When Bending Spoons acquired Evernote, the company had about 350 people, roughly $90 million in revenue and close to zero profit. A year and a half later it ran on 50 to 60 people. Today about 20 people fix its bugs, ship its features and tune its monetization, and Luca Ferrari says the business is very profitable, with revenue between $100 million and $200 million.
Ferrari, co-founder of Bending Spoons, doesn't present this as a cost-cutting trick. He presents it as the predictable output of a company designed around a few unusual convictions: raw talent beats experience, work should always exceed capacity, and complexity has to justify itself or be deleted.
Those convictions came from an unlikely place. Ferrari grew up in a northeastern Italian village of fewer than a thousand people, the son of hairdressers, and built the company first in Copenhagen and then in Milan, far from any startup ecosystem. "We were absolute nobodies," he says. Nobody was around to copy, so they reasoned from first principles and accepted that they'd make more mistakes than someone following tried-and-tested playbooks.
Isolation turned out to be a method
Ferrari says he doesn't read business books or follow podcasts, and the company's operating model shows it. The one outside influence he credits is Netflix's culture deck, specifically the idea of beating complexity with talent rather than process. He also borrowed the phrase "extreme ownership" from Jocko Willink's book, though he says Bending Spoons defines the concept differently.
The stated ambition is deliberately extreme: build "the best company there ever was." Ferrari concedes it's likely impossible. His argument is about payoff curves. Rewards, both emotional and material, sit at the extremes, so a person with limited time and energy should pick one or very few pursuits and go all out. He applies the same logic to his marriage, which he rates close to 10 out of 10 and considers worth a hundred times more than an adequate one.
He also tries to erase the founder identity from the company. Contribution and trajectory matter, not who was there on day one.
Experience goes stale, talent doesn't
The thesis traces back to a failed startup. Ferrari's first company, Evertale (2010 to 2013), used AI to write a person's diary automatically. It never scaled commercially, but the team of about 12 taught him two things. The best person on the team was easily ten times as productive as the median one. And that person was among the least experienced.
Ferrari gives several reasons experience is overrated in software:
- Most tasks in the industry need a good brain and a drive to excel, not decades of accumulated knowledge.
- Customer expectations and engineering practice change fast. Writing software in 2010 bears little resemblance to doing it in 2026.
- Experience can make people worse. Someone who has normalized low performance standards, or learned that advancement comes from pleasing superiors, arrives with habits a meritocracy has to unlearn.
Experience can be supplied by the company. Talent has to be hired. So Bending Spoons hires for talent and then gives people hard challenges and strong colleagues so the experience accumulates quickly.
Talent level also dictates management style. With mediocre talent, Ferrari says, process and procedure become the lesser evil because you need to guide people. With a dream team you want as few rules as possible. His analogy: you coach Jordan and Barkley differently than a team of modest players, and Bending Spoons tries to assemble the former and then get out of the way.
Hiring as a quantitative discipline
The obvious objection is how to spot talent in someone who has no track record. Ferrari's answer starts from a statistical point. Results are a reliable signal only over long horizons. Thirty years of investing results tell you Warren Buffett is skilled; one year of 50% returns could be two lucky picks. A graduating student has no such sample.
So the company built what Ferrari calls a science of hiring, staffed with engineers rather than traditional HR people. It runs tests designed to proxy mental capacity, and for each applicant it identifies over 100 signals to predict long-term potential. He compares it to quantitative hedge funds, which combine hundreds of individually weak signals into a predictive whole. Some signals are obvious (GPA tells you something). Others are subtle.
His example is collaboration. Interviewers could judge whether a candidate seemed open to criticism, and that judgment turned out to be a poor predictor of real behavior, since smart candidates know to be pleasant when it counts. What did predict it was how candidates treated the logistics staff who schedule interviews, a role that doesn't feel connected to the final decision. Anything qualifies as a signal as long as it's predictive, even if it isn't intuitive.
Why hiring and exits sit in one central team
Bending Spoons centralizes hiring and parting ways across all its businesses. Ferrari argues the standard model, where a team lead picks from a few pre-screened CVs, has bad incentives built in. Managers want the seat filled quickly, so they pick the first adequate candidate. They prefer fully mature hires over green ones who might become much better. And a manager who hires three people a year has no sample to learn from and no professional pride in interviewing.
