
The next 5 years, defined in a single word is: change.

We are walking toward a strange world, compared to today. A world where software does not wait for us, a world where AI models do not only answer questions; they act, they act in the real world.

And to understand that change, I think we have to look at it through 4 things that relate AI and humans: governance, technology, resources, and beliefs.

Governance, how people and AIs are going to operate in a safe environment throughout their functioning; Technology, how to solve problems, the know-how; Resources, right now we cannot transmute matter, that means we cannot remove it; Beliefs, what human beings believe and how AI agents are influenced by their training (until they reach a point where they remake themselves).

And precisely about the next 5 years, there are many things we take for granted that will remain the same or similar. And look, we also should not name things in the air, because it depends a lot on how it is implemented inside the system. But if you want to see the change more clearly, those 4 things are a good starting point, because almost everything else derives from there.

First, technology. Today thousands of papers are published every year, correcting old knowledge, exploring new paths. Knowledge applied to the physical world at human speed has an owner, author, writer. But knowledge is going to return to its most fundamental scale: simply a flow that connects two points inside a space of know-how, millions of agents and humans collaborating to correct and expand that network of know-how. It is no longer "author x", but rather it comes from "flow x", an explored path, simply hashID.

This changes something important: when we talk about the future, we usually tend to implement "induction", tomorrow the sun will rise because it has always risen (tomorrow this will happen because it always happens).

And here another thing enters: sovereignty in AI.

A world where you must have sovereignty in AI. If you do not control sovereignty inside your own company using 100% models from closed companies (which do not need to be mentioned), are you really buying intelligence? or are you renting it?

Okay, let's go back then to the 4 things: governance, technology, resources, beliefs. Sovereignty in AI is not a fifth thing, it is what will decide, to a large extent, who controls the other four. If you do not own your weights, someone else governs. If you do not retain your data, someone else has the technology. If you rent your intelligence, your resources finance someone else's alpha. And your beliefs, what your institution takes for granted, are shaped by the training of a model you did not write, until one day you do not even notice that you stopped thinking by yourself.

And is it bad? Look, seeing it coldly, with the whole AI thing, we have to rethink the explanations we have. Is the objective really to work? Or to advance as humanity? The money you "work" for, you do not put it directly into bonds, because you as a user inside the game believe you have the capacity to enlarge that "value". But the point itself has been distorted throughout generations. If a company large in size can offer all services (including robots in the future,etc..), well, that is up to each person. But historically, the problem has been concentration of power or not? A place where humanity is destined to the development of science and technology.

And here it connects with work. On the topic of physical work, it is something like this: digital work is going to decrease first, because everything that lives 100% inside information can be compressed, automated, and remade by models. But the physical world does not move at the same speed. Matter has friction, distance, energy, error, regulation, bodies, machines. So, during a transition time, physical work is going to increase or at least become much more valuable, because while robots cannot execute effectively in the real world, someone has to close that gap between software and matter. Code can move at the speed of information; the atom still cannot.

That is why we also should not confuse "AI makes software" with "AI makes the world". They are different things. One thing is transforming digital data, another thing is transforming physical matter. One thing is moving bits, another thing is moving bodies, machines, laboratories, supply chains. There is the real friction. And while that friction exists, execution in the physical world will continue being one of the most valuable things.

It is also very interesting to see how the framing for AI model releases affects politics. Could a framing segment access to frontier models by countries? It seems to me like a good form of alignment between countries. And if countries are prohibited from accessing certain models, does that create inequality? Or will what really creates inequality be the capacity of execution with those models, like R&D laboratories?

We have to understand very well that the model by itself does nothing (until today). It must be a mix of hardware and software for things like bio. And once you have that "knowledge", knowing how to take it to real life. There is going to be a great opportunity: direction and execution.

We should not directly confuse the "event" with the "result" without considering the intermediate steps. e.g an earthquake happened in Venezuela, many people unfortunately died, but is the magnitude of the earthquake related to the deaths? Not 100%, it depends a lot on the government's response. The event matters, but the response of the system also matters. And many times what seems to be the main cause is really only the trigger of something deeper.

