#66 - Raphaëlle d'Ornano - Behind Decoding Discontinuity 

Speaker A: Um, the moats and Anthropic with cloud code. OpenAI is doing the same thing. Have really moved from. We have technical force, of course, but we have very strong and unique m orchestration capabilities that actually makes AI work in the enterprise. And which is why just this week we've seen, I think it was this morning, OpenAI announced that they were doing a big venture on consulting, deploying forward deployed engineers in the enterprise and had raised billions for that. Last week, dropped topic, made the same announcement. So right now the question is not whether AI works or not or whether AI is real or not. I don't think that's the question anymore. It's how do I actually make that work in the enterprise? And that's the 11% locket.

Speaker B: Welcome to Finscale, the show that takes you to the heart of finance transformation. Every week, Solene Meterkorn and her guests explore new initiatives shaping the future of the industry. Enjoy the show.

Speaker C: Hello, everyone, and welcome on Finscale. I'm, um, really glad today to have Rafael Dornado to talk about AI and hedge funds and many other things. Hello, Rafael.

Speaker A: Hey, Solene. How are you?

Speaker C: I'm fine. Thank you so much for your time. You're based in New York, correct?

Speaker A: Correct. In West Chelsea.

Speaker C: Okay. So, uh, Rafael, I know, uh, you from. I guess I heard about you like five, five or six years ago. Uh, you're an essec. Uh, you're a member of the Paris Bar. You started at Deloitte, then at Aid Advisory, and, uh, you founded, I thought it was an investment firm, an investment bank, a few years ago. And then while I was, uh, talking with you a few months ago, I realized that was completely different from my perception. So I guess that would make sense to stop a little bit, uh, on your professional path that you take, like five, six minutes to talk about your. Yes. Uh, what you've done so far.

Speaker A: Of course. So, uh, I've done a lot of studies. I love studying. I've studied finance, I've studied law, I've studied computer science. So I started my career, uh, indeed in the world of finance, working at Deloitte in Los Angeles, in Paris. Uh, after graduating from essec, I spent a couple of years there. I joined another, uh, boutique advisory firm doing the same work of financial diligence. And very quickly, it appeared to me that the way we were looking at business models that, uh, at the time, it was really the burgeoning of software in Europe. Back when I started the firm, we were not approaching this as I wanted to approach it and maybe we were lacking a bit of curiosity. So I was 27 and I said, okay, maybe I'm an entrepreneur. I did not even know what entrepreneur meant at that time. And I started my firm. I was just 27 years old. So I started, uh, the firm called Doranado and Co at the time, that is now, that has now been rebranded Decoding Discontinuity. And very quickly we built an approach to diligence companies in the field of technology in a different way. So this was financial and strategic diligence. And very rapidly we became one of the leading firms in Europe. So I completed over 800 diligence engagements across every single field of technology. Software, Internet, deep tech, et cetera. And, uh, three years ago, I decided to come to New York as we were really expanding our business with very big tech firms. And when I came to New York, I realized that I didn't want to be an advisor anymore. I had very strong convictions, and I was going to put capital behind my convictions, because in the US People tell you, well, if you have a conviction, you need to put your money where your mouth is. So I decided that I was going to go into the field of investment and since I had those very strong convictions, start, uh, building an investment firm based on the research work that I'm sure we'll be talking about.

Speaker C: Perfect. Yes. So now you have skin in the game. Uh, Rafael, um, can you maybe tell us a bit more about Decoding Discontinuity and what you decipher? Very, uh, very often. I guess once a week. Uh, just a few words about this newsletter because I believe that the audience knows a lot about AI but you go one step or two step. Very, very interesting. And I really, I, um, I tell my audience that they should really subscribe your newsletter.