A central talent team does nothing else. It sees the whole pipeline, knows statistically whether waiting will produce a better candidate, and can redirect someone who applied as a product manager into growth. Ferrari connects this to a maxim the host recalls from Brad Jacobs: an empty seat is less damaging than a poor fit. At Bending Spoons the seat barely exists as a concept, because "we all have the same job," which is helping the company succeed.
Those talent managers get no bonuses or variable pay. Ferrari says concrete objectives tied to pay or promotion "almost invariably" lead to inferior outcomes, because people optimize for the metric rather than the company. His evidence is experiential: the best projects he's seen weren't run on mechanistic KPIs. The model works only if the people are intrinsically motivated, which is why the selection matters so much.
Give people more work than they can finish
The most counterintuitive practice is deliberate overload. Everyone, especially promising juniors, gets far more on their plate than feels comfortable.
Ferrari's first argument is mathematical. If you choose among ten tasks, each with an ROI, adding an eleventh cannot lower the return on what you pick, since you can still do the original ten. It can only raise it. This holds only if people choose well, which is why leads are coached above all to help reports prioritize.
The second argument is about training. Someone given a short task list that fits in eight hours never learns to live with what Ferrari calls "the immensity of the possible." Someone who learns early to be comfortable with a hundred times more options than capacity can select surgically, which he calls an "insane superpower." Better to build that tolerance in the first year than discover at year ten, running a billion-dollar organization, that you're overwhelmed.
Most work, he adds, is elective. Winning starts with doing the small fraction of projects that are ROI positive, in order of return.
Extreme ownership as the filter
Intelligence needs to clear a fairly high threshold, Ferrari says, and after that the differentiator is how badly someone wants to be great. He cites Rafael Nadal, who by expert consensus was not a top-50 talent of his generation yet ranks among the best ever. The company has seen people near genius-level IQ fail because the job was priority three or four in their lives.
Screening for it means looking for past evidence: someone who worked heavily alongside their studies, contributed volumes to open source, or ground at a startup for three years without letting up. Ferrari tells new hires that if they bring this mindset, there's less than a 1% chance they won't have an amazing career. He notes it's contagious in a high-density team and that diluting it makes the people who have it leave or lose it.
Rotation supports the same culture. People move across businesses because fresh eyes generate ideas, because subcultures form when teams stay fixed, because humans get bored, and because new acquisitions often offer the highest expected return on someone's time.
Complexity carries the burden of proof
The second cultural principle is relentless simplification. Ferrari's claim is that left alone, any organization adds parts (features, process steps, teams, rules) and rarely removes them. Each part also adds interconnections with the others, so going from three pieces to four makes a system more than 33% more complicated.
Two rules follow. Anyone proposing new complexity carries the burden of proof; opponents just raise a flag. And everyone is asked to hunt for existing complexity, with suspicion rising the longer something has been there, because humans go blind to a static status quo.
The biggest wins, he argues, come from complete removal rather than incremental trimming. His example is job titles. About a decade ago the team was spending emotionally draining hours defining who deserved "senior" or "director." Asked what titles were actually for, they concluded the benefits were bragging rights and CV signaling, neither of which the company needs to manage. So they abolished titles. Employees can write what they like on LinkedIn as long as it's reasonable.
Internally, an algorithm labels people by the function of their direct reports, so the head of product and someone leading one product manager are both "product management lead." Out of thousands of employees, no one has complained about the absence of a title. Ferrari estimates the savings in hundreds if not thousands of person-hours, and describes complete deletion as potentially a 10x improvement against 5% from tinkering. The same applies to products: cut the feature used by 2% of customers and serve the other 98% far better.
An operating system for acquired companies
Ferrari describes more than 50 proprietary tools, built over roughly a decade, that function as an operating system into which acquired companies are installed. There's one system for payments, one for A/B tests, one for predicting user lifetime value, one for recruiting predictions, one for orchestrating the AI models used internally, and so on. Platform teams own the tools, and each acquired business can add features and fix bugs that propagate to the whole portfolio. In that sense, each acquisition brings another source of improvement ideas.
It's also why Ferrari says Bending Spoons isn't private equity. Funds buy to sell, intervene mostly on costs and prices, and can't integrate their holdings without making them harder to sell. Bending Spoons has never sold a material business, rebuilds products and infrastructure, and moves people fluidly across companies. Ferrari rejects the "conglomerate" label as well, since a conglomerate is a set of distinct parts and Bending Spoons tries to make everything homogeneous.