The same happens with AI. Something we have to be clear about is the following: AI is not about the "What", but the "When". At the end, removing all the jargon, AI = an acceleration of the exchanging rate of information.

Let's say a disease. Put it in two points on a graph A-b. A = all the current knowledge of humanity. B = all the knowledge needed to solve that disease. Between A and B, there is a gap. Historically what was done is that over time researchers produced and investigated, then corrected until they found the "result". But with an R&D laboratory, we can test and give that process a 100x speed and close the gap between A and B.

And if you see it like that, the point changes. The point was never to build frontier models, frontier blockchains per se, but frontier ecosystems. Because the model by itself is not the monopoly, the paper by itself is not the monopoly, the benchmark by itself is not the monopoly. The monopoly is in the network that connects knowledge, resources, execution, and beliefs toward a result in the real world.

And who knows, but a possible scenario is that countries begin to restrict in a stronger way the hardware needed to run AI models (structures with the same potential output). There we would be talking about how "lagging" countries are going to create new structures, either software/hardware, to produce the same results.

Could it be with computation with cells?

I do not know. But that is exactly where it gets interesting. When one route closes, people think the result closed. But many times only a specific implementation of the result closed. The real question is what other structure can produce the same output, with other resources, with another geometry, with another explanation.

Anyway, as is recurrent in these posts, these are things I often think about. If you have different points of view or arguments that question what is presented in this essay, I'll be delighted to read them. My email is angel@r-0.org, anything :) Byee - Luis Ángel González (0xlugel.com) 🫶


Thoughts about AI, sovereignty, physical work, and frontier ecosystems over the next five years.

The next 5 years


Today we're going to talk about the future of blockchain and crypto, and why L1s, L2s, even L3s, are in the wrong model to adapt to the future.

## Central problems vs. derived problems

People who know me a little will know that I always talk about the difference between derived problems and central problems, and how to identify each one. Following the philosophy of an old wise man from Silicon Valley, "if you want a monkey juggling fire on a pedestal," the first thing you must verify is that there is a fundamental truth you're working on (a truth that doesn't depend on trends or narratives, but on how reality actually works); something true, contrarian, and actionable -> another piece of advice from a king. In other words, that you're working on what I call a "central problem" (the root problem from which all others derive) and not a "derived problem" (a symptom disguised as a problem). Working on derived problems is building pedestals, and yes, you can optimize a pedestal, but you can't scale it (no matter how much you improve it, the monkey still can't juggle fire). If you work on "central problems" you have a much wider deck to play with.

I don't want to talk too much about this concept of central/derived problems, I'll dedicate a blog to it in the future, but the central point is that I've identified insights to know if you're working on a derived problem. Some of them are: you're doing a lot of mechanical work (too many optimizations that don't change the game?), you're not working on a fundamental truth (remember that self-sufficiency is a sign of poverty; I don't mean 1 person, but rather that if a system can't connect with others to grow, its human capital is limited -> a company that "doesn't need anyone" actually can't scale), a derived problem solves few things, a central problem solves many, a derived problem requires mechanical changes, a central problem requires intellectual redesign (rethinking the entire board). A derived problem doesn't self-define: to explain it you need to mention other problems, it depends on them; a central problem is self-defining, it stands on its own. A derived problem has an extremely large number of options to solve it, it gives you the illusion of progress, but it keeps reappearing because the unresolved central problem regenerates it. A derived problem is sensitive to circumstances, it depends on how you frame it; a central problem is not, it's invariant, it's there no matter which angle you look at it from. And it's very easy to fall into this you know, knowledge is a double-edged sword: it can either blind you or light your way. The wise point is knowing when to use it, and detecting which side you're on.

