Luca Ferrari
Co-founder of Bending Spoons
Where to hear and read about Luca Ferrari1 episode or test
Luca Ferrari is a co-founder of Bending Spoons, which buys software businesses such as Evernote and Vimeo to hold forever and rebuilds them on one shared platform run by a small, rotating core team. The lesson most worth taking from him is that being the best operator is a strategy in itself: hire raw talent with extreme ownership, strip out complexity, and the ability to outbid everyone follows.
If you emulate what most people are doing, you're pretty much guaranteed to, at best, be mediocre.
Inside Bending Spoons: Finding Talent, Leveraging AI & Driving Operational Excellence | Luca Ferrari
Luca Ferrari: the story and the lessons
1How Luca Ferrari started Bending Spoons
Ferrari's first startup, Evertale, was an AI diary app that failed. Evertale ran from 2010 to 2013 and wrote diaries automatically, but it never scaled.
Its lasting lesson came from the team: the best of about 12 people was roughly 10 times as productive as the median and was one of the least experienced. Ferrari's later habit of using cutting-edge technology also traces back to this AI work from 2010.
He funded Evertale by working nights while employed at McKinsey. Ferrari took McKinsey's offer, told the partner he planned to leave, and was hired anyway. He worked on the startup from midnight to 2 a.m., raised about $500,000 a year later and quit.
Isolation in Copenhagen and Milan forced the founders to think from first principles. Bending Spoons started in Copenhagen and soon moved to Milan, with no ecosystem, no advisers and founders nobody knew.
They accepted more mistakes and slower progress in exchange for a few genuinely new ways of working, and being far from startup hubs kept them away from the mantras those hubs preached.
Ferrari goes all in or does the bare minimum, nothing in between. He argues a 10 out of 10 marriage is worth 100 times an adequate one, and he treats work the same way, keeping one or very few pursuits and cutting the rest.
Aiming at the greatest cathedral rather than a nice church is, in his view, more energizing and attracts better people, even if the goal is close to impossible.
He tries to erase the idea of a founder inside the company. Only contribution and trajectory count, whether someone started Bending Spoons or joined later.
He calls the shared ambition of his closest colleagues a Bending Spoons thing rather than a founder thing. He openly admires fanatically focused founders like Adam Foroughi of AppLovin, whom he rates 10 out of 10 on rationality and effectiveness.
2Why Luca Ferrari hires talent over experience
Ferrari believes experience goes stale fast in technology. How software is built in 2026 bears little resemblance to 2010, and most work is not rocket science: it needs a good brain and a desire to do well.
Experience can even hurt when it normalizes low standards or office politics. Experience can be given to talent later through good challenges and strong colleagues, but talent cannot be given to experience.
He treats hiring as a statistics problem built on weak signals. A talent team scores over 100 signals from academic records, personal projects, recruiter emails and tasks, each marginally predictive but strong together, the way a quant fund works.
Results mislead for young candidates because a year or two of output is too small a sample. One signal: how a candidate treats logistics support staff predicted collaboration better than interview behavior. "Anything really anything qualifies as a signal and we just care that it's predictive."
Ferrari took hiring away from hiring managers. A central talent team that does nothing else handles hiring and exits. Last year it received 800,000 applications and hired fewer than 300 people.
Managers want seats filled fast and prefer mature candidates over green ones who could become far better. The central team sees every role, knows when waiting will likely pay off, and can redirect a candidate, for example from product manager to growth manager.
Extreme ownership is the trait he selects for above all. Past a fairly high intelligence bar, Ferrari says success is almost all about how badly someone wants to be amazing. Some near-genius people failed because the job was only their third or fourth priority.
He looks for proof such as heavy open-source work, crafts learned alongside studies, or a startup someone ground at for three years. He bets there is less than a 1% chance an extreme owner fails to build an amazing career there, citing Rafa Nadal, who was not a top-50 talent but became one of the greatest.
He keeps the team dense with owners because ownership is contagious. Owners are paranoid about what could go wrong, forget nothing and do more than asked, which sets off a positive escalation among peers.
Dilute the group and those people leave or lose the drive, so Ferrari says a high-performance team cannot have more than a small fraction who lack it.
3How Luca Ferrari organizes and pays people
A strong team needs as few rules as possible. With mediocre talent, process and checks are the lesser evil. With strong talent, Ferrari gives leeway and trust, comparing it to coaching a Jordan and Barkley dream team.
Netflix's culture deck influenced him most: beat complexity with talent, not process.
Ferrari pays no bonuses, not even to hiring managers. Bending Spoons has no variable pay because tying concrete targets to pay or promotion pushes people to tick boxes.