Speaker A: Thank you. Thank you so much. So look, Decoding Discontinuity has really come to be my, I mean, my passion and what I spend like all of my time with, except for my family, of course. But I do spend a lot of time on Decoding Discontinuity. Look, shortly after I came to the U.S. um, this thing called ChatGPT came to the public right in November 2022. And at that time, after having done hundreds of diligence work, I said to myself, well, this is not just a simple disruption. This is something that is way more disruptive, like a tectonic shift. And it's going to have an impact that goes much beyond the technology itself. So I started doing a research around this topic to understand first where was the value Going to accrue in generative AI, huh? And how was that value going? How was value going to be dislocated? Much beyond generative AI software and the broader economy. So I started this platform on substack that has now come to be one of the, I'm very happy, one of the most read newsletters in the field. And this platform is where every week I expand, express my thesis in real time. So I use this as the grounding of my convictions for my investment fund where every week I develop. Well, this company has an amazing moat. This new technology is going to have a massive impact. This is the SaaS apocalypse, et cetera. And so at a point where I have now rebranded my firm, both my advisory and my investment platform behind decoding this continuity, because this is really at, ah, the forefront of how I view the world. It's like basically my glasses are on substack every week.

Speaker C: Okay, so, um, about your convictions, uh, you've done something that is quite interesting because you published your um, your investment thesis, uh, Orchestration, uh, economics, uh, it's publicly available, it's 280 pages. Quite amazing. Uh, congrats for the job. Um, question to you. Why did you decide to make it uh, public, uh, and to give access to everyone. It's very detailed, uh, very interesting. But I guess for your competitors it's also a source of inspiration.

Speaker A: Well, I mean that's obviously a very good question and I've been challenged on this point. Uh, look, I think that if you look at how AI is evolving with the rise of open source, I think there's a very strong parallel here. It's what matters is what you actually do with the ideas, the ideas themselves. Like everyone can have good ideas. The way I develop my conviction based on ideas that I connect together, like this is really about connecting the dots between so many things happening. I would love if everyone took the manifesto and copied my ideas, that would actually amplify my thinking. So that would be amazing actually. But thank goodness my alpha is not what I have published. My alpha is I have a worldview. My worldview is agenda. AI is a paradigm shift in which humans are not performing work solely anymore, but machines, software is performing work. So when you have a point where you have intelligence that becomes abundant intelligence that converges, where does the value go? The value goes to the companies that are able to orchestrate that intelligence and to use that intelligence to put it in a system and to actually produce tangible business outcomes. And that applies to software, but that applies to every single company in the S&P 500, CAC, FTSE, whatever stock index you choose. And so basically 100% of the economy is either the companies that are building the AI infrastructure or the companies that are being impacted by AI positively or negatively. So what I wanted to stress with this manifesto was that the way we think of moats, you know, the term that Warren Buffett helped popularize, imagine this castle. The way we have thought about moats up until now is you have a castle. And, uh, to protect the castle, you need to close. You need to close the door so that no one can enter the castle. But with generated AI, the attacks are coming from the sky, the helicopter that no one had thought about, uh, and they're coming from the foundations of the castle that are shaking from below, from above. So this is what is behind my world. The way we think of a business, of any business's competitive advantage, and whether it has a moat, meaning whether it has value, is wrong. So I want to correct that. I want to correct that by saying we first need to assess whether a business is architecturally resilient to generative AI. Is it on the wrong side of history or is it on the right side of history? If you're a software company or any other service company and you're being abstracted by Cloud or OpenAI, you're on the wrong side of history. What is the point of looking at competitive advantage, at looking at the financials with the most granular detail, if your terminal value is zero? So the approach that I, uh, propose in this manifesto is really a worldview of how architectural resilience trumps any other form of assessment as a preliminary filter. And that's what I want to help popularize, if that makes sense. And then I have the secret sauce on how you actually apply that.

Speaker C: You, uh, published an interesting, um, substack, ah, article called the 11 paradox. Um, and that says that, uh, 11% of enterprise builders switched AI providers last year, despite the switch, uh, being technically trivial. Um, and, uh, that is your core argument on orchestration. Lock in. I guess that would make worse digging into this 11% paradox, of course.

