Pinned Tweet
I'm planning to be more intentional on X in 2026, so here's who I am and what I do:
I'm Greg.
- Running 4 companies with ~40 people, 0 outside investment
- 1M+ transactions processed on Easytools, our creator stack
- Bootstrapped a Polish AI education company to $8.2M
- 20,000+ graduates from Samsung, Allegro, PwC, Orange
- 50%+ course completion rate (industry average: 5-15%)
Most of this happened in the last 3 years, and almost all of it came from ignoring the playbook everyone else copies.
I grew up in a Polish apartment block. Built websites in Flash at 14 for Quake clans. Learned to code by retyping examples out of physical books because YouTube didn't exist yet.
That whole arc, from retyping code in a Polish apartment as a kid to running 4 companies between Mallorca and Warsaw, is the lens I see everything through today.
Here's what I'll be posting about:
- Lessons from 200K customers across 18 yrs of selling online
- Running 4 companies async with a small team + automations
- What it actually takes to build a serious company without a single round of VC
- How giving course authors the majority of profit became our biggest moat
- Why education and tools belong in the same business, not separate ones
- The founder traps that quietly kill bootstrapped companies before they scale
Most founders I meet are quietly burning years building the wrong thing the wrong way. VC traps, cofounder traps and "best practices" that only work at scale you'll never reach.
My goal for 2026 is to give you the unedited version of what actually works.
Follow if you want the full playbook.
I'll link every thread I write under this post.
Let's go.
I wrote down how we're building MEGA: a program unlike any we've done, 2,000+ developers in already.
↳ challenges as tools inside your own harness
↳ education rebuilt agent-first, from scratch
↳ a certificate you can fail to earn
↳ a story nobody will explain
welcome to our kitchen 👨🍳
read here: brave.inc/post/how-we-build-…
#megadev
@_overment @joelhooks @megadevhq @theo @techgirl1908 @kentcdodds @johnlindquist
I built my radio engine (onair.aiatlas.work/#music)
are mp3s public somewhere? I want to have @levelsio playlist 🥁 🙏
It worked
Of course it worked!
It never ever doesn't work right?
I vibecoded my own Spotify that works with Amazon Alexa and plays all my own MP3 music files 😊
Claude Code calls it the 🎶 Jukebox
I put some coffee jazz MP3s (that I purchased) in an R2 bucket, made an Alexa skill in the Alexa Developer Console with help from Claude Code and now it plays it if I ask "Alexa, coffee jazz"
It wrote a package.json and index.js, then I saved and built it on the console. Then I had to manually add a Routine to Alexa so that "Alexa, coffee jazz" became "Alexa, ask the jukebox to play coffee jazz"
It also has Bossa Nova and Lofi and I can just drop music in whenever I want and it auto plays it
Spotify with Alexa was giving me really terrible recommendations of music and always the same and the only thing I used it for was this!
So I guess I can cancel Spotify now too 🙏
Greg retweeted
"the application of taste is essentially a differentiator"
🔥 new MEGA podcast with @mattpocockuk & @joelhooks: youtu.be/0BlAHq8D79U
Greg retweeted
"it is time-consuming, but it unlocks so many possibilities that it's simply worth it" - check out our drop by @_overment on moving 924 pages from Webflow to Astro, with 15 migration steps inside 🤫
first time my saas analytics show clarity
there's a huge leverage in building internal analytics tool. I could never get the right data in place when using third party apps.
now it just works, is free and unlocks so many other things like event driven, agent led onboarding
Theo wrote that he feels sick about charging $1,300.
Adam wrote that he locks himself away until the job is done.
Joel wrote he was f'n nervous to pitch.
I'm the one who set the price, asked Joel and Adam if they'll do it. So let me be real too.
I'm in education for 20+ years, my whole adult and professional life. Over the years we've trained over 400 000 people, helping build the culture of value, skills, and curiosity in Poland.
In many ways education feels like giving back to the community that shaped me. The thing I learned is that the price is not where the trust is won or lost. It's won in week two, when someone realizes it was worth it, or lost in week two, when they realize it wasn't.
So here's what the $1,300 pays for
Five amazing Authors each one working on their own projects, pushing the limits of practical ai in development. @kentcdodds @techgirl1908 @johnlindquist @theo @_overment were all brave enough to trust us and we'll always put their reputation above ours.
A team of 20+ that spent months building a program that runs inside your own harness, because we wanted you working, not watching.
A few things we built that I've never seen in a course before because pushing limits is fun.
Who should buy it: a developer whose company pays for training. A senior developer who already works with agents every day and wants the next level, paid from money that isn't rent. That's who we built this for.
