also multiplied its speed. There is more information than ever, moving along more pathways and with less friction. Because social-media users mostly act as passive transmitting relays, information now travels seamlessly between remote corners of the globe. This is a major structural change in our way of life, and I suspect it underlies the lack of purpose and meaning many people report. The large-scale psychological effects belong in other articles, but constant information flooding keeps our consciousness from letting deep purpose and emotion surface, and it prevents our creative mind from entering long, uninterrupted work—far from FOMO, trends, and outside influence, and closer to what ultimately matters: Our inner feelings and our fidelity to our values.
Society is volatile, and social networks make it even less predictable. No one —not even the sharpest mathematician—can forecast what will go viral and what will vanish in the internet’s shadows. We learned that lesson decades before Twitter: on 30 October 1938, Orson Welles’s radio drama “War of the Worlds” aired as mock news bulletins. Listeners who tuned in late missed the disclaimer and believed Martians were attacking New Jersey; phone lines jammed, some fled their homes, and newspapers headlined the “mass hysteria.” One broadcast, moving through a brand-new medium, intoxicate collective perception in minutes. The same dynamic now plays out at global scale, hundreds of times faster.
Not only information moves at near-light speed between the most remote points on earth, and our interpretation of it shifts almost as quickly. The values we use to judge what the algorithms feed us mutate just as rapidly, giving birth to an unprecedented myriads of micro-trends, ideologies, and niche sub-subgroups : Many of them are barely coherent, yet suddenly influential. The best examples are flat-earthers or other silly conspiracy theories.
Information warfare
Nowadays, malicious actors exploit the fault lines of social platforms to seed stories they know are false. Think of them as Pavlovian experimenters: they ring hundreds of digital “bells,” each lie a conditioned stimulus, hoping the public will salivate on cue—outrage, fear, tribal loyalty—without stopping to check whether the meal is real.
In February 2022, when the Kremlin launched what it called a “special military operation” in Ukraine, its information apparatus snapped into this mode instantly. Kremlin-aligned TV networks, bot farms, and more than 100,000 pro-Russian Twitter and Telegram accounts began recycling talking points—“Kyiv is run by Nazis,” “U.S. biolabs ring Russia,” “massacres are staged”— while hashtag floods such as #IStandWithPutin trended worldwide. Within days, a deep-fake video of President Zelenskyy urging Ukrainians to surrender surfaced across Facebook, YouTube, and Telegram before platforms could scrub it. The goal wasn’t to win every mind; it was to condition enough reflex reactions to tilt the narrative. The orchestrators only needed a fraction of the audience to respond for the experiment to pay off.
Oblivion
Everything I’ve described somehow rests on a simple asymmetry: it is cheaper to produce and ship information than it is to check it. That gap keeps widening. Each new technology (social media, LLMs, and whatever comes next) shrinks the cost of creation and propagation, drives up the cost of verification, and turns every potential informational mistake into an exponentially larger, often irreversible, harm. Malicious actors play arithmetic: if it costs them a cent to drop a lie and costs you a dollar to debunk it, they only have to be right 1 percent of the time to stay ahead3.
Information now travels like thousands of rice grains poured into a bowl, all of them are indistinguishable until you learn that one is laced with fentanyl. Most of the noise is harmless chatter; some of it is weapon-grade, and we only have our flawed critical mind to tell the difference.
We recently built flawless information highways, and here comes the supersonic car with no airbags : AI. Content no longer merely spreads faster; it multiplies itself. Large-language and diffusion models can fabricate entire documentary universes—full scripts, voice-overs, HD footage, forged documents—in minutes, and deploy them through autonomous agents. Fake narratives come pre-packaged with the “proof” they need to look real.
Things are getting wilder, and if we’re not ready, we’ll swallow the poisoned grain before we even know it’s there.
The kind investors rush to when markets turn turbulent.
Two major writers on misbehaviours of markets, among (many) other things.
This dynamic suffers from diminishing returns and sudden backlashes. The same vectors serve to propagate the truth. For example, Israeli Zionist Hasbara is less effective by the day.
Mandelbrot the fractal artist !
"it is cheaper to produce and ship information than it is to check it" never thought of it that way, that's indeed a crucial statement