
Tech • IA • Crypto
Les avancées en intelligence artificielle transforment rapidement le développement logiciel et accélèrent un basculement plus large des systèmes d’information centralisés vers des systèmes décentralisés.
Les progrès rapides des grands modèles de langage transforment la programmation, d’une compétence spécialisée en une capacité largement accessible. Des tâches autrefois confiées à des ingénieurs juniors et intermédiaires peuvent désormais être automatisées, certains systèmes d’IA pouvant générer 95 à 99 % du code dans certains flux de travail. Cette évolution compresse les parcours de carrière traditionnels et redéfinit le rôle de développeur.
La compétence émergente n’est plus l’écriture manuelle de code, mais la capacité à formuler clairement les problèmes et les résultats attendus. Les développeurs deviennent des coordinateurs, guidant les systèmes d’IA plutôt que d’écrire chaque ligne. Cette transition valorise davantage le jugement, la conception de systèmes et la maîtrise des outils d’IA.
Le développement assisté par IA augmente fortement la production. Des projets qui prenaient autrefois des années avancent désormais en quelques mois, certains signalant plus de progrès en six mois qu’au cours des trois années précédentes. Cette accélération élargit les ambitions et transforme des objectifs gérables en projets bien plus vastes et complexes.
La domination des plateformes cloud a créé des risques structurels. Quatre problèmes persistants sont identifiés: atteinte à la vie privée, censure, hausse des coûts et vulnérabilités systémiques en cybersécurité. Ces enjeux sont considérés comme inhérents aux architectures centralisées, difficiles à corriger par régulation ou ajustements progressifs.
Un mouvement croissant promeut des systèmes décentralisés où les individus exploitent leurs propres serveurs et gardent le contrôle de leurs données. Ce modèle reflète l’architecture de Bitcoin, qui remplace le contrôle monétaire centralisé par un système fondé sur le consensus distribué et l’énergie.
Bitcoin est présenté comme un exemple précoce de la manière dont des systèmes d’information décentralisés peuvent bouleverser des structures établies. Son succès suggère que des transformations similaires pourraient toucher d’autres domaines de l’informatique et de la gestion des données.
À l’ère numérique, contrôler l’infrastructure informatique revient à gagner en autonomie. Ceux qui gèrent leurs propres systèmes gagnent en indépendance, tandis que la dépendance à des fournisseurs externes crée une forme de subordination. Cette dynamique rappelle des périodes économiques où la maîtrise des technologies dominantes conditionnait la mobilité sociale et économique.
Les outils d’IA réduisent les barrières à l’entrée, permettant à des non-experts de créer des applications et des systèmes. Cette démocratisation ouvre la participation à un public plus large et pourrait transformer l’innovation et l’entrepreneuriat.
Le remplacement de rôles traditionnels suscite des inquiétudes quant à la sécurité de l’emploi et au sens du travail. Beaucoup tirent identité et reconnaissance de leur activité professionnelle, et l’érosion de ces rôles pourrait avoir des effets psychologiques et économiques étendus.
S’adapter implique d’apprendre à utiliser efficacement les systèmes d’IA, notamment la gestion des entrées, des contraintes et des sorties. Des qualités humaines comme le goût, l’esprit critique et la prise de décision restent essentielles et pourraient devenir les principaux facteurs différenciants.
L’intelligence artificielle ne transforme pas seulement le développement logiciel; elle accélère aussi une transition vers un contrôle décentralisé de l’information, redéfinissant le travail technique et les structures de pouvoir numérique.