Speaker A: So look, when in May last year. So one year ago, when we were like, really in the middle of this race with OpenAI and Tropic, XAI, Mistral and all of the big AI labs, I published on Decoding Discontinuity, a thesis in which I said, the most of the AI labs is no more technical superiority of the model, but it's actually the orchestration. And that was like, everyone looked at that article and I was like, well, what matters is how you actually put the LLM into a production. How does the LLM call tools, how does the LLM fetch the context, how does the LLM talk with another LLM, et cetera. And this is what I call, and what I call what everyone calls orchestration or the hardness, which is a very popular term. You hear a lot about hardness engineering right now. So what we see with this 11% lock in is that companies will maybe switch between models. But if you are able to be the orchestrator in the technical sense, so if you're able to actually make that intelligence function in an enterprise environment, well then the switching becomes much more complicated. And this is what we have seen over the past year and really into like H2 of last year, where the Moats and Anthropic with cloud code, but OpenAI is doing the same thing, have really moved from, we have technical force, of course, but we have very strong and unique orchestration capabilities that actually makes AI work in the enterprise. And which is why just this week we've seen, I think it was this morning OpenAI, like announced that they were doing a big venture on like consulting, deploying forward deployed engineers in the enterprise and had raised billions for that. Last week Tropic made the same announcement. So right now the question is not whether AI works or not or whether AI is real or not. I don't think that's the question anymore. It's how do I actually make that work in the enterprise. And that's the 11% lock in.

Speaker C: Okay, thank you very much. Uh, you decided to launch a hedge fund. And why is that? Because I would have believed that you would have raised rather a venture, um, capital fund, uh, something like that. But why did you decide to go for a hedge fund structure, of course?

Speaker A: Well, so at first I was pushed to do a venture capital firm, which I refused to do. So I can tell you more on that story. Look, my thesis again, if you take my glasses of like this image of the castle and how architectural resilience is more important than anything else in this current environment. My, my conviction beyond my thesis, because before a thesis you have a conviction is the whole economy is being dislocated. Some companies will be immense winners, will go to the moon, and many of them, and some companies are going to be going to zero by 2030. Uh, so when I wake up in the morning, that's how I view the world. Companies that are going to the moon, companies that are going to zero. Based on that, my methodology allows me to assess that In a quantifiable way and to actually put all of the relevant data points based on which companies are going to the moon and which companies are going to zero. So by having a headshot which trades on all of public assets in the us, In Europe, in uh, Asia, being able to see wool, I think those 30 companies are the winners and I think those 30 companies are the losers. Software, non software and AI infrastructure is for me the best expression of my thesis. Venture capital. I'm very close to many great venture capital funds who I have super interesting conversations with and who read decoding this continuity. And I have a very valuable relationship by really learning on um, what are they investing in. Where is the innovation going? Why is I uh, don't know like uh, vocal uh, AI or like this new uh, neuro symbolic AI etc. Understanding where venture capital dollars are going, huh Is key in my, my framework because if all of the dollars go in one specific segment, you know that the companies are going to be indirectly impacted in the public market. So I would not be able to work if I didn't have this vision first my background, because this is what I was diligent in before and this ecosystem, I think it's one of the key components of the fund.

Speaker C: And you're concentrating these investments in on the US market S&P 500 or other markets?

Speaker A: No, I mean I have approximately uh, 85% of my portfolio which is in the US I have uh, many stocks in Japan and in Asia where I think there is fantastic opportunities. Those are mostly long positions.

Speaker C: Okay, and um, you are ah, first fund manager. This is the first. You don't have credentials. Um, so how uh, were you able to raise. Of course you have a very strong community and very brilliant people reading you. But there is a very important step, uh, when you raise a fund and people ask for uh, credentials, uh, how did you uh, convince them that you were a good choice?

