Playbook duScale

The fundamental question isn't "how much do I make?" but "what game am I playing?". The world rests on a simple social contract: 1% who create, 9% who comment, and 90% who endure. Each league has its upsides and downsides.

Money isn't the benchmark. A VP at Goldman Sachs can earn more than an entrepreneur — but he hasn't taken any fundamental risks. What matters is the alignment between the game you choose and how you play it.

If your business survives without ever taking off, every extra year grows your regret logarithmically. Oussama had already given Olivier the same diagnosis three years earlier. The fact that nothing has changed since is proof that three years of life haven't been lived to their full potential.

What's going to make the difference in the world to come is two things: will and execution on one side, taste on the other. Knowing what's right or not. Knowing what's good or not. You can be more or less intelligent, but intelligence won't make the difference anymore. AI has commoditized it. It's an absolute paradigm shift.

It's action that creates information. You don't learn things to do things — you learn things because you do things. Stop asking questions whose answer is a test. Oussama calls this the "Catechism according to Ouss": the phrase to repeat every morning when you wake up.

Oussama directly confronts an entrepreneur downplaying his product on stage: "You should all go get burned for being ashamed of selling." Shame about selling is a fatal handicap in entrepreneurship.

During the live, Ahed, an IT consultant, explains that he has "no commercial skills" and that big corporates "will never do a deal with a company with 1,000 € in capital". Dry reply: "They don't give a shit about your capital. What they care about is whether you solve an exciting problem or not."

Oussama's decision matrix in four boxes: Irreversible head decision → think hard. Irreversible heart decision → avoid as much as possible. Reversible head decision → don't think at all. Reversible heart decision → think about it. Most entrepreneurs invest brain time in questions that aren't worth it.

The mammoth hunter was at the top of the social hierarchy — his job has no value today. Nobody would have predicted we'd pay someone $50,000 a day to act like a fool on Snapchat. Every era creates new, unpredictable forms of value. Rather than waste time predicting the future, stay on top of the wave in the present.

There is no merit in entrepreneurial success: there is only randomness and luck, which can take more or less laborious forms. The universe doesn't care about merit. This view frees the entrepreneur from personal justice narratives that paralyze action and generate guilt. What matters isn't deserving success but creating the conditions where luck can operate.

People don't pay because a product is good. Price is perception. Perception is storytelling. Storytelling runs in marketing. It's the intensity of the marketing that justifies the price — and the price that funds the budget to build a good product. The good product, you build it for yourself, not for the market. You have an engineer's ethics, an engineer's habitus? Good for you. But understand it's a personal matter. The market buys a price, not a product.

At The Family, Oussama would walk into a room and tell a random entrepreneur: "Double your prices." 99% of the time, the founder doubled under pressure and lost zero clients. Only once did it create a disaster. The ratio speaks for itself. Bruno, founder of a rental-management SaaS at 5 € per month, is told he should be at 1,000 € per year paid upfront.

If you sold at 1,000 € per year instead of 5 € per month, you'd have 990 € of acquisition budget to go grab the next client. That's the acquisition loop. A high price funds the ads that fund the growth that fund the quality. A low price creates a vicious circle where you never have enough budget to grow.

Every morning, a checklist. 10 minutes LinkedIn, 10 minutes newsletter, 10 minutes Twitter, 10 minutes cold email. What works organically, boost it with paid ads. Never separate organic and paid. Acquisition has to become a routine as mechanical as brushing your teeth. When a piece of content performs organically, push it a bit with ads. That's where the magic happens.

The 4 quadrants of acquisition: Free direct (organic content), Free indirect (word of mouth), Paid direct (Meta, Google ads), Paid indirect (sponsoring, events). Test everything to find the channel on which you put 80% of your effort — 20% on the rest. But first, test everything.

Call your best American client. Tell him you're struggling, that your company might not survive, and that you'd appreciate some intros. Reaction in France: the guy hangs up. Reaction in the US: magic happens. Transparency and vulnerability create trust and unlock networks.

Why build something you haven't already sold? Oussama tells Ahed, IT consultant: "Go to GenSpark.ai, produce slides that pretend the product exists, and go find 4 clients ready to pay 100,000 € for something that doesn't exist yet." Paradoxically, selling a non-existent product to a big corporate is easier than selling an existing one — because it goes through the innovation department, not procurement.