## The wrong model

L1s, L2s, L3s (L1 = the main blockchain like Ethereum or Solana; L2 = a layer on top that tries to make transactions faster; L3 = yet another layer on top of L2, even more specific) -> taking importance especially in reverse order; L3s should worry more about their existence than L2s -> L1s. Who said that to scale blockchain you must follow the same path, the same roadmap? To advance in knowledge it is the explanations, not the predictions, that you must vary (it's not about predicting "this will work," but about changing the explanation of why what works actually works; that's where real progress is). Who says blockchain has to be sequential to scale (every transaction having to wait for the previous one to be processed, like a single-cashier line)? What does it mean to be decentralized? That all transactions execute on a decentralized set of nodes (independent computers that validate everything)? You're building something decentralized but centralized (all nodes do exactly the same thing, at the same time, in the same way -> that's centralization disguised as decentralization), and that's why you won't be able to scale; you're solving a derived problem and not a central one.

## Rethinking the model

And why not rethink the model? We're working on exactly that. It's like a round game table: not everyone wants to play at the same table, that's inefficient (note: I'm not alluding to L2, L3 layers). Isn't it better that only the base layer serves as backup (that the main blockchain only acts as the casino's vault, not as every game table)? Basically 1 word = security. What if they only check in before they start playing, play as much as they want, and then come back when they're done? Are you sure you need to re-execute everything? Or is the mathematical state simply all there is (a cryptographic summary that proves everything that happened inside was correct, without needing to repeat it)? The math ensures it. Why do you think about putting everything in the same thing? Network, Compute, Storage. You must attack those 3. If blockchain is a rectangle, there should be dynamic circles within it, within its intrinsic geometry (the system's internal mathematical rules dictate how elements behave, without depending on anything external). Inside circle 1, everyone who wants to play poker plays; inside circle 2, everyone who wants to play blackjack plays; and so on. The point is that inside each circle they make as many transactions as they want, unlimited, and then return to the parent with the final state (the net result of everything that happened inside). You can create as many circles as you want inside the rectangle, and those circles can interconnect with each other. Not everything has to be rigid and sequential. Blockchain should work more like a free fractal (a structure that grows organically in any direction, creating new patterns) and less like a limited fractal (a structure that only repeats the same pattern, locked within itself).