Speaker A: Well, so first I was very fortunate to uh, meet a very well known investor here in the US who uh, has allowed me to do uh, consulting for him and whom I really been able to test my ideas and uh, help run a portfolio to really like test this live and not just go from generating ideas to actually selecting names and uh, running a live portfolio. So I've been very fortunate on that. I think that when you have paradigm changes like the one we are living in, having a first principles approach is actually what matters the most. And I'm seeing in all of the conversations that I'm having with key institutional allocators that when you arrive and say, I have a worldview. This is how I see the economy moving. And this is not because I, I woke up this morning thinking that, but because I have a very grounded approach that I have been building and showing to the public for the past three years, which is, by the way, delivering very strong performance on the portfolio. People at first think that you're not serious and then they very quickly start listening to you. So it's been, I mean it's very frustrating because I don't come from, oh, I was not working at Tiger, at Quatru, at whatever. I'm French, I'm a woman though. I don't. I have not seen that as a problem up until now. I mean, I still don't see it as a problem, but I think it's like, okay, so she's different, but I can look at her ideas. I can see that, oh well, she said things on Figma, on Core Weave, on Fermi. Everything turned out to be right. So maybe we should listen and you know, by actually being able to back test your work, to say, look, I said that on Core Weave, I said that on um, Sigma. And like actually look at the report. You should, you would not have invested in the IPO if you had read my report or you would have invested differently. Then you build your track record like that and it's a long term game. I've realized that this is, this is going to take time.

Speaker C: Mhm.

Speaker A: I have a very strong ambition in this project and it's not happening overnight. So I think when you choose your race, when you're firm on your convictions, on your approach and you don't even. There's many days where you're wrong, where the economy goes in the wrong direction. So you're not like, it's not always working well. But when you're firm on, okay, this is my vision, this is what I want to do and I'm working super hard to get this going. You will find people who will trust you and who will support you in this journey. You're the one who needs to be very firm.

Speaker C: Yeah. M. And many entrepreneurs, I guess, supporting you.

Speaker A: Right.

Speaker C: Um, Rafael, you have a dual culture. So you're based in the US but you are French, as you said. Um, I saw you spoke at the Milken Institute, uh, on eu. In your opinion, what is the most striking difference, um, with allocators, when you discuss with allocators in the US and difference with Europe, not in terms of uh, uh, capital allocation, but rather in the logic.

Speaker A: Look, I think that um, the U.S. it's so much harder, I would say that to get someone's attention. In the US it's so much harder. And a lot of people from Europe struggle with that because here it's, you need to be the best at, uh, what you are. And if you're the best at what you are, then people will forget that you didn't come from this environment, that you didn't. I mean, of course where you come from matters, what you've accomplished. I mean, like in any country. But I would say here, when you're able to defend a conviction and have very strong ideas and again, do the work here, I mean, people work very hard to succeed. I would say that's the first difference. It's not, oh, I know this person, I know that person. Now, uh, here it's, you don't know anyone because it's, it's. I mean, the fight for attention, I would say, is, can be surprising for many people who move to the US I think allocators are, I don't actually see, I mean, not that I don't see a difference, but I think allocators here maybe are, they're actually funding this AI revolution. They're putting depth behind coreweave, behind Fermi, behind all of these gigantic AI projects. So I would say there is a difference in allocators. It's the fact that maybe we're a bit, uh, a bit like in advance from a time perspective, because this is where the AI labs are. This is where the massive amount of spend is being done. So I would say the awareness on this whole AI and agentic AI is not that people are smarter, but that because of the proximity to the AI labs and to the capital, it's closer,

Speaker C: it's m. More mature in the understanding, I think.