For high-ticket businesses: host a monthly dinner, 20 people max, in the 3 key cities of your market. Find a friendly sales rep in each city who organizes and invites. Budget: roughly 5,000 € per dinner. Re-invite the same people. If out of 60 people reached per month, 3-4 order, it's done. Infinitely more effective than 20,000 cold emails that don't convert.

Call 50 of your clients on a video call. Record with Firefly or equivalent. Ask questions. When they say a powerful phrase, cut the clip and ask: "You said that during our call. Can I use it as a testimonial?" 90% say yes. You turn 300 users into 100 testimonials, i.e. three videos a day on social. Social-proof carpet bombing.

When you sell to different networks or volumes at very different prices, the trick to avoid comparisons: publicly display a high price, grant confidential discounts to big networks, and get every member to sign an NDA on the negotiated price. This is exactly how Oracle sells its databases at -90% to some clients without others finding out.

Scaling is doing more with less. It's not "doing bigger", it's not just growing. It's increasing returns to scale. Every successful company delivers decreasing quality — that's the tragedy of scale. The goal is to slow that decline through automation.

Complexification is the enemy of scale. Oussama draws a chart live: complexity rises like this, management capacity is like this, and the intersection of the two is where chaos appears and companies die. This is exactly what a SaaS is for: getting out of middleman complexity, standardizing, avoiding the bottleneck.

Amir and Abdel Ali, with their facility management company, get told: "Build the "nth SaaS", build the "nth marketplace". Stop believing that being different is the key."

The brain solves conversion rate, catalog qualification, pitch. The muscle solves lead volume, cold email, follow-ups. The brain should only kick in once all the muscle solutions have been deployed. Most entrepreneurs do the opposite: they think the right idea will save them, when they haven't even ground it out yet.

It's not enough for your product to match the market. Your way of building has to match who you are. Building in public isn't for everyone. Cold calling isn't either. AI lets you build a system personalized to your personality: if you prefer writing, write; if you prefer talking, talk. But don't copy someone else's process.

Stop the fantasy of "let's hire the best". An entrepreneur's dream shouldn't be to have the best team in the world, but to make money despite having the most average team in the world — thanks to process, thanks to AI. It's normal employees aren't passionate. That's why we all want to replace them with AI.

Claude Code does the work of an engineer at 10,000 € per month — for $200 per month. And you still have mental barriers? The French-speaking ecosystem suffers from a disease: penny-pinching. Counting tokens, hesitating, hunting for free. Even in the worst situation, you can work four days as a dishwasher at 50 € to pay the subscription. Your problem is a problem of will, not of means.

The goal isn't to re-read what the AI did. It's to define objectives, tests, success criteria, and let the AI run infinite loops. Coding with AI isn't coding like a developer. That's why the creator of Open Claw isn't an engineer — and why OpenAI buys it for a billion after 96 days of work. The point is to be on equal footing with the AI on architecture choices.

Every quarter of an hour, for a week, write down what you're doing. On Sunday, give everything to Claude and look at what the AI can automate. Start over every week. Week by week, you optimize, you automate more. It's as dumb and as powerful as that.

Oussama explains Norbert Wiener's cybernetics (1960s): a radiator has a state (the temperature), a thermostat (the target), an output (heat) and an input (the ambient temperature). It produces heat until it reaches the target, then stops. AI systems work exactly the same: input, state, output. Three vector variables X, Y, Z. Mastering this self-feedback loop logic has become mandatory.

The richer your context files, the more the AI knows you. Oussama reveals his structure live: "core" folder (identity, anti-patterns, constraints), "psychology" (interpersonal dynamics, full psych profile), "work" (portfolio, investment thesis), "intellectual" (books, concepts, films, worldview), "style" (writing in French, in English, rhetorical weapons), "technique" (stack, tools, workflow), "team" (his AI employees: Alan Turing for backend, Loki for chief of staff, Steve Jobs for brainstorming). Far more powerful than one-off prompts.