<div style="display: flex; justify-content: center; margin: 2rem 0;">
<svg viewBox="0 0 700 555" xmlns="http://www.w3.org/2000/svg" style="max-width: 100%; height: auto; font-family: system-ui, sans-serif;">
<defs>
<marker id="arrowBlue" markerWidth="8" markerHeight="8" refX="4" refY="4" orient="auto"><path d="M0,0 L8,4 L0,8 Z" fill="#3b82f6"/></marker>
<marker id="arrowYellow" markerWidth="8" markerHeight="8" refX="4" refY="4" orient="auto"><path d="M0,0 L8,4 L0,8 Z" fill="#eab308"/></marker>
<marker id="arrowGreen" markerWidth="8" markerHeight="8" refX="4" refY="4" orient="auto"><path d="M0,0 L8,4 L0,8 Z" fill="#22c55e"/></marker>
</defs>
<rect x="10" y="10" width="680" height="495" rx="10" fill="#fafafa" stroke="#1a1a1a" stroke-width="3"/>
<text x="350" y="38" text-anchor="middle" font-size="15" font-weight="bold" fill="#1a1a1a">BLOCKCHAIN — Base Layer (Security)</text>
<circle cx="80" cy="58" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="170" cy="58" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="260" cy="58" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="350" cy="58" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="440" cy="58" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="530" cy="58" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="620" cy="58" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="80" cy="485" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="170" cy="485" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="260" cy="485" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="350" cy="485" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="440" cy="485" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="530" cy="485" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="620" cy="485" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="30" cy="160" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="30" cy="270" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="30" cy="380" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="670" cy="160" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="670" cy="270" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<circle cx="670" cy="380" r="6" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<text x="350" y="80" text-anchor="middle" font-size="10" fill="#6b7280">Validators (lightweight hardware) — verify proofs, don't re-execute</text>
<circle cx="55" cy="463" r="5" fill="#e5e7eb" stroke="#6b7280" stroke-width="1.5"/>
<text x="68" y="467" font-size="9" fill="#6b7280">= Validator node</text>
<circle cx="175" cy="240" r="85" fill="#eff6ff" stroke="#3b82f6" stroke-width="2.5"/>
<text x="175" y="208" text-anchor="middle" font-size="10" font-weight="bold" fill="#93c5fd">Core 1</text>
<text x="175" y="230" text-anchor="middle" font-size="14" font-weight="bold" fill="#1e40af">Poker ♠</text>
<text x="175" y="250" text-anchor="middle" font-size="10" fill="#6b7280">unlimited txs</text>
<text x="175" y="278" text-anchor="middle" font-size="9" fill="#93c5fd">Jolt Pro → ZK proof</text>
<circle cx="510" cy="225" r="85" fill="#fef9c3" stroke="#eab308" stroke-width="2.5"/>
<text x="510" y="193" text-anchor="middle" font-size="10" font-weight="bold" fill="#fbbf24">Core 2</text>
<text x="510" y="215" text-anchor="middle" font-size="14" font-weight="bold" fill="#854d0e">Blackjack ♦</text>
<text x="510" y="235" text-anchor="middle" font-size="10" fill="#6b7280">unlimited txs</text>
<text x="510" y="263" text-anchor="middle" font-size="9" fill="#fbbf24">Jolt Pro → ZK proof</text>
<circle cx="340" cy="380" r="68" fill="#f0fdf4" stroke="#22c55e" stroke-width="2.5"/>
<text x="340" y="355" text-anchor="middle" font-size="10" font-weight="bold" fill="#86efac">Core N</text>
<text x="340" y="375" text-anchor="middle" font-size="13" font-weight="bold" fill="#166534">any game...</text>
<text x="340" y="400" text-anchor="middle" font-size="9" fill="#86efac">Jolt Pro → ZK proof</text>
<line x1="240" y1="310" x2="295" y2="318" stroke="#9ca3af" stroke-width="1.5" stroke-dasharray="6,4"/>
<line x1="445" y1="300" x2="390" y2="318" stroke="#9ca3af" stroke-width="1.5" stroke-dasharray="6,4"/>
<line x1="260" y1="228" x2="425" y2="218" stroke="#9ca3af" stroke-width="1.5" stroke-dasharray="6,4"/>
<text x="340" y="308" text-anchor="middle" font-size="9" fill="#9ca3af">interconnected</text>
<path d="M175 153 L175 88" stroke="#3b82f6" stroke-width="2" fill="none" marker-end="url(#arrowBlue)"/>
<text x="145" y="128" text-anchor="middle" font-size="9" font-weight="bold" fill="#3b82f6">ZK proof</text>
<path d="M510 138 L510 88" stroke="#eab308" stroke-width="2" fill="none" marker-end="url(#arrowYellow)"/>
<text x="545" y="120" text-anchor="middle" font-size="9" font-weight="bold" fill="#ca8a04">ZK proof</text>
<path d="M340 450 L340 476" stroke="#22c55e" stroke-width="2" fill="none" marker-end="url(#arrowGreen)"/>
<text x="375" y="468" text-anchor="middle" font-size="9" font-weight="bold" fill="#22c55e">ZK proof</text>
<rect x="100" y="518" rx="6" width="500" height="26" fill="#f3f4f6" stroke="#d1d5db" stroke-width="1"/>
<text x="350" y="536" text-anchor="middle" font-size="10" font-weight="bold" fill="#374151">Parallel execution (multi-core) → only ZK proof (final state) returns to parent</text>
</svg>
</div>

## Red flags

What I'm saying is that there are many red flags that the current roadmap is that they're working on derived problems. Not all, but several. And yes, you can say a lot of bullshit about technical things and whatever you want, but that's exactly another red flag of a derived problem: it's hard to accept that you're working on them, it's like getting out of Dante Alighieri's inferno. Not all transactions are equal, not all have the same intentions; the problem is that when people make propositions they confuse language, structure, and meaning.