"We find that setting highly specific concrete objectives to which career progression or pay are tied almost invariably leads to bad outcomes."
Everyone at Bending Spoons has the same job: help the company win. Ferrari dislikes sharp job descriptions, so the idea of a seat barely exists and empty seats are not a worry.
Of 100 possible projects perhaps 10 are ROI positive, and many companies do 40 things while skipping good ones. Winning means doing only the ROI-positive work in priority order.
He deliberately gives promising people more work than they can finish. Adding an 11th task to ten can never lower the return on what someone picks, and it may turn out to be the best one.
It trains people to handle the immensity of what is possible, which Ferrari would rather test in someone's first year than discover at year ten running a billion-dollar organization. A manager's top job becomes helping reports choose the right tasks.
People rotate across businesses often. Fresh eyes bring new ideas, rotation prevents subcultures, varied challenges keep people learning, and resources follow each new acquisition's higher returns.
Pay must be high enough that nobody thinks about it. A two-month job-ad test showed only modest uplifts, not enough alone to justify paying about 20% more, but the company raised pay anyway.
Ferrari reasoned that a reputation as a top payer builds over months or years at universities and on pay sites, which a short test cannot capture. He admits he cannot prove the raise was right.
4Luca Ferrari's rule of relentless simplification
Anyone adding complexity carries the burden of proof. Going from three to four pieces can make a system 40 or 50% more complicated, not 33%, because of new connections. Whoever opposes an addition only has to raise a flag.
Ferrari blames this drift for modern bureaucracy: each rule was meant well, but nobody counted the cost of its connections. "The burden of proof is on those who want to add complexity."
The longer something has existed, the harder he questions it. Inertia bias makes people assume the status quo is fine, and Ferrari says knowing that weakness is about half of beating it.
He adds that simplicity buys speed and scalability.
Ferrari scrapped job titles entirely rather than refining them. About a decade ago, defining who counted as senior or director was draining and disappointing people. The only benefits turned out to be bragging rights and CV signaling.
People may now pick any reasonable title for LinkedIn, and an algorithm labels roles internally: anyone managing product managers is a product management lead, with two reports or 200. No one has complained, and hundreds to thousands of hours were saved.
Full removal beats trimming. Incremental simplification might give a 5% lift, while dropping titles was closer to a 10x improvement.
The same applies to products: cut a feature only 2% of users touch and the other 98% are served far better within a year. He advises doing it outright, not through a slow migration.
5The Bending Spoons operating system and its use of AI
Over 50 proprietary tools run every business Bending Spoons owns. They cover payments, A/B tests, lifetime value, recruiting, AI orchestration, access control and data. Serious investment began about ten years ago, when the company was five people.
It works like an internal open-source community: businesses fix bugs and add features, and the improvements spread across the portfolio. "Adding businesses actually makes us better as a whole."
Ferrari keeps the tools private. They are a competitive advantage, too advanced for most digital businesses, and so tightly integrated that sold singly much of their value fades.
Selling them would also cost focus.
AI tools remove slow handoffs between people. Diagram, an in-house design tool, follows the head designer's guidelines, reads the codebase, writes code and creates A/B test segments, so product managers and engineers can do design work.
"We humans are insanely inefficient at exchanging information." A request to a designer can cost three days and still come back wrong. The team has used AI since 2018, first to predict user lifetime for A/B tests.
An agent called Old Spooner does multi-person work in one request. It lives in Slack with the same access as the person asking. Evernote's general manager used it to find a bug's root cause, propose a fix and alert the lead engineer in about three minutes.
Ferrari got monthly-active-user graphs in minutes instead of waiting a day or two for an analyst.
Headcount is a result, not a target. The company earns well over $4 million in revenue per core team employee and would hire more if it could find enough people.
Ferrari points to Instagram being built by about a dozen people, and warns that large organizations pile on teams and rules until caring and getting things done both get hard.
6How Luca Ferrari buys and rebuilds companies like Evernote
Ferrari makes operational excellence the whole strategy. About 99% of resources go to being the best operator and remarkably little to the acquisition side.
"By definition, mathematically, if a business is better off with you than with everybody else, you should be the highest bidder." He does not claim to spot what others miss, only that businesses do better under his ownership.
He bought Evernote for its predictability. About a quarter of a billion people had used it, paying customers had stayed 5 to 10 years, and most revenue was subscriptions. It earned about $90 million and roughly broke even, and Ferrari believes the roughly $200 million offer was far above the next bid.
He prefers betting on existing customers over new-user acquisition, which swings with competition and ad costs. In hindsight he says the team should have negotiated the price better.