Hi everyone, my name is Matt Hill. I'm a co-founder and CEO of Start 9 Labs. We make operating system and router uh operating system software that enables normal people to take pretty gritty advanced control over their information systems. Um, I have been handwriting code, which is kind of a novelty at this point, for about 11 years. And I'd say as of 2 years ago, I would have told you that my maximum perceived uh advancement of artificial intelligence in relation to coding capabilities um would have been scripting, maybe a bit of rudimentary application development, but that it certainly would not replace uh senior level developers. It may give some mid-tier developers a run for their money and sure junior developers might be out looking for a job. Um over the last two years, as many of you may have observed through one avenue or another, the advancements in artificial intelligence have been shocking. Um the last time I had this kind of world view uh shock and alteration was Bitcoin. Uh I think a lot of us when we discovered Bitcoin after dismissing it the first or second time eventually came around to realizing that it was a uh a very important technology that could and probably will fundamentally alter the architecture of money and finance on earth. And that took a long time for me to make that transition, right? To go from my prior worldview, which was fundamentally libertarian, but ultimately uh you I was a gold bug, a Ron Paul revolutionary gold bug. And um it took me a couple years to really wrap my head around the depth of the the change that Bitcoin proposed. And um I got really settled into that and I found a new purpose for myself, right? I I got involved with a Bitcoin company early on. Um I did coding for them and eventually became the CTO of that company and uh after departing decided that I wanted to do something really really gritty in the Bitcoin space. Um I wanted to do an uncompromising the most cipher punk angle I could think of where we wouldn't be dealing with the regulators where we wouldn't be dealing with the lawyers um where the team would be predominantly engineers building uncompromising technology in and around Bitcoin and uh we set out to do that and we got into a groove and we've been doing it for 7 years and I think many of the people at this conference probably use our technology at this point uh about 25% of the nodes codes on the Bitcoin network are running on Start 9 servers, Start OS operating system. And about 6 months ago, I had a um a shock, right? An equally impactful, if not more impactful shock than I had when I first discovered Bitcoin. And that was a revelation of where we were at and potentially where we were going with artificial intelligence. And it might even be played out at this conference so far that more people were talking about AI than Bitcoin. Um my talk was supposed to be become a Bitcoin citizen and run a node, right? which is something I talk a lot about and I think everyone should do. If you're a Bitcoiner and you care about the Bitcoin network, it's a very good idea for you to run a node for yourself to validate your own transactions to ensure your ability to access the network at any time. But if we're being honest, it's not what I've been thinking about a lot recently. um something bigger is taking place here that Bitcoin is an integral part of but it is broader and what it is is information systems are changing right Bitcoin showed us the way of how to go from a centralized information system that produced money by fiat to a decentralized information system that produced money by consensus and energy And what I find fascinating is that many of our other information systems are centralized as well. And the consequences of that centralization are not as acutely felt today as the monetary centralization was. We have as a society and civilization have sort of just rolled recklessly into the cloud computing era. Right? The information era, information age began in earnest with the early personal computers. There was some leadup to that, but the digitization of our lives and our economy began with the personal computers. And then we transformed into a platform based economy, the platform era where we took all of the computing infrastructure and information and began shutting it up onto centralized clouds. This was a convenient and very natural progression in my mind. I understand why it happened. This wasn't some huge mistake. It was a necessary stepping stone towards what? Well, at present, more and more people are beginning to realize the consequences of the information systems that we've all come to depend on. When you use central service providers for your basic computing needs, four acute problems arise from that. One is invasion of privacy, one is censorship, one is cost. They have to get paid. So, you're either paying subscriptions or you're paying with your data. And in most cases, you're paying with both. And the last one is the inherent insecurity of pooling everyone's data onto a single server and expecting it not to get hacked through social engineering of the people with the keys. All four of these problems are inherent and irrevocable with the system that we have designed. You can stack layers of application security regulation oversight. You could even have somebody in Washington think they could pencil a law that could prevent a server from being hacked. But the reality is is that as long as we are all using centralized information systems, invasion of privacy, censorship, forever rising costs, and inherent cyber security vulnerabilities will never go away and they will continue to get worse. And every one of those