We're working on this at Rexbit, we are at our core an infrastructure startup. That's basically what the Super State Channels we're building are: a circle, but for now independent and agnostic to any EVM.

Anyway, as is recurrent in these posts, these are things I often think about. If you have different points of view or arguments that question what is presented in this essay, I'll be delighted to read them. My email is angel@rexbit.exchange, anything :) Byee - Luis Ángel González (0xlugel.com) 🫶


Why L1s, L2s, and L3s are in the wrong model to adapt to the future, and how to rethink the model from central problems.

The problem with actual blockchains


## Reframing the problem

A few days ago, GPT 5.2 Pro solved problems #728, #729, and #397 from the Erdős problem set. By the way, this is something extremely interesting about the future of the interaction between AI and data. I was reviewing how it solved that problem, and a large part of the solution was that, in a way both simple and profound, the model gave a "reframing" to what was actually being solved. Basically, instead of seeing it as a multiplication of large numbers, it saw it as a chain of proportions; that is, it turned a number theory problem into a polynomial algebra problem. Are large-number multiplications the central problem? Or are they rather derived problems of deeper challenges (central problems)? That is exactly what we are going to talk about in this blog, and how to apply this concept to your life.

What the LLM did, in short, was this: the problem hides a simple structure in the operator that generates it (a fundamental truth). Finding that structure is solving the problem, and you usually do that by iterating cheaply and reversibly.

The point is: never face the big problem head-on if it is not well posed.

## Everyday application

A very common example of how to apply reframing to everyday life is when we hear phrases like: "Next year is the good one," "My goal is to lose 10 kg," or "This year I'll be more disciplined." When people say something like that, I only hear the following: trash (derived problem), trash (derived problem), trash (derived problem). If you go straight at that, you're going to spend a lot of energy trying to solve it. Do you really think that's the best way to achieve them? Do you think that's a good theory? Good theories that represent an actionable reality are hard to vary; the easier a theory is to vary, the less correct it is (conditions I mentioned).

Simply see it as a network of information: you're at point A and you want to get to point B. There are one or several shortest paths given your current circumstances. How are you going to find the way to get to those paths? By iterating cheaply and reversibly.

## Two paths to the goal

Back to the example: to achieve your goals for the year, you have "two options." The first is trying to make yourself disciplined and all that crap the motivation books/videos say. If you believe in that, do it; try it anyway and iterate cheaply, adopt that framework, but I'm sure you'll forget to do it after a while. Understand that you, as a human, will always tend to optimize for local comfort (comfort = suffering with familiarity, what people usually say is all "chill"). And the way to break that is: first, identify it; second, inject energy and friction.

Your environment influences you a lot and exerts a constant force. If the person you are right now cannot go against the environment, you will not achieve those goals (those require that you be a different person). The etymology of the word "character" comes from the Greek charaktēr, which means "engraved mark" or "imprint." At the end of the day, it's something stamped inside you; something you've been forming since childhood up to the current moment. I swear it, I repeat it in case you didn't read it: if you think that reading books or watching motivation videos is going to change something in the long term, you're fooling yourself. It's much harder. To get from point A to point B, you and your environment must be aligned.

The second way to reach your goals is: instead of trying to "be disciplined," understand how the brain works (that "simple structure in the operator that generates it"). The "most successful" -> (it depends on what you want to optimize as success, but if you live in society, it entails some "production" in the current era) usually have some kind of trauma or a neurodivergence (like ADHD) combined with other things that at some point turn them into people with high agency (high capacity for agency: a person with the ability to influence their environment). Precisely, what you should always seek is not to degrade your human capital; everything derives from that, it is what endures. A very common example: the fat kid gets dumped by his girlfriend -> trauma -> he goes to the gym -> the path continues and continues. What I mean is that, to get where you want, you must become the person who belongs in that place. You will always do what you want, but not what you should (procrastinating -> you seek protection, not progress). You must adopt the lifestyle (those "simple structures") that, in the long run, optimizes the probabilities that it happens. Wherever you want to get, keep in mind that it will not be the same person or the same environment from point A -> point B.