Open Claw, well-configured, works without you being in the loop. Claude Code / Cowork works with you, on your instructions. An agent can use a tool — your Open Claw can use your Cowork. Also, match the model to the need: Claude Opus for complex tasks, Sonnet for day-to-day, Codex to challenge, Gemini for SEO. Stay curious, test every model constantly.

Developers and engineers who try to copy personal branding or storytelling approaches fail because they're playing outside their strength zone. The right approach is to treat go-to-market exactly like an engineering problem: build automated systems, optimization loops, agents that test copywriting. Today you can download the "brain" of a marketing expert via an AI's skills.

French-speaking culture has a passion for "good content". But content alone isn't enough. You need mass. You can't have 30 original ideas per day, so you have to find what works and repeat it. Copy the Americans: they're under pressure in a massive market, so they're naturally better. What works in the US works in France.

Record a one-hour face-cam video on YouTube. AI agents turn it into LinkedIn posts, tweets, newsletter, Instagram carousels — 100% automatically. This is the system Oussama recommends to Adam, 16 years old, who has no time between school and his startup. For 1 hour a week, you get 7 days of intense content.

Robin the cabinetmaker has soccer players following him on Instagram without knowing it. AI could have identified them, tagged them, surfaced them as premium prospects. Better than a form: have AI talk to 100% of people interacting with your account. An American craftsman got Oussama to buy furniture simply because his bot sent him a DM after he liked four videos.

Abruptly switching from French content to English on an existing account risks "polluting" the algorithms with French-speaking followers who penalize the geolocation. The right strategy: create a private American account, hit 10,000 followers before going public, and never announce the switch to your French audience to keep them from going to check out the English account.

To get the first 1,000 subscribers, the best psychological hack is to act as if the newsletter were paid and "offer" it for free. The basic human mechanism: if something looks like it has value (and is normally paid) and you get it for free, you're much more inclined to sign up and open it. This technique also boosts open rates.

When AI commoditizes intelligence and execution, everything will become like fashion. We no longer really tell the difference between real quality and perceived quality. A Dior t-shirt costs more than an H&M one not because it's better, but because Dior has a mystique and a history. SaaS will follow the same path: people won't buy them for the tech but for the brand.

Y Combinator wrote it in their Request for Startups. People aren't interested in how they buy things — they're interested in the outcome. As software becomes a commodity, the outcome isn't. Build agencies that are SaaS in disguise: internal automation, human ceremony on the front.

The ceremony — the meetings, the documents, the reassurance — is what people buy, even when AI does 90% of the work. It's the difference between a 3 million logo and a 20 € logo on 99designs.

Counterintuitive but brutally effective. When you're lost between multiple strategic choices, pick one at random and dedicate yourself only to it. If it wasn't the right one, you'll find out faster than by staying in the fog. Start over. Method inspired by Bernard Werber in The Encyclopedia of Relative and Absolute Knowledge: flip three coins to determine the combination. Execution gives you information — not the other way around.

Work on all your projects every day, even 5 minutes each. What doesn't fit into the day shouldn't fit into the week. No backlog: recurrence and routine. It's Christophe Delone's lesson on multitasking: never push things across multiple days. If you have 8 projects, you work on all 8 projects every day.

Agency — the capacity for autonomous action — is the most important muscle of the next ten years. It's trained through repetition and time constraints. Set yourself rules: build something in one day, launch a site in one day, write a song on Suno and publish it on Spotify in one day. Like in a video game, side quests build up the XP for boss day.

Dopamine issues, thyroid, testosterone? We underestimate biology. There's no point flagellating yourself if the body is malfunctioning. Go see a functional doctor. Once the mechanics are checked, motivation becomes secondary behind discipline. Discipline is doing things even when you don't feel like it.

Life has four "burners" — family, friends, work, health — and you can never have more than two lit at full power at the same time, not enough gas to fuel all four. Wanting to excel in entrepreneurship AND in your studies at the same time is an illusion: you have to choose which burner you light up high and accept the trade-offs.

A brilliant idea without execution is worth absolutely nothing. The startup graveyard is full of revolutionary ideas that never saw the light of day. What makes the difference is the ability to turn that idea into reality, to overcome daily obstacles, and to adapt to market feedback.

Don't protect your idea, share it. The real value lies in your ability to execute it better and faster than others. Action is the only true indicator of progress.