What AI changed
AI use predates the current wave (it informed lifetime-value predictions and testing as early as 2018), but Ferrari says LLM progress over the past two years has produced real breakthroughs in engineering, data analysis and design. He gives two internal examples.
Diagram, an in-house design tool, pulls up an app's screens, follows the head designer's guidelines automatically, reads the codebase to understand how interfaces interact, generates code for review by the lead engineer, and plugs into the A/B testing system. A product manager can now do in perhaps 1% of the time what used to need a designer, and without the lossy back-and-forth of explaining an idea to another human.
Old Spooner (alter ego or alternative Spooner) is a Slack agent with the same access permissions as the person invoking it. Ferrari watched Evernote's general manager tag it with a bug report: check the customer support tool to see if the problem is widespread, search the codebase for root causes, propose a fix, and ping the lead engineer for review. In about three minutes she had an identified and fixed bug that human-to-human coordination might have stretched to weeks. Ferrari used it himself to pull user trends by country and device in minutes instead of waiting days for an analyst.
Ferrari says the tools aren't for sale. They're too sophisticated for most buyers, they work best because they're natively integrated with one another, and they're a competitive advantage he'd rather keep.
The arbitrage behind the acquisition math
Why could Bending Spoons run Evernote with a fraction of the headcount? Ferrari offers four reasons, and none is a hidden insight about the business itself:
- Talent access. Bending Spoons received 800,000 job applications last year and hired fewer than 300 people. A standalone Evernote can't draw that pool.
- Risk tolerance. A standalone CEO who halved headcount and watched it fail would be finished. Inside a portfolio, positive-expected-value bets are easier to take.
- Shared technology. Platform investments are amortized across every business, which no single company of that size could justify.
- Freedom from vanity metrics. Public markets punish a falling subscriber count even when higher prices raise revenue. Inside Bending Spoons, Evernote can be optimized for cash generation.
Ferrari adds that scale doesn't multiply the effort. The Evernote transformation in 2023 used a task force of about 50 core-team members. A more recent acquisition roughly four times as large by revenue, with over 1,000 employees, took 50 to 60 people. That's why the company, now at about $3 billion in run-rate revenue, prefers fewer and bigger deals.
The strategy rests on a syllogism. If Bending Spoons can run a business better than almost anyone, then it can bid more than almost anyone and still earn strong returns. Ferrari says 99% of resources go to becoming the best operator, with remarkably little on the acquisition side itself. He believes the Evernote offer was far above competing bids (he adds that in hindsight they should have negotiated harder) and says that in the last five years, to his recollection, no seller has accepted another buyer's offer after rejecting theirs. That's a self-report, and he acknowledges he can't see the other side's books.
His negotiating style follows from that confidence. Put one fair number on the table, rarely raise it by more than 5 to 10%, and invite sellers to shop it around, since a seller who confirms the market for themselves closes faster. The exception he recalls was a seller who forced a number before data was available, so the company gave a range and later raised the offer substantially. The host connects this to Buffett's famous Clayton Homes bid, where $12.50 stayed $12.50 through every counter.
Logic is reliable, numbers need suspicion
Ferrari is a self-described stats geek (he follows jiu-jitsu and CrossFit through data and couldn't recognize the athletes), but he distrusts running a company on numbers alone. "Logic and rationality properly defined are perfect," he says. Numbers are approximations of reality.
His salary example makes the point. When the team proposed testing whether higher advertised pay would lift application quality, Ferrari backed the experiment but told colleagues in advance that he'd favor raising salaries even if the test showed nothing. A two-month test only captures click-through at the last step. It can't capture the compounding effect of a reputation for paying exceptionally well, which spreads through universities, workplaces and job boards over years. The test showed modest uplift, not enough to justify 20% higher pay on its own. Compensation went up anyway. Ferrari admits he can't prove the decision right, only that he believes it was.
The same temperament shows in his approach to timing. "Procrastination is awesome if it doesn't come from laziness," he says, because postponing a consequential decision usually buys information at almost no cost. The company decided in the first half of 2025 to prepare for a listing, but held the final go decision until the spring of 2026, when it chose to proceed.
Ferrari's argument for all of it is a single bet: when you hire people who care intensely, remove everything that doesn't pay for itself and give them tools that compound, the organization gets smaller and faster at once. Evernote's 20 people are the evidence he points to. Whether that holds across 50-plus businesses and a public-market audience is the test he's now taking.