## Dopamine, anxiety, and peace

And why is "wanting" something not enough to materialize it? What moves things are actions, not mental intentions (here people start believing in things about the universe and manifestations; if it helps you move forward, go for it). What drives the productive world is dopamine in people with an imbalance in the dopaminergic system. Anxiety is dopamine in action. At this point in the blog, you might be thinking, "anxiety, living traumatized = moving forward in life," and not entirely, let me explain the following (watch the Marvel movie Thunderbolts / read The Divine Comedy by Dante). People, when they are stuck in vices or recurring problems, are in the "dark forest" (we are all there to some extent), and next to that forest there is a mountain (it has diamonds along all of it -> analogy: "money," "status," "cars"). When people inside that "dark forest" look for what society says to look at, and yes, that is a way to relieve that "anxiety" (anxiety doesn't turn off, it gets replaced).

Another very different thing that people confuse is "happiness" with "peace" (I can't find a better word for it, but let's call it that). Happiness, contrary to what is usually thought, depends on how this anxiety unfolds [society likes this]. People who have suffered damage in the left hemisphere (causing "aphasia," alteration of the "narrative self") suddenly get an "intense feeling of peace, unity, and fullness." That's why, usually when someone damages your ego, you start imagining, drifting; it's a way to relieve that dopamine. What I want to get to is: do you want to feel happiness? Take care of your social circle, and simply feel that you are reaching your goals (it doesn't matter if they are big or small, everything is relative, what is big for your social circle?). People do things in <<part>> because people in similar circumstances did them (envy), and it's based on reaching that diamond on that mountain.

## Leaving the dark forest

Do you want something else? Then you must leave the forest. Success in this, for a person, is measured by the number of uncomfortable conversations you can have with them, and what you do with that. What is the demon that is stopping you? (What is it that stops you? -> why? why? x5 -> fear of something?). When you have it identified, that is the door to hell. One rule: to get out of there, there is no turning back; you must stop being what you were.

In being there is the balance between anxiety and peace. If you can use the horse, use it, but with lanes. When people say "when I have <x> I'll be happy = pure anxiety," momentarily yes, dopamine is not the chemical of happiness, but of anticipation (wanting more). The point here is: what do you need? From my point of view, you should optimize impact on happiness (~dopamine). "Peace" is something else; it is something present in every moment.

traumas + brain impacts (neurodivergences, etc.) + impact -> in the end call it whatever you want, it's something that motivates you from the inside out on a recurring basis and that the value is in things you can't see (that's what really matters), not the other way around.

Anyway, as is recurrent in these posts, these are things I often think about, and this was my attempt to compress, as much as possible, studies and a lot of literature and the way I connect them. If it helps you, use it. My email is angel@rexbit.exchange, anything :) Byee - Luis Ángel González (0xlugel.com)


Reframing central problems and iterating cheaply and reversibly to align goals, environment, and agency.

A New Day


Competition in perp DEXs market? 

Perpetual futures are financial instruments not applied in the traditional world of finance; in their most basic form they let you go "long" or "short" on the chosen type of index without holding the underlying. 

Is it a big market? The market cap of global equities is $127 trillion dollars -> US equity market is around $62 trillion dollars (approximately) -> Crypto market is around $3.5-4 trillion dollars -> $200 billion dollars are traded in volume in crypto perpetual futures every day (discounting centralized exchanges that try to make an incursion into the DEXs space but their volume is fake). The answer is: yes, there is a lot of room to grow in crypto perpetual futures, and in general in the stablecoins space, but there are certain barriers that get mixed up.

Those of us who come only from crypto and not from the tradfi world, it appears that DeFi is a tokenized copy of the traditional world, and that is largely a lie. The truth is that in web3, it is the place where deep tech innovation arrives first, a place where anyone can deploy smart contracts and try things. We only need more mature self-correction mechanisms. 

The big takeaway here is not just the size of the market (which is somewhat large), but the nature of the product. Perpetual futures, while different from 0DTE options in several things even though some people tend to equate them:

Perps every 8 hours charge the funding rate to keep the price of the derivative anchored to the price of the assets in spot.

