🏠 casa.dev is closing this week. 🌅
I started Casa about 18 months ago on the tailwinds of testing ideas for ML/AI infra platforms. During that time, I was inspired to explore and re-imagine the space our software – and the content around our software – lives in.
Since then, I built and launched two products, failed at fundraising and accelerators, landed two customer pilots, and had hundreds of engineers and a couple of teams use Casa. It's been polytropic.
Conviction is vital fuel for entrepreneurs and founders. Without conviction, the entrepreneur cannot truly invest themselves in their company. Once the conviction fuel tank dips into the red and stays there for too long, it's time to move on.
I began to question my conviction in Casa around the time a VC I really respect talked me through some questions about life and work. During this time, a customer pilot was wrapping up, and it didn't convert. After I analyzed their feedback, I decided to try and find other solutions in the market that solved my users' problems so well that continuing to work on Casa would feel like a waste of time and energy. If my search turned up empty, that would be a signal to continue on with Casa.
That exploration led to two very important discoveries and conversations.
The first uncovered lots of products GitHub was working on in closed beta trials. Some of those products revolved around the work of GitHub Next.
Projects like Copilot Workspace and Spark hit sweet spots for contextual integration and amplifying software production abilities. With Copilot Workspace, anyone can create a plan to direct technical work alongside active workstreams in GitHub. With Spark, anyone can rapidly test ideas and experiment alongside code running in production.
These products nudged users to locate more technical documentation inside GitHub itself, thus making AI interaction on the platform more robust and context rich. More quality content (inputs) means more quality context (outputs). A classic tale.
For me, the most impactful product was GitHub Copilot Extensions, and accessing GitHub Copilot directly from the browser. That was a key problem my users had, and I thought copilot extensions solved it to great effect.
Developers wanted the ability to say to a non-technical coworker or manager "All you have to do is open up GitHub and ask how things are going". No need to manually aggregate content across services to assess progress; it's summarized right there for you across pull requests, issues, and discussions. Furthermore, extensions took this to the next level by enabling agents with access to specific external context (such as tasks, design, and docs) right in copilot chat via an extension. This meant with a simple @jira, @figma, or @google-docs in a copilot chat, you could retrieve information from those sites about important context pertaining to the software you were working on, all while staying alongside your code in GitHub! Brilliant!
With that problem solved, we still have our work cut out for us concerning data privacy, security, compression, and sheer scale. Not all large organizations may allow their agents to exist alongside other agents, or in another organization's infrastructure. If they do, how do multi-tenant platforms ensure data privacy at such a massive scale? I believe this points to major evolutions in design, systems, and infrastructure for AI adjacent software development. What's not going to change in the next 10 years is the need for fast and secure interactions with complex software systems. This need is amplified by scale and security concerns AI brings to disconnected and proprietary data when merging with end-user applications.
For Casa, this wasn't the vision, and I didn't want to make that pivot. It wasn't features like multiplayer editing, automatic preview environments, or GPU access in CI that mattered to users; what mattered was simplicity, security, and actually solving their problems. Both pilot customers didn't want to move their code away from GitHub in the end! Larger teams of engineers (100+) already have so much in place in terms of security, compliance, CI/CD, and more on GitHub that it just didn't make sense to switch without parity and peace-of-mind.
Other related products like Replit Teams, Pierre, and Codeanywhere are innovating in the space too, and they'll likely face similar challenges of critical mass and momentum.
Which brings me to the main takeaway of my second conversation: critical mass.
This conversation was with a technical founder of a very successful devops and cloud infrastructure company, and his main suggestion was to double down on where people currently are, and begin solving problems for them where they are. From there you can innovate in new ways and experiment. If I wanted to continue on with Casa as is, his question hit home; "do you really want