About 20 people now run Evernote, down from about 350. A task force of about 50 rebuilt the codebase, infrastructure and monetization, and product improvements now come at least three times as fast.
Being part of a group lets businesses take bets standalone owners cannot. A standalone company is punished for falling subscriber counts, so it may skip a price rise even when revenue would grow. Ferrari's example: 30% fewer customers each paying twice as much is a better business.
A failed bet is not fatal inside the group, lessons carry across businesses, and acquired products attract top graduates the way McKinsey does, offering 12 to 18 months rebuilding AOL or Evernote.
Ferrari prefers fewer, bigger acquisitions. Vimeo, at about $400 million in revenue and over 1,000 people, took a task force of 50 to 60, roughly what Evernote needed.
Transformation effort does not grow with revenue and the platform keeps improving. At about $3 billion in run-rate revenue, a deal must be much larger than Evernote to matter, though more businesses still strengthen the tools.
He insists Bending Spoons is not private equity. It has never sold a material business, rebuilds product and technology, and integrates every business into one platform and core team.
Ferrari has never seen private equity reinvent a product, and integration rules out the piecemeal sale a fund needs. "We don't buy to sell... We buy to hold and operate forever."
7How Luca Ferrari negotiates and decides
Ferrari names one fair, firm number and does not haggle. Raises have almost never exceeded 5 or 10%, and Bending Spoons has not been outbid in at least five years.
His model is Buffett's Clayton Homes bid, answering 17, 15 and 14 with the same $12.50. In one recent case a seller forced a number before sharing data, so he gave a range and later raised the offer substantially.
He encourages sellers to shop his offer. A seller sure it has the best offer signs faster, and some who walk away return 6 to 9 months later at the same price.
Weak bidders, by contrast, push for exclusivity within five days because they must win on timing.
Ferrari delays big decisions and makes small ones at once. Bending Spoons chose in early 2025 to prepare for an IPO but left the final call as late as possible, deciding definitively in spring 2026.
Waiting often brings more information or makes a decision moot, while small decisions are made immediately to free up mental RAM. "Procrastination is awesome if it doesn't come from laziness."
He trusts logic over numbers. A self-described stats geek, Ferrari still treats numbers as approximations that mislead when taken at face value.
"Logic and rationality, they never fail you."
Henry Singleton is his model for capital allocation. Acquisitions remain the best use of capital, so buybacks are a far-off option.
Ferrari admires how Singleton paid with richly valued stock for undervalued businesses and later spun them off without dogma. He sees Buffett and Singleton as great pickers but not exceptional operators, since they avoided businesses needing heavy fixing.
Inside Bending Spoons: Finding Talent, Leveraging AI & Driving Operational Excellence | Luca Ferrari
Bending Spoons buys software businesses to hold forever and rebuilds them on a shared platform of proprietary tools, run by a small core team picked for talent and drive over experience. The result is businesses like Evernote run by about 20 people, plus acquisition bids that have not lost to a rival in five years.
Ferrari's first startup, Evertale, was an AI diary app that failed. Evertale ran from 2010 to 2013 and wrote diaries automatically, but it never scaled.
Its lasting lesson came from the team: the best of about 12 people was roughly 10 times as productive as the median and was one of the least experienced. Ferrari's later habit of using cutting-edge technology also traces back to this AI work from 2010.
He funded Evertale by working nights while employed at McKinsey. Ferrari took McKinsey's offer, told the partner he planned to leave, and was hired anyway. He worked on the startup from midnight to 2 a.m., raised about $500,000 a year later and quit.
Isolation in Copenhagen and Milan forced the founders to think from first principles. Bending Spoons started in Copenhagen and soon moved to Milan, with no ecosystem, no advisers and founders nobody knew.
They accepted more mistakes and slower progress in exchange for a few genuinely new ways of working, and being far from startup hubs kept them away from the mantras those hubs preached.
Ferrari goes all in or does the bare minimum, nothing in between. He argues a 10 out of 10 marriage is worth 100 times an adequate one, and he treats work the same way, keeping one or very few pursuits and cutting the rest.
Aiming at the greatest cathedral rather than a nice church is, in his view, more energizing and attracts better people, even if the goal is close to impossible.
He tries to erase the idea of a founder inside the company. Only contribution and trajectory count, whether someone started Bending Spoons or joined later.
He calls the shared ambition of his closest colleagues a Bending Spoons thing rather than a founder thing. He openly admires fanatically focused founders like Adam Foroughi of AppLovin, whom he rates 10 out of 10 on rationality and effectiveness.