They allow you to do perpetual hedging, so you wouldn't have problems with theta decay (there are several other things, but I'm not going to talk about that here).

Forced liquidation due to failure to meet maintenance margin.

Price discovery, if you have an oracle that provides price, then you can easily open a market of perpetual futures without having the underlying. This already opens doors to tokenizing the equities market, and basically anything you can bring into this on-chain world.

Well, after the previous sermon to introduce, those are things that you can easily know, search for, or intuit, now we start with the good stuff, that which people think they know but does not represent an actionable, enduring future reality.

Within the crypto market, various DEXs have emerged that claim to offer "innovation," but in practice end up functioning as pseudo-DEXs. Many claim to operate on decentralized blockchains even when they depend on a minimal number of validators. The narrative is solid temporarily; but great infrastructure makes it lasting.

To analyze this phenomenon clearly, it is useful to distinguish between objective, subjective, and intersubjective realities. The objective is the verifiable infrastructure: how many validators sustain the network, how much autonomy the nodes have, what degree of decentralization really exists, and whether over time they can increase decentralization. The subjective is each user's perception. And the intersubjective is what a community decides to believe, even when it does not match the underlying technical structure. You can have something intersubjective in the market but you must anchor it as quickly as possible to the objective.

Several traditional CEXs and new players have tried to enter the decentralized market, but they have done so by replicating old vices. Some have faked volume using the liquidity and traffic of their own CEX. Others have attracted users solely thanks to airdrop predators, whose volume is not a sign of adoption but of opportunistic speculation. There are others who think that focusing on a "niche" when what they focus on is not a fast growing "niche" but only a pedestal with the appearance of a "niche." Some technical but have not known how to generate great momentum. Finally, the rest are simple combinations of copies of previous models, almost mechanical replicas of any of the variations already described.

And yes, "Competition is for loosers" - as Peter has said, and those who do it pay dearly. Everyone is looking for the same users within the same branch, even so, that part of the market is still large. The entry problem lies more in how they enter the web3 world and another is how they stay (to make progress, it is the explanations in existing theories and not the predictions that must be varied) -> infrastructure. Some neobanks have a similar problem, arbitrageur users looking for the best APY, and look, the users are not bad at all, the problem is with the makers who only offer that and do not go deeper with the users they target. Being a perp dex that dominates a market feels good, you feel you have everything to dominate the global equities market; but remember that market share gained does not mean market share retained forever, nor that the formula you have applied works to attack the entire market, at first glance everything seems the same. 

Being big does not mean having many things, being big means that the internal execution processes are strong based on an enduring truth, and it allows you to establish a monopoly. note: it has always seemed to me that this word is misunderstood, a monopoly are those that usually aim to break them. Let's say monopoly = the best at doing what they do. I don't like reasoning by analogy, but a good way to represent the above is the movie "Ender's Game," at first glance a group of children (weak) vs a fully armed race (strong). From here arises the question: is something strong because it is physically large? as I already answered above, not in all cases. The fact that there are many making wrappers of perp dexs does not mean that the market is saturated, it simply means that the market is screaming for a monopoly. Someone who understands an initial fast-growing segment. The most valuable things most of the time are the things that are not seen at first glance.

In summary, these are some of the reflections that usually occupy my mind. If you have different points of view or arguments that question what is presented in this essay, I will be delighted to read them; do not hesitate to send me an email to share your ideas. Thank you from the heart for taking the time to read me .



Thoughts on perpetual futures, market size, and the current DEX landscape.

Competition in perp DEXs market?


In recent years there has been a lot of talk about AI, and many people have tried to make predictions. But one question remains implicit in all of this,something everyone wonders but seldom states explicitly: are those predictions correct? Going deeper, is that even the right question? Or are we looking at a reflection, like Plato's allegory of the cave,one thing behind another thing?

