How an Amazon Engineer Earning $500k a Year Quit and Made $310,000 From 16 Hours of Work

Most people think the trap is the job.

It is not. The trap is the story you tell yourself about why you cannot leave it.

Daniel Vassallo had the $511,000 salary. The promotions. The respect of every engineer in the room. The clear path to principal engineer and then beyond that. He had everything the system promised you if you worked hard enough and waited long enough.

He left anyway. Not for another job. Not for a funded startup. Not for anything with a guaranteed outcome.

He left because he stopped believing the next year was going to feel different from the last one.

What happened after he left is one of the cleanest demonstrations of what happens when someone with real expertise stops trading that expertise for a salary and starts building small things from it instead.

A book he co-wrote in a few weeks made $140,000. A 100-minute video course he put together in 16 hours made $310,000. A Discord community he built as an afterthought generated over $800,000 in its first two years. In April 2025 he sold that community to Gumroad for $3.6 million.

None of it was a startup. None of it had venture capital. None of it required a team.

Every single thing he built started as a small bet. And the small bets kept compounding.

The Setup

Daniel Vassallo grew up in Malta.

Small island in the Mediterranean. Population around 500,000. The kind of place where your options feel defined by geography before you are old enough to question them.

He got hooked on computers at five years old. Started programming early. Spent his childhood on hardware and magazines and whatever he could afford to keep learning. Dropped out of university in Malta partway through a physics and computer science degree and shifted to a correspondence program through the University of London instead. Took whatever work he could find between studying. Waiting tables. A school lab. Sales for laboratory equipment. QA for a winery.

His first real programming job was part-time at a sports betting company in Malta. Around 2005. He walked in with a fake-it-till-you-make-it approach and found the problem immediately. Their developer had let them down and the team was manually inputting massive amounts of data every day. He automated it. Saved them hours of work before anyone expected him to.

That moment told him something important. The people who make themselves genuinely useful do not have to wait for permission to matter.

In 2010 he joined Amazon. Entry level developer. Building tools for AWS, Amazon's cloud infrastructure division. Within three and a half years he had been promoted twice to senior engineer. The track to principal engineer was essentially guaranteed if he stayed.

He stayed for eight years.

His salary started at $75,000. By his final year it was $511,000. Most of that in Amazon stock, because Amazon caps its cash compensation at $160,000 and pays everything above that in RSUs that vest and land in your account like cash.

$511,000. A family in Seattle. Two small kids, two and four years old. By any reasonable external measure, the story was going well.

The Constraint

Here is the thing about a $511,000 salary at Amazon.

It is not just money. It is a set of decisions that have already been made for you. Where you live. How you spend your days. What problems you work on. Which promotions you pursue. Which meetings you sit in. Which bureaucracies you navigate.

Vassallo was good at all of it. That was part of the problem.

He had been promoted. Respected. Listened to. He was regarded as an expert and a leader in his field. People looked up to him. And none of that was making him want to show up anymore.

He looked at the people above him. The managers. The senior executives. The people on the path he was supposed to want. He did not admire their lives. He did not want to become the version of himself that arrived there.

"I wasn't motivated enough to do another year. Despite getting rewarded repeatedly with promotions, compensation, recognition, and praise."

The motivation problem is not something a promotion fixes. It is something a promotion reveals. Every new level brings more reward and the same emptiness, which means the problem is not the level. The problem is the system.

He had saved approximately five and a half years of living expenses before he left. He is honest about that. The savings made the decision easier. But he has said plainly that he believes he would have left even without them. That if the savings had not existed he would have found a part-time remote arrangement to cover bills while he figured out what came next.

He left Amazon in February 2019.

He left almost a million dollars in unvested stock options on the table when he walked out the door.

The Opportunity

Vassallo did not leave Amazon with a plan.

He left with a direction. And there is a real difference between those two things.

The direction was this. Stay self-employed. Build things from what he already knew. Do not bet everything on one outcome. Keep the downside small enough that a failure does not end the experiment.

His first attempt was Userbase. A SaaS product for developers. End-to-end encrypted user authentication that made it simple for developers to build secure web apps without managing the complexity themselves. Technically interesting. Real problem. He got 1,000 people on the waitlist. Hit number one on Product Hunt on launch day. Made the front page of Hacker News.

And then almost nothing.

The product made around $10,000 a year. Not per month. Per year. For something that launched at the top of every developer platform simultaneously.

A different kind of person doubles down here. Raises money. Hires a team. Runs the playbook. Pushes through because the alternative is admitting that the first bet did not work.

Vassallo did something different. He accepted the evidence. He kept Userbase running as a small side thing and did not let it become the thing that consumed him.