Ferrari believes experience goes stale fast in technology. How software is built in 2026 bears little resemblance to 2010, and most work is not rocket science: it needs a good brain and a desire to do well.
Experience can even hurt when it normalizes low standards or office politics. Experience can be given to talent later through good challenges and strong colleagues, but talent cannot be given to experience.
He treats hiring as a statistics problem built on weak signals. A talent team scores over 100 signals from academic records, personal projects, recruiter emails and tasks, each marginally predictive but strong together, the way a quant fund works.
Results mislead for young candidates because a year or two of output is too small a sample. One signal: how a candidate treats logistics support staff predicted collaboration better than interview behavior. "Anything really anything qualifies as a signal and we just care that it's predictive."
Ferrari took hiring away from hiring managers. A central talent team that does nothing else handles hiring and exits. Last year it received 800,000 applications and hired fewer than 300 people.
Managers want seats filled fast and prefer mature candidates over green ones who could become far better. The central team sees every role, knows when waiting will likely pay off, and can redirect a candidate, for example from product manager to growth manager.
Extreme ownership is the trait he selects for above all. Past a fairly high intelligence bar, Ferrari says success is almost all about how badly someone wants to be amazing. Some near-genius people failed because the job was only their third or fourth priority.
He looks for proof such as heavy open-source work, crafts learned alongside studies, or a startup someone ground at for three years. He bets there is less than a 1% chance an extreme owner fails to build an amazing career there, citing Rafa Nadal, who was not a top-50 talent but became one of the greatest.
He keeps the team dense with owners because ownership is contagious. Owners are paranoid about what could go wrong, forget nothing and do more than asked, which sets off a positive escalation among peers.
Dilute the group and those people leave or lose the drive, so Ferrari says a high-performance team cannot have more than a small fraction who lack it.
A strong team needs as few rules as possible. With mediocre talent, process and checks are the lesser evil. With strong talent, Ferrari gives leeway and trust, comparing it to coaching a Jordan and Barkley dream team.
Netflix's culture deck influenced him most: beat complexity with talent, not process.
Ferrari pays no bonuses, not even to hiring managers. Bending Spoons has no variable pay because tying concrete targets to pay or promotion pushes people to tick boxes.
"We find that setting highly specific concrete objectives to which career progression or pay are tied almost invariably leads to bad outcomes."
Everyone at Bending Spoons has the same job: help the company win. Ferrari dislikes sharp job descriptions, so the idea of a seat barely exists and empty seats are not a worry.
Of 100 possible projects perhaps 10 are ROI positive, and many companies do 40 things while skipping good ones. Winning means doing only the ROI-positive work in priority order.
He deliberately gives promising people more work than they can finish. Adding an 11th task to ten can never lower the return on what someone picks, and it may turn out to be the best one.
It trains people to handle the immensity of what is possible, which Ferrari would rather test in someone's first year than discover at year ten running a billion-dollar organization. A manager's top job becomes helping reports choose the right tasks.
People rotate across businesses often. Fresh eyes bring new ideas, rotation prevents subcultures, varied challenges keep people learning, and resources follow each new acquisition's higher returns.
Pay must be high enough that nobody thinks about it. A two-month job-ad test showed only modest uplifts, not enough alone to justify paying about 20% more, but the company raised pay anyway.
Ferrari reasoned that a reputation as a top payer builds over months or years at universities and on pay sites, which a short test cannot capture. He admits he cannot prove the raise was right.
Anyone adding complexity carries the burden of proof. Going from three to four pieces can make a system 40 or 50% more complicated, not 33%, because of new connections. Whoever opposes an addition only has to raise a flag.
Ferrari blames this drift for modern bureaucracy: each rule was meant well, but nobody counted the cost of its connections. "The burden of proof is on those who want to add complexity."
The longer something has existed, the harder he questions it. Inertia bias makes people assume the status quo is fine, and Ferrari says knowing that weakness is about half of beating it.
He adds that simplicity buys speed and scalability.
Ferrari scrapped job titles entirely rather than refining them. About a decade ago, defining who counted as senior or director was draining and disappointing people. The only benefits turned out to be bragging rights and CV signaling.
People may now pick any reasonable title for LinkedIn, and an algorithm labels roles internally: anyone managing product managers is a product management lead, with two reports or 200. No one has complained, and hundreds to thousands of hours were saved.
Full removal beats trimming. Incremental simplification might give a 5% lift, while dropping titles was closer to a 10x improvement.
The same applies to products: cut a feature only 2% of users touch and the other 98% are served far better within a year. He advises doing it outright, not through a slow migration.
