Hello!
Whats up.
This is a long one, if you want to read about the nitty gritty details, read on.
It’s been an extremely productive last few weeks. Main update: me and Sharif are working on Tidbit — it’s like ChatGPT, but, with channels.
We went to Japan for 14-days — no idea what we were gonna build.
Here’s us eating yakitori:
14-days later — we are back in SF and we got paying customers, users, a ton of attention, a product we’re really excited about, and a lot of clarity. Kinda crazy.
Let me rewind a little.
About 6-weeks ago, I wasn’t feeling it. I was just tired of making stuff by myself.
I don’t know how to say it, I just wasn’t having fun. And if I’m not having fun I don’t think I’ll build good things.
I found I got so much energy when I worked with others.
So, I said fk it and attempted to start looking for a co-founder.
I didn’t expect anything to come from it. But, I thought I’d learn something about myself from the process at least. I wrote this note about the type of person I was looking for, and sent it to 15 friends who I felt might know someone I could work with.
Most were surprised. I was a solo founder at my last company.
I was solo founder throughout the last 5-years.
No one really expected me to ever have a co-founder haha.
But, those 5-years were painful. The amount of mental gymnastics I had to pull on myself to stay sane was honestly impressive. Shoutout to past me. Perhaps one day I’ll talk about it more.
Anyways, I thought I could learn from my past.
And one thing that would have been better, was a co-founder.
One person I sent the note to was Sharif.
I was surprised by his reply:
Some lore:
I met Sharif nearly a decade ago now — online on a HackerNews post.
I was 21 and had just finished college in Orlando and was working on training ai models for video games. He was 19 and still in college and had recently launched a distributed GPU product named VectorDash.
I was extremely impressed by it.
He was just a random internet guy with like 100 followers like me.
So, I DM’d him saying it was cool, expecting nothing from it.
Over the next few months we became really good internet buddies.
We’d Discord call all the time and talk about random stuff in the DMs. We’d mostly talk about GPUs, machine learning, and whatever else 20-somethings talk about lol:
Sharif then decides to drop out of college to work on Vectordash.
He decides on SF.
We end up settling + rooming in SF together.
Honestly, looking back we were just kids.
It’s important to say we both had (and still have) an insane amount of motivation and ambition. But, we were also both nobodies from the suburbs of the east coast and had no idea how SF worked. We came to SF with nearly no money, and no real plans.
So, even though we had literally never met irl we just kinda trusted.
Over the next 7-years, I end up building out kanga, then zipschool + buildspace as a solo founder. He ends up building out vectordash, then debuild + lexica as a solo founder.
We both build a lot of things.
We both go through a lot of tough times.
We both stay really good friends.
But, we never worked together. It was always a timing thing. But, also a maturity thing. Back when I was 21, I was an ass. I had a massive ego. I thought I was the best. It was embarrassing. It took a very long time for me to change and learn to work with others.
Today I’m 29, Sharif is 27.
That’s a long time in startup-land. You grow up fast.
So when Sharif hit me up to work together for a few weeks on something I think it just felt right. And in classic Sharif + Farza fashion — we decide to go to Tokyo for 2-weeks to figure out wtf we wanted to make. We decide on a Sunday. We’re on the flight to Japan on Tuesday.
Why?
No reason. We just like Japan.
Also, we both believe in the importance of changing up your environment to get new ideas and thoughts.
We booked the ticket without any plan or idea.
But, we didn’t worry much at all though. Both me and Sharif have the same philosophies when it comes to building stuff — you gotta try stuff to find stuff.
So, we were gonna aggressively try stuff. That was the plan.
Our first spot was in Setagaya City, Tokyo — a quiet suburb:
We also had a supermarket near by, so, we were able to buy a bunch of fresh Japanese food easily. All this was pre-made from the grocery store. We ate so much Yakitori.
Protein all the way.
So, we were able to focus purely on making stuff!
Day 1 at 9AM in Tokyo the plan was just to individually prototype something and then demo it to each other at 3PM. We both talked at length about how a new interface on top of these models could exist, so we both made prototypes based on that convo.
Sharif built a desktop app that looked exactly like ChatGPT, except with Discord-like channels.
I built a Discord bot where you could essentially prompt the model with one other person and actually do real work together, alongside the model. We actually came up with the title to my latest YouTube video together, alongside Claude.
(Initially, it was quite bad. But, the vibe was there. And we were amazed how fun it was to come up with ideas together alongside AI with this simple prototype).