While building Userbase he had been sharing everything on Twitter. The process. The decisions. The numbers. The problems he was hitting and how he was trying to solve them. He was not doing this as a marketing strategy. He was doing it because transparency was the natural way he processed what he was building.

His Twitter following grew. The audience he was building was not fans. It was people who recognised themselves in what he was sharing. Employed engineers. Developers with expertise and a job they were not sure about anymore.

That audience was not a business yet. But it was going to be.

The Playbook

He packaged what he already knew and sold it before he had time to overthink it.

By mid-2019 Vassallo had been building on AWS for over eight years. He knew which services actually mattered and which ones were noise. He knew how to make technical decisions without getting paralysed by optionality. He knew things that took a decade inside Amazon to learn and that most developers outside it were guessing about.

He and his former Amazon colleague Josh Pschorr turned that knowledge into a book. They called it The Good Parts of AWS. 173 pages. A deliberately opinionated take on which AWS services you would be foolish not to use and which ones you could safely ignore. Not a comprehensive reference. A specific, experienced perspective.

They launched it on Gumroad on December 25, 2019, priced at $65.

In the first 14 days it made over $45,000.

Total lifetime revenue crossed $140,000. Over 13,000 copies sold. For roughly 160 hours of work between two people.

The Christmas Day launch was not accidental. Developers are off work. They are online. They have time to read. They are thinking about what they want to do differently next year. The timing was a small deliberate decision that most people would not make because it feels wrong to launch something on a holiday.

Most right decisions feel slightly wrong before you make them.

He built his Twitter audience by sharing the work instead of talking about the work.

Here is what Vassallo did not do on Twitter. He did not post motivational content about entrepreneurship. He did not write threads about how to be successful. He did not build a persona around having escaped corporate life.

He shared what he was actually doing. Revenue numbers. Product decisions. Things that worked and things that failed. The real texture of trying to stay self-employed when you have two small children and you just walked away from $511,000 a year.

That specificity is what built the audience. Not volume. Not consistency for its own sake. Specific, honest, observable detail about a life people wanted to understand.

By the time he launched his Twitter course in April 2020 he had over 24,000 followers. Not tens of millions. Not even hundreds of thousands. But a very specific group of people who had been watching him figure this out in real time and trusted exactly what he would teach.

He made $310,000 from a 100-minute video he recorded in one take.

In April 2020 Vassallo released Everyone Can Build a Twitter Audience on Gumroad. Priced at $25.

He put it together in 16 hours. One take. No editing. No production studio. A screen recording of him talking through exactly what he had done to build his audience from zero to 24,000 followers.

It made $100,000 in the first two months.

Total revenue crossed $310,000. Over 13,000 sales.

Sit with the arithmetic for a moment. $310,000 from 16 hours of work. That is more money per hour than almost any knowledge worker on the planet earns from a salary, including the $511,000 Amazon salary he left behind.

The product did not succeed because it was polished. It succeeded because the person selling it had built the exact thing they were teaching in real time, transparently, in front of the people buying it.

The credibility was already established before the product existed. The product was just the packaging.

He failed at SaaS and accepted it fast enough that it did not cost him everything.

The Userbase story is the one most people skip because it is the uncomfortable part of the Vassallo narrative.

He built a technically sophisticated product. It launched at the top of every relevant platform simultaneously. And it did not work. Not in the way a business needs to work to justify the hours and the focus.

He did not hide this. He shared the numbers publicly. $10,000 a year. Not enough. Moving on.

That transparency did three things. It kept his audience's trust. It freed him to move resources to the things that were actually working. And it became the most concrete possible proof of his core thesis. That the portfolio approach protects you from any single failure. That a SaaS that does not work is just one bet that did not pay off. It does not end the experiment.

He turned a course into a community by accident and it became his biggest bet yet.

In November 2021 Vassallo ran a cohort course. A Portfolio of Small Bets. Six live Zoom sessions over two weeks. Around 30 people per cohort. Teaching exactly what he had learned about building small things from expertise instead of swinging for big outcomes.

He ran 29 cohorts.

As demand kept coming he realised the value was not in the six sessions. It was in what happened between them. The Discord. The community. The members who stayed and kept sharing and building together.

He packaged it differently. Join the community. The course is included. Lifetime access, one-time fee.

Between November 2021 and October 2023 the Small Bets community generated $824,000 in revenue. $600,000 in profit. Over 4,500 members. Growing by word of mouth from people who had built real things and attributed the thinking to what they learned inside the community.

In April 2025 he sold Small Bets to Gumroad for $3.6 million. Cash and options. He kept running it.