will hit every one of us at different times and to different magnitudes. So we need something else, right? We need an alternative to a centralized information system and that is obviously a decentralized information system. What does that look like? Well, it means running servers. Servers are where all the data are processed and stored. And if you're not using your own, you're using somebody else's. And if you are using your own, then you control it. And to get around late in the game here to the essence of this talk, when you control information in the information age, you have power. In every prior human era, whatever the most advanced technology was in terms of efficacy, even violence, if you were a master of that technology, you were a sovereign being. You were free. You were powerful. You had the option to choose your purpose and direction in life. But if you didn't master the technology of your era, you were always a secondass citizen. You were always somebody who had to answer to someone else, right? You till the land instead of owned the castle. You worked the factory line instead of ran the factory. Just doesn't mean you're a bad person. It just means you're not as free as the person who ran the factory. >> Welcome to predict. The world is a market. Everything is a market. Get a 100% cash back up to $100 on your first predict bet if it loses. predict where everything is a market. >> And the information age offers us all the same opportunity. And like the ages before it, very few of us will recognize the technology in its essence, the leverage it affords, and take the time to master it. With the rise of AI, even my prior conceptions of what it meant to be successful in the information age were shattered. They said, "Learn to code." I learned to code. I got good at it. People who did that named their price. They named their location. They named their job. They were a form of aristocracy in the information age. And in about one year's time, not an exaggeration, they have all been wiped out. Junior engineers don't exist anymore. I will never hire a junior engineer for the rest of my life. Mid-tier engineers gone within the last couple weeks. Senior level engineers might hold out for a little bit while longer, but large language models will perform 95 to 99% of all code going forward and anyone can do it. Those of us who have experience with it will do it better for now, but it's a totally different skill set. And as more and more of our systems are in essence code, the ability to produce code which previously required you to write code by hand and now just requires you to be very explicit about what it is that you want or what the problem is. And if you can be very good at communicating those two things out the other side of it, we'll pop 10,000 lines of code that do exactly what you want it to do. It's very impossible to overstate the power of that. At the moment, I feel like a wizard, right? That's often used in software development. I thought I was going to build a monument in my lifetime. and that Start 9 and our server and router operating systems and soon security cameras and other smart devices that are going to make the world around you intelligent yet under your control was going to be my monument. And over the last 6 months, I've made more progress in that direction than I had in the previous 3 years combined. And that's not an exaggeration. I suddenly feel like I'm doing this and buildings are popping up out of the ground. And some of you have had similar experiences with vibe coding, right? Where before you couldn't do it at all and now you can just pop out an application. So what's happened for better or worse is the to-do list just expanded by 10x. Now I can't build a monument. Now I have to build a city. So, the to-do list is getting longer. We're not going to run out of things to do. This technology just takes what only previously a few people could do with any degree of efficiency and sort of showered fire upon the people. And so, what I'll challenge you to do here, and I started a couple minutes late, so I'm going to go about one over if you don't mind. Um, is to find purpose. now because it's all threatened to an extent. Okay, a lot of jobs are very threatened right now and so many of us derive our purpose and reason for waking up in the morning through our skills, our knowledge, our work. The paychecks we receive are evidence and validation that we matter and have value. A lot of this is going away. It is a seismic shift in the history of humanity, I think. And so I will challenge you to not reject this and cling to an ancient era that was last year and get with the modern times that is this year and start leveraging these technologies, learning how to use them with great efficiency. Manage context windows. Manage the budget. Be the agent. have the taste and judgment that the AIs don't yet have and may never have. Hone those skills, identify something in the world that doesn't quite write work the way you want it to and have the confidence to know that even though you weren't a trained developer, you can probably code it up now and it's really fun. And I can tell you right now, I got a laundry list of things to do that I will never get through. So if you think sovereign computing is cool and writing code is fun, I'll give you a list of 10 things you could build. Um, so anyway, have fun and uh, thank you. >> Every year this community comes together to celebrate, to debate, to build what comes next. And every year the stage gets bigger. Sound money center stage. So where do you go to celebrate the next chapter in Bitcoin history? You come home. Nashville. July 2027.