Over 50 proprietary tools run every business Bending Spoons owns. They cover payments, A/B tests, lifetime value, recruiting, AI orchestration, access control and data. Serious investment began about ten years ago, when the company was five people.
It works like an internal open-source community: businesses fix bugs and add features, and the improvements spread across the portfolio. "Adding businesses actually makes us better as a whole."
Ferrari keeps the tools private. They are a competitive advantage, too advanced for most digital businesses, and so tightly integrated that sold singly much of their value fades.
Selling them would also cost focus.
AI tools remove slow handoffs between people. Diagram, an in-house design tool, follows the head designer's guidelines, reads the codebase, writes code and creates A/B test segments, so product managers and engineers can do design work.
"We humans are insanely inefficient at exchanging information." A request to a designer can cost three days and still come back wrong. The team has used AI since 2018, first to predict user lifetime for A/B tests.
An agent called Old Spooner does multi-person work in one request. It lives in Slack with the same access as the person asking. Evernote's general manager used it to find a bug's root cause, propose a fix and alert the lead engineer in about three minutes.
Ferrari got monthly-active-user graphs in minutes instead of waiting a day or two for an analyst.
Headcount is a result, not a target. The company earns well over $4 million in revenue per core team employee and would hire more if it could find enough people.
Ferrari points to Instagram being built by about a dozen people, and warns that large organizations pile on teams and rules until caring and getting things done both get hard.
Ferrari makes operational excellence the whole strategy. About 99% of resources go to being the best operator and remarkably little to the acquisition side.
"By definition, mathematically, if a business is better off with you than with everybody else, you should be the highest bidder." He does not claim to spot what others miss, only that businesses do better under his ownership.
He bought Evernote for its predictability. About a quarter of a billion people had used it, paying customers had stayed 5 to 10 years, and most revenue was subscriptions. It earned about $90 million and roughly broke even, and Ferrari believes the roughly $200 million offer was far above the next bid.
He prefers betting on existing customers over new-user acquisition, which swings with competition and ad costs. In hindsight he says the team should have negotiated the price better.
About 20 people now run Evernote, down from about 350. A task force of about 50 rebuilt the codebase, infrastructure and monetization, and product improvements now come at least three times as fast.
Being part of a group lets businesses take bets standalone owners cannot. A standalone company is punished for falling subscriber counts, so it may skip a price rise even when revenue would grow. Ferrari's example: 30% fewer customers each paying twice as much is a better business.
A failed bet is not fatal inside the group, lessons carry across businesses, and acquired products attract top graduates the way McKinsey does, offering 12 to 18 months rebuilding AOL or Evernote.
Ferrari prefers fewer, bigger acquisitions. Vimeo, at about $400 million in revenue and over 1,000 people, took a task force of 50 to 60, roughly what Evernote needed.
Transformation effort does not grow with revenue and the platform keeps improving. At about $3 billion in run-rate revenue, a deal must be much larger than Evernote to matter, though more businesses still strengthen the tools.
He insists Bending Spoons is not private equity. It has never sold a material business, rebuilds product and technology, and integrates every business into one platform and core team.
Ferrari has never seen private equity reinvent a product, and integration rules out the piecemeal sale a fund needs. "We don't buy to sell... We buy to hold and operate forever."
Ferrari names one fair, firm number and does not haggle. Raises have almost never exceeded 5 or 10%, and Bending Spoons has not been outbid in at least five years.
His model is Buffett's Clayton Homes bid, answering 17, 15 and 14 with the same $12.50. In one recent case a seller forced a number before sharing data, so he gave a range and later raised the offer substantially.
He encourages sellers to shop his offer. A seller sure it has the best offer signs faster, and some who walk away return 6 to 9 months later at the same price.
Weak bidders, by contrast, push for exclusivity within five days because they must win on timing.
Ferrari delays big decisions and makes small ones at once. Bending Spoons chose in early 2025 to prepare for an IPO but left the final call as late as possible, deciding definitively in spring 2026.
Waiting often brings more information or makes a decision moot, while small decisions are made immediately to free up mental RAM. "Procrastination is awesome if it doesn't come from laziness."
He trusts logic over numbers. A self-described stats geek, Ferrari still treats numbers as approximations that mislead when taken at face value.
"Logic and rationality, they never fail you."
Henry Singleton is his model for capital allocation. Acquisitions remain the best use of capital, so buybacks are a far-off option.
Ferrari admires how Singleton paid with richly valued stock for undervalued businesses and later spun them off without dogma. He sees Buffett and Singleton as great pickers but not exceptional operators, since they avoided businesses needing heavy fixing.