The Amazon job paid $511,000 a year and required 8 years to reach.

The Small Bets community made $3.6 million and required approximately four years to build from a cohort course he started as one more small bet.

Tools Used

Gumroad for distribution of every product. No custom checkout. No complex e-commerce setup. The product was the expertise. The platform was chosen for its simplicity.

Twitter as the primary audience-building channel. Not for virality. For transparency. Daily sharing of real decisions and real numbers over years, not weeks.

Discord as the community infrastructure once the cohort course evolved into something people wanted to stay inside of beyond the six sessions.

Zoom for the live cohort sessions. Not a sophisticated platform. The simplest tool that let him teach in real time to a small room.

A word processor and PDF export for the AWS book. No design agency. No publisher. A document saved as a PDF and put on Gumroad for $65.

Screen recording software for the Twitter course. One take. 100 minutes. No editing.

The pattern across every tool Vassallo chose is the same. The minimum amount of infrastructure that lets the expertise reach the buyer without adding complexity he had to manage.

Timeline

2005: First programming job at a sports betting company in Malta. Automates manual data entry. Realises that genuine usefulness does not require permission.

2010: Joins Amazon as an entry-level developer. Builds tools for AWS.

2013: Promoted twice to senior engineer within three and a half years. Compensation growing consistently.

2019, February: Leaves Amazon after eight years. Salary had reached $511,000. Walks away from approximately one million dollars in unvested stock options. Begins building Userbase.

2019, December 25: Launches The Good Parts of AWS on Gumroad with Josh Pschorr. Priced at $65. Over $45,000 in the first 14 days.

2020, April: Launches Everyone Can Build a Twitter Audience on Gumroad. Priced at $25. 16 hours to create. $100,000 in the first two months. $310,000 total.

2021: Userbase acknowledged as a failed bet. Making roughly $10,000 per year. Vassallo pivots focus. Begins freelancing. $220,000 in income that year.

2021, November: Launches the first cohort of A Portfolio of Small Bets. 29 cohorts total.

2022: Small Bets community transitions from cohort course to ongoing community. $720,000 in revenue that year.

2023, November: Small Bets community crosses $824,000 in total revenue since founding. $600,000 profit. Over 4,500 members.

2025, April: Sells Small Bets to Gumroad for $3.6 million. Continues operating it.

Mistakes and Lessons

The Userbase failure is worth examining beyond the surface level conclusion that SaaS is hard.

Vassallo has talked about what went wrong there clearly. The product was technically sophisticated. The launch metrics were strong. But the business model required acquisition, retention, and recurring revenue at scale. It required him to compete in a market where the incumbents had years and resources he did not.

He had picked a big bet disguised as a small one.

Userbase demanded the kind of commitment that SaaS demands. Long development cycles. Customer support infrastructure. Ongoing maintenance. It was not a product he could build in days and let run. It was a product that needed him fully and constantly and still produced $10,000 a year.

The lesson he took from it is the one that became the core of everything he built afterward.

A good bet has limited downside. It requires a small investment of time that you can actually afford to lose. It produces something that can earn money without requiring your constant active management. And it does not require you to be right about a specific large outcome. It just requires you to place enough small bets that one of them catches.

The AWS book took 160 hours across two people. The Twitter course took 16 hours. Both made six figures. The SaaS took years of full commitment. It made $10,000 a year.

He was not smarter when the bets worked. He was betting with a structure that let the wins compound without requiring the failures to be catastrophic.

The second mistake he identifies is waiting too long to believe the numbers.

When the AWS book made $45,000 in 14 days most people would have treated that as luck. He watched it carefully and then built the Twitter course using the same logic. Package specific expertise. Sell it to the audience that has been watching you build. Keep the production simple enough that the downside is 16 hours of your time. Let the market tell you if it was worth doing.

The market told him. He listened.

The Psychology

Three things drove Vassallo that most breakdowns of his story miss.

He optimised for optionality, not outcomes.

He did not leave Amazon to build the next billion-dollar company. He left to remain self-employed. That sounds like a smaller ambition. It is actually a more sophisticated one.

When you are optimising for staying in the game, every decision changes. You do not double down on a failing bet because doubling down might end the experiment. You build in small units so that no single failure can eliminate you. You choose structures that generate income while you sleep so that you are never one bad month away from having to go back.

The $3.6 million exit was not the goal. It was the output of hundreds of small decisions made while optimising for a completely different thing.

He used transparency as his compounding asset.

Every number he shared publicly was a deposit in an account that earned interest over time.