I loved Sharif’s demo of channels.
He loved my demo of a multiplayer interface for these models.
We combined it into a rough product vision by drafting a bunch of one-liners and potential launch tweets. For us, if we could describe the idea in one line or make it sound exciting in a tweet — then we’d be on the same page.
Here are a bunch of our drafts.
From there we worked backwards from the one-liner and got to work.
For the rest of the days, it was the same routine:
Every morning around 9AM we’d spend 30-60 minutes to figure out what we wanted to do by 9PM. Do the thing. And by 9PM hopefully do what we said we’d do.
We both coded a lot. (My code was worse)
And slowly, things started coming together — channels, invites, context management, etc. A majority of the code we wrote for the entire trip was written by Claude Code.
(We paid the price in SF though, turns out, Claude just stacks technical debt lol).
We had many of these at the local Dennys + local cafes.
Japanese Dennys is amazing.
Japan in general was an inspiring place to work and we could just walk outside and get inspired or get a breath of fresh air. We were in a really quiet part of town. We were the only foreigners.
These morning check-ins ended up being extremely important.
Again we had never worked together.
We’d check-in each day around how we were working together + talk through any issues we were having around working together and just figure them out immediately.
We’ve been friends for a while, and neither of us are sensitive to personal feedback, so it ended up being easy.
Main goal of trip became obvious by like Day 5 — build a product that we really loved using. If we went back to SF with a product we used everyday. That’s a massive win. To us, this was the main thing.
We also agree to start posting daily demos immediately. We were gonna be shameless. Didn’t care about polish. And we also agree to get more real people using the product. We were optimizing for learnings. Not a user number, or a graph, or a money number.
The more we learned, the closer we could get to a product we loved.
That same day we shipped a demo to Sharif’s Twitter. It did really well.
More importantly: we had a few of our friends show interest, a really good sign.
Again, at this point we had no idea if this was interesting. So, every message from a person showing interest gave us a good push.
A day later we shipped another demo to my Twitter. We recorded the demo at like midnight and we were hallucinating from jet lag but shipped anyways.
“Demo driven development” was the goal. Forced us to get our thoughts in order. We didn’t wanna put out absolute shit. But, a demo was always cool.
This time we focused more on the channel aspect. And from there momentum just came.
People loved it. Pretty soon, we collectively had about 1000+ DMs.
Founders who wanted to use it with their teams + co-founder to brainstorm. College students who wanted to collab with other students. People who wanted to use it with their wives/family/friends to plan trips in a more collaborative way with AI.
And and a ton of ChatGPT and Claude power-users that simply loved the long-context channel idea because they were tired of re-explaining themselves to Claude.
These were use-cases we never even thought of.
This was amazing, people were coming up with the product for us.
Some people didn’t even follow the rules and just DM’d me “Access”.
But I still gave these folks access cus I’m a nice guy ofc ofc.
By this point, it was clear we had…something.
From there, I booked like 20 user calls and grinded convos with users and Sharif made the product better every 2-hours. We had a really fast loop going.
My main goal of the convos was to get a clearer and clearer idea of why people were excited about it. For me, showing people the product and talking to them essentially prompts my brain in new ways and gives me ideas.
Sometimes I’d just give access to free to someone and closely track the user in Posthog. Sometimes I sent a Stripe link in the DMs before hopping on a call with them.
It was random.
Bunch of experiments. Again, optimizing for learning.
And tbh stuff like this continued until the end of the trip.
I’d make noise, get people in the flow, we’d talk to users, we’d improve product. Me and Sharif were just next to each other 24/7.
And that’s where we’re at now.
It’s extremely obvious that users are excited about the idea + the problems it solves.
It’s clear that people want longer context convos and are tired of reexplaining themselves to Claude/GPT. It’s clear they want to more efficiently collab with their team + the model instead of passing around GPT and Claude links 24/7 in Slack.
This week we’re taking a lot of the feedback and just making the product better. I also have like another 25 user calls so that’s good. Me and Sharif gonna keep the same loop.
So Japan? Massive success.
We’re cooking.
Will me and Sharif keep working together? Yeah. We’ll just keep feeling it out. Keep shipping. Let it all happen as it should. And, see what happens :).
See you next week.
Also — I wrote this essay, Money or Dreams.
And, I shipped this YT video:
Lastly, highlight of trip (apart from jumpstarting a company lol):