Speaker A: So there was an interesting, um, uh, tweet this weekend with someone who said, okay, when you're with the AI labs, it's zero. When you're in San Francisco, you have three months difference. When you're in New York, it's six to nine months. And then when you're in Europe, it's like more. And I mean, there is actually a part of truth with that. It's like if you're literally living with the AI labs, you will know more things than if you're in SF two miles away or in New York. But anyways, you're still closer than if you were on the other side of the Atlantic.

Speaker C: And in terms of size, um, and performance, what are your expectations? I would say in five years it's Hard. I know it's hard. Two years.

Speaker A: No, look, I mean, uh, things are going the right direction, of course. I mean, I'd like this fund to be, I, uh, mean over a billion. But it's. What matters is the performance. The performance has been, uh, very, very strong. I cannot disclose performance, but I can say it's been, uh, very superior to many other funds. Uh, partly because I, uh, was maybe ahead on the SaaS apocalypse. So I was, uh, short on a lot of software names starting in November. And so obviously that played out very well over the beginning of the year. Not when everyone was realizing that software had a problem, but before. And with a nuanced approach where I think a lot of software companies are actually in a good position. When a lot of software companies are effectively going to zero. And when people wanted to say, oh, the SaaS apocalypse is over, I said, no, the SaaS apocalypse is not over. The SaaS apocalypse is just beginning. So performance has been very strong. It needs to continue being strong. Again, this is a game where the numbers are the numbers and so you can have the best ideas. If you don't deliver performance, you need to find an average job, which is, I mean, that's the case of hedge funds. Right? So, I mean, I'm very aware of that. Uh, again, I'm here to build a very strong franchise. It's what I did with my advisory firm, uh, over the past, um, 12 years, brought it from Paris to New York, was able to work with top tier investors starting from nothing. So again, here I started from nothing. I was able to find tremendous support. I have a lot of people helping me in this journey. I'm very confident. I just learned to be patient. This takes time. And it's not by saying, oh, I want to have 10 billion, 1 billion, no, it's okay. Am I able to deliver results constantly? Am I able to work under pressure? Am I able to work with all of the volatility that is on the market? And I'm learning my way into this, uh, into this. Not a new job, because I've been doing this for a long time now. But I'm. Look, it's a fascinating journey, clearly.

Speaker C: Let's stop a few minutes. On the SaaS, uh, ecosystem. You said it perfectly. But then you have been following SaaS for years now. Um, would you say there is a topology of SAS founders or SaaS, uh, companies that will, um, win rather than the others that will lose? So, uh, what are the main differences between the losers and the winners?

Speaker A: Well, look, first I want to Say something that is important and I think we tend to forget it's that agents are software. Like what is an agent? An agent is fundamentally a unit of software. So when we say software is at risk, what we're saying is software as a service, SaaS as a way of delivering SaaS is at risk. But software is of course I mean agentic AI is software. So I think, I mean it's the first very important distinction because we tend to put everything in the same category. I think that what is happening with uh, software and not SaaS is that we used to have a split between application, uh, software infrastructure software and within application software. We had a split between horizontal, vertical and the silos that helped shape how we were approaching companies, even intelligence at the time. I think those all go away in the new agentic AI world. If you think of how companies are deploying software now. Companies are deploying fleets of agents, fleets of agents that need to produce a uh, business outcome, that are working together, that are fetching context, that are talking to a user through a ah, different front end. That implies that a software company reorganizes itself as both infrastructure being the ground for the agents thinking about how it actually delivers the application value to the end uh customer through an interface or as uh Salesforce has done with Headless360 directly through an API, which is a very new way of delivering software. So I think that is what is happening right now. It's the nature and the purpose of a software platform is where can you deliver real work in the enterprise with agents that are able to again make the right decisions because they're grounded correctly, able to talk to each other, the agents of the agents, to talk to a human and to render result. That is what is at stake. Thinking that software is going to be delivered as on a per seat basis through a regular interface and that that is not at risk. You can be vertical, you can be horizontal, you can be whatever you want. That is going away for sure. So it's a question of it's not oh everything dies in software, it's software reimagines. And I am very wary of a lot of defensive positions on um, software right now that I'm hearing even from top tier companies which are saying LLMs are probabilistic. The enterprise needs deterministic software and so software is going to be a layer between enterprise work and um, probabilistic LLMs. I personally think that that is wrong because of course the AI labs are going to go into making this work in some way or the other, I mean they're smart. And so the value of a software platform really needs to be of, uh, how do I get work done in the enterprise and what do I bring that is unique and that allows to actually accomplish an end result? That's for me what companies should be thinking about. Not thinking that there is this opposition between LLMs and real work in the enterprise. I think that that is going to be short lived.