The person who watched Vassallo post about Userbase failing and then saw him launch the AWS book and make $45,000 in two weeks had context no advertisement could create. They had watched him try something real, fail at it, adapt, and try something else. When he put a product out they already trusted the man behind it.

That trust is not something you can manufacture. It is built slowly through honesty about the full picture, not just the wins.

Most people only share the good parts. Vassallo shared the spreadsheet that showed Userbase making $10,000 a year. That honesty is what made the Twitter course worth $310,000.

He understood that randomness is not the enemy. It is the game.

The Small Bets philosophy is not a productivity framework. It is a theory about how outcomes actually work.

Most people believe that if they pick the right idea and execute it with enough effort they will get the outcome they are aiming for. Vassallo decided early that this is not how it works. That there is too much randomness in the world to predict which specific bet will pay off. That the correct response to that randomness is not to make bigger, more confident bets. It is to make more, smaller, better-structured bets.

When the game is random, you win by being in the game longer with more positions. Not by betting everything on one outcome you have convinced yourself you can predict.

That insight is worth sitting with longer than it takes to read it.

The 2026 Builder Translation

Vassallo's story is from 2019. The tools he used are still available. The principles underneath them are even more powerful in 2026 than they were then.

Here is what the same playbook looks like if you started it today.

The expertise you have is more valuable than you think it is.

Vassallo had eight years of AWS experience. He packaged a fraction of it into a 173-page PDF and made $140,000.

You have expertise from your job right now. Domain knowledge that took years to accumulate and that people outside your industry are actively trying to find. The question is not whether you have something worth selling. It is whether you believe it is worth selling. Most people do not. That is why the market for specific, hard-won expertise is not saturated.

Pick one thing you know better than 95% of people who have not done your job. Write it down. That is the first version of your product.

The production bar is lower than it has ever been.

In 2019 Vassallo recorded a 100-minute screen recording in one take and made $310,000 from it.

In 2026 you have AI-assisted writing tools that compress the time from expert knowledge to finished document dramatically. You have short-form video platforms where a 60-second clip can build an audience faster than a year of blog posts. You have platforms that handle payment, delivery, and customer communication for a small percentage fee. The infrastructure that used to take months and a team to build is available to a solo builder in an afternoon.

The 16-hour course is now closer to a 6-hour course. The 160-hour book is now closer to an 80-hour book. The production constraint that once made this feel impossible has almost disappeared. What remains is the expertise. Which you already have.

Building in public on X is the highest-leverage thing you can do before you launch anything.

Vassallo did not have an audience when he left Amazon. He built one by sharing what he was actually doing. By the time he had a product, thousands of people were already invested in whether it would work.

In 2026 X has more people looking for exactly what he was sharing than it did in 2019. The feed moves faster. The competition for attention is higher. But a person sharing real numbers, real decisions, and honest failures in a specific domain still cuts through because that kind of content is still rare.

Start posting before you have a product. Document the expertise. Share the decision-making. Let the audience form around the process. Then the product launch is not a cold start. It is a delivery to people who have been waiting.

Replace the goal with a direction.

This is the one that sounds simple and is genuinely hard.

Vassallo did not leave Amazon to make $3.6 million. He left to stay self-employed. That direction gave him permission to fail at Userbase without it meaning he had failed at the larger project. It gave him permission to try a book and a course and a community course and a Discord without any single one of them having to be the answer.

In 2026 most builder content tells you to define your goal, reverse-engineer the steps, and execute the plan. That is how you build a career inside a company. It is not how the best small builders have built things from nothing.

Pick a direction. Stay self-employed. Keep the experiment running. The specific outcome emerges from the attempts, not the plan.

Modern Opportunity Radar

If this playbook is making you think about what to build, here are three real opportunity spaces that share the same DNA as Vassallo's small bets approach in 2026.

Expertise-as-product in overlooked professional domains.

Vassallo sold AWS knowledge to developers. The same model applies to any professional domain with complexity that takes years to understand from the inside. Logistics. Healthcare operations. Manufacturing supply chains. Regulatory compliance in any industry. The people inside these fields assume their knowledge is too boring or too specific to sell. That assumption is wrong. The more specific the expertise, the smaller the competition and the higher the willingness to pay.

One-take video knowledge products.

The production quality of a one-take video course is no longer a disadvantage. It is increasingly a signal of authenticity. A polished corporate training video feels like a company. A one-take recording of someone who actually did the thing feels like a mentor. The market for the latter is growing and the tools to produce and sell it have never been more accessible to a solo builder.

Community as the product, not the add-on.