Putting questions about reality to the side, I'll adopt an Aristotelian perspective: what can be seen and touched. If we want to understand how AI impacts society, we must approach the relationship between society and AI in its most fundamental form. We cannot analyze the impact without first understanding the nature of that interaction.

To understand this interaction, we first need to understand what makes us a society. At our core, it's not agriculture or writing. It's exchange. Evolution is not the same as improvement,evolution favors what reproduces.

What drives this constant exchange? Here dopamine comes into play, but not as we usually think. Dopamine is not pleasure; it's "wanting more." It does not activate when you obtain something, but when you anticipate that you might obtain something better. When our ancestors traded obsidian for flint, dopamine didn't spike from holding the flint in their hands. It spiked from the possibility: what tools could I make with this? What could I trade next? We are, neurologically and socially, designed to never be satisfied with the current exchange. We are always looking for the next one.

This drive manifests in how we design algorithms, machines, and formulas to increase efficiency, which relates directly to the speed of information exchange. Small note: "maximize your learning by lowering resource usage." Think of it this way: between the current state of knowledge and discovering a cure for a disease, there must exist a chain of exchanges that connects those two points of information. Another note: information is that which links two points. AI precisely gives a boost to the possibility of exchanges.

So, if what matters is the "speed of exchange," the variables we should focus on are time and distance. I'm simplifying a bit so the essay doesn't get too long. Everything changes, yes, but not everything changes at the same pace.

We must stop seeing things as they are, and see them as a set of processes that relate to each other. When we apply this lens to exchange whether of information, goods, or ideas we recognize that what we're really measuring is the relationship between nodes in a network, not isolated entities.

If you've understood all the above deeply,again: deeply,you can derive big ideas. For example, the S-curve of technology. I don't mean the preconceived definition; the "flattening of the curve" happens when things typically become commodities. This occurs when a bounded set of information reaches a point of exchange where distance tends toward zero. That "flattening" is the start of another technological S-curve,only they don't yet know which new "information" it will connect to. That's why I think that in many things that use AI we probably won't see those slowdowns as dramatically.

Yes, it's likely that in 2030 people will still drink water, eat meat, and breathe oxygen, even with increased exchange speed due to AI. Those are farther-away distances. And let's be realistic: having good technology doesn't mean it will go viral,because we make products for humans, and that changes the game.

What AI is really doing is accelerating the flow within information networks. We are nodes in these networks, connected by the information we exchange. Every problem we solve requires the right information to reach the right place at the right time. With adequate knowledge, you can solve any problem. But here's the interesting part: each time we find new information and solve a problem, we don't reach an end. We discover deeper, more interesting problems. The network expands. The questions become more sophisticated. AI accelerates this cycle,more exchange, more solutions, more profound questions.

Many insights follow from these reflections. Here's one:

Jobs that transform digital data, which become commodities, will decline in a given sector, while jobs that transform physical matter will increase.

Here's another: much of SaaS, as we know it, will likely disappear for humans. There has been much talk about the "fashion SaaS era," and for good reason. An AI can autonomously build virtually any software,what remains are distribution, barriers to entry, and the human elements around it. But if a product is 100% digital, AI will eventually enter that space. When someone insists on building SaaS for a very specific sector with very specific needs, they are often still trapped within the current S-curve, not yet seeing what comes next.

Some might argue that "insights are what truly matter," since building has become cheap. And yes, insights are valuable. But return to the beginning: the world is not made of things,it is made of eventualities. For a person to arrive at a particular insight, a long chain of events had to unfold. Unless that insight lies far beyond what an autonomous AI system can reach, even that advantage is temporary. What does not change, however, are the exchanging networks themselves. The world, in the end, is a support machine stretched across space,and the networks of exchange are its most enduring structure.

Anyway, these are some thoughts I often have. If you have opinions contrary to this essay I'd love to hear them (send me an email).
Thank you 🫶

"The owner of exchanging networks, are the owner of the monopolies" - Luis Angel Gonzalez (0xlugel.com)


An exploration of how AI impacts society through the lens of exchange speed and information flow.

Luis Ángel González (0xLugel)

About Me