Speaker C: Okay, um, Rafael, um, one last question because I'm always worried about that and this is something I cover on my other podcast called from within the Soul of Business. Considering that many companies will go to the M moon and some will go to the floor, um, do you think that companies address the human impact on themselves? Um, meaning upskilling people, ah, reskilling people. And bearing in mind that it will have a human impact, uh, on the economies in general, is it something that you see, uh, as considered or completely forgotten in the strategy of these companies?

Speaker A: Look, I would say that we're, we're seeing an evolution and I'm kind of, I would say I'm seeing more positive evolutions over the past weeks and months than maybe what we had at the end of last year. I think that now the model capabilities and what can be accomplished with AI is starting to be more and more clear. So we are not in this unknown anymore of is this even real, is this actually going to deliver value? I mean again, there is a long way to go. But I would say people are more okay with what, what does this bring and are more realistic. So because of that, they can now say, okay, now how do I make that function with humans, us in the loop and how do I make this actually work where I have humans that are very necessary, that are going to do their work in a way that is very different. So I would say that companies that are acknowledging that AI is going to be front and center of every single workflow, again with the appropriate orchestration, with the right harness, with everything that makes it ready to work in the enterprise. I, I'm assuming that this will be the case in one way or the other, then you start to rethink every single one of your workflows with intelligence at the core, intelligence that has been abundant and again that is converging in capability. Once you have resolved also the cost issue, which has been a big debate over the recent weeks, and then you start to rethink, okay, how do I leverage the humans in my organization to allow for this human agent chain? Maybe I'm very naive, but I think that Companies that are, that are seeing the truth in front and are able to say, okay, this is how I'm reorganizing the company. And it's going to take time because this doesn't happen overnight. And this is again, true for software, but it's true for banks, for insurance, for any company you can imagine. Like, even the company with, like, physical operations. Of course, I would say we're in that phase right now where humans are key, but we're not hiding our faces anymore saying, oh, AI will never work. That's rubbish. Blah, blah, blah. No. Which was the case, but which was still the case at the end of

Speaker C: last year, in my opinion, this is my personal opinion. I believe that we have not envisaged the impact that it will have either on users, ah, a human and its, um, cognitive impacts, but also on, um, the companies themselves. And for me it's really a good, um, thing in the sense that it will force us as humans to really think, what's my, as a human, what's my added value to the company? And to really concentrate on our real added value and not as, uh, I mean, concentrate on the who we are rather than what we do. Rafael, it was super interesting. Thank you very much. I have one last question. As you know, um, I always, uh, ask my guests a book recommendation or podcast. What would you like to recommend today?

Speaker A: Uh, so as a book, I'm reading, um, the Love Letters from Albert Camus. And it's amazing. It's 1,200 pages, but, uh, it changes from AI and it's a wonderful story.

Speaker C: I love that. Thank you. Thank you so much. Uh, very useful. And I invite my, uh, audience, uh, to go, uh, on your website and check the manifesto and the investment thesis, which is really super interesting.

Speaker A: Thank you so much. Thank you.

Speaker B: Thank you for listening to finscale. Don't hesitate to subscribe to the newsletter on Substack to stay up to date with the latest innovations. You can also find the podcast in audio format on all listening platforms and in video format on YouTube. See you soon with Solen Metercorn.