Vassallo built his Discord community as an afterthought to a course and it became his most valuable asset. In 2026 the same pattern is available in every professional niche. The people who are building things, struggling with specific problems, and looking for others who understand their situation are on Discord and Slack and Circle right now. A well-curated community around a specific domain of expertise, with real events and real practitioners sharing real experience, is a product with higher retention than any course and higher perceived value than any book.

How You Can Replicate This

You have expertise from your job. Real expertise. The kind that took years to develop and that most people outside your field do not have access to.

Here is the sequence that works.

Write down the single most useful thing someone could know in your domain that they cannot learn from a Google search. The insight that took you years to understand. The mistake everyone in your field makes that you learned to avoid. The framework you use that nobody taught you because you built it from experience.

That is your first product. Not the full system. Not the comprehensive guide. The one thing.

Open a free Gumroad account. Write it down. Make it a PDF. Price it at $29 to $49. Share it on X to however many followers you have, including if that number is small. Let the people who buy it tell you what they want more of.

While you are doing that, post on X about the process. What you know. What you are building. What the numbers look like. Not the highlights. The full picture.

Do this for ninety days before you evaluate whether it is working. Ninety days of one thing at a time. No pivoting at week three because the numbers are small. The numbers are small at week three for everyone.

When something gets a signal, the next bet gets clearer. Vassallo's AWS book told him his audience would pay for specific expertise. The Twitter course told him they would pay for expertise about building the audience itself. Each small bet made the next one more accurate.

You do not need to know what the fifth bet is when you are making the first one. You just need to make the first one small enough that you can afford to find out.

Related Playbooks

The Elston playbook covers how a JP Morgan employee built a web hosting business by finding the specific audience the giants were ignoring. The distribution mechanics are different from Vassallo's but the structure is identical. Pick a specific group of people being failed by what currently exists. Build exactly what they need. Charge before you have finished building it.

The Minecraft playbook covers how Markus Persson used community as a distribution engine. Vassallo did the same thing with Twitter. Different platform, same principle. The people who watched him build in public became the people who bought what he built.

Every playbook in this series starts with the same question. Who has expertise that the market is not currently reaching. And what is the smallest possible way to package that expertise and find out if someone will pay for it.

Premium Insights

Here is the number most retellings of Vassallo's story get wrong.

The Twitter course did not make $310,000 because it had a sophisticated launch strategy. It made $310,000 because of what happened in the two years before anyone bought it.

Vassallo had been posting on Twitter about the Userbase experiment. About the AWS book. About every failed small bet and every successful one. By April 2020 when he posted the Twitter course for sale he had an audience of 24,000 people who had watched him try to stay self-employed in real time for over a year.

Those 24,000 people were not followers in the conventional sense. They were witnesses. They had watched him fail at Userbase and keep going. They had watched him make $45,000 in 14 days from a book. They had watched him share every spreadsheet and every decision with the kind of transparency that most people are too afraid to attempt.

When he released a product about how to build exactly the kind of audience they were already a part of, the purchase was almost a foregone conclusion.

The lesson is not about the product. It is about the two years before the product.

In 2026 you can start building that audience today. Not after you have something to sell. Not after you have figured out what the product is. Now. By sharing what you are actually doing and thinking and learning in whatever domain you understand from the inside.

By the time you have the product, the audience will be waiting. By the time you launch, the trust will already exist. The 16 hours of work will produce $310,000 not because 16 hours of work is worth $310,000. But because the two years of transparency before it were worth every minute.

The second insight is about the word self-employed.

Vassallo has said clearly that he did not leave Amazon to build a bigger thing than Amazon. He left to remain self-employed. That is a completely different optimisation target than the one most entrepreneurial content tells you to aim for.

When your target is self-employment rather than scale, a $140,000 book is a win. A $310,000 course is a win. A SaaS making $10,000 a year is a data point, not a catastrophe. The metric is staying in the game with your autonomy intact.

That reframe changes every decision you make about what to build, how big to let it get, and what success actually means when it arrives.

Most people aim at a number. Vassallo aimed at a life. The number turned out to be $3.6 million but that was never what the bet was about.

Your Move This Week

Open a blank document. Write down the single most valuable thing you know from your job that took you more than a year to fully understand. Not a topic. One specific insight. One framework. One mistake and the correct way to avoid it. Write it down in full. That document is the first version of your first product. Give it a price between $29 and $49. Do not edit it for more than two hours. Put it on Gumroad. Tell ten people. See what happens.

* * *

Vassallo did not leave Amazon because he was brave. He left because he was honest enough to admit that the next promotion was not going to fix the thing that the last one did not fix.

The salary is never the trap. The story you tell yourself about what you would lose by leaving is the trap.

He stopped telling that story in February 2019. What came after is the record.