Who shouldn't: the person Theo described. If you're early, scared of falling behind, and this would come from savings, don't. We are building a vast library of free resources: drops, skills, podcasts. The assessment is free. The Kickoff on Monday is free. Take those, stay with us we'll be shipping tons of free stuff, consider MEGA next year.
2,000+ developers are already in. It's the largest first edition we've ever run and the first global one, and I'm nervous in exactly the way Theo and Adam are: not about whether they'll come, but about whether we'll be worth it.
I think it will. Let's find out together.
And if you're not joining, wish us luck anyway ✌️
"I trust him with my life."
I'm the guy responsible for the concept, outline and core content of the upcoming cohort.
over the past three and a half years, my courses have reached 16k+ developers and thousands of digital creators and business owners in Poland. 4.5k developers joined the last edition alone.
after that, Greg and I sat down with Joel and decided to make something together. that's how the idea of MEGA was born.
at the same time, we knew we couldn't make it happen alone. or maybe we could, but we didn't want to. so we reached out to the best people we could think of and told them about MEGA.
now, you need to know that here in Poland, people know me. they know what to expect from my projects, and many of them have been with me since the very first one.
they trust me, and I don't want to let them down.
and yet, I never know if they find enough value in my work, or if they enjoy what they get. so every single time, I lock myself away and put my whole heart into it until the job is done.
but this time there's something more. I feel responsible not only to the people who know me, but also to Theo, Angie, Kent, John, Joel, Greg, our entire team, and everyone who trusts us enough to join the first edition of MEGA.
today, I have no idea what the outcome will be. what I do know is that we're doing our best to build something we've always wanted to build. something that will inspire all of us to become even better engineers and humans... to become MEGA developers.
lastly, I agree with Theo.
don't buy this course.
get your boss to buy it.
and whether or not you join us,
wish us luck.
hard and scary is the way value is forged
i peddle (badass) courses for a living, and if you've followed theo for any length of time you know he's traditionally not a huge fan of "courses" in general
*i* was super f'n nervous to pitch him on the idea of doing a course
he insanely busy (like if you think you get how hard this man is locked in, you are not even close, it's more locked in than that), he's intensely focused, he definitely isn't going to pivot to course creator
BUT all of that is exactly why working on this with him is appealing in the first place
if you watch theo regularly you'll also know that he is hands down one of the best educators in the AI space, his videos are topical and "in the moment" but if you watch they are also deeply informative and educational (it's why i love them)
i'm stoked for this cohort and think we are lucky that it exists
but also, holy shit, that's not all
angie.
john.
kent.
yall.
a full month of soaking it up. lfg.
this is also why we're working with @theo
you don't hear things like this from course authors
THE ATTITUDE 🙏🙏🙏
Obligatory call out:
Yes, the course is expensive. That's because we want your company/boss to pay for it.
If you can't expense it, please do NOT buy this course if you aren't making Real Dev Money™.
It's super focused on "how to apply your existing engineering knowledge to building with AI". Genuinely not that useful if you're not a full time dev yet.
Seriously: DO NOT PUT YOURSELF IN A BAD FINANCIAL POSITION TO TAKE THIS COURSE. My free YouTube videos will help you more with getting your first job, I promise 🙏
saas founders, how's your churn lately??
through most my apps churn is accelerating from feb-march.
it's high time to build more moats
can you share? prompt to generate graph in the comment 👇
prompt to generate this:
I'm a SaaS founder. Generate a churn chart from my subscription data (Stripe export, CSV or a connected source) that I can compare with another founder.
Data: only paid subscriptions (no trials, no free plans) from the last 12–18 months. Group customers into cohorts by the month of their first payment.
Compute and show:
1. Monthly logo churn %: customers who cancelled or expired during the month ÷ active paying customers at the start of the month.
2. Monthly revenue churn % (MRR lost to cancellations and downgrades ÷ MRR at the start of the month), plus net revenue retention if upgrades are available.
3. Cohort retention curves: the % of each cohort still paying in month 1, 2, 3 … 12, plus an average curve across all cohorts.
Chart: a line chart of retention by month since first payment (Y axis 0–100%), with a thin line for each cohort and a thick line for the average. Below it, a second chart with monthly logo churn and revenue churn for each calendar month.
Normalise everything to percentages, with no absolute customer counts or amounts, so it can be compared safely. Add a short table: average monthly logo churn, average revenue churn, month-3 retention, month-6 retention, month-12 retention, and the number of customers in the sample. List the assumptions (e.g. how you treated refunds, pauses and annual plans).
the new way of work
we've been serving free, top-notch content every week for the past 14 weeks on mega.dev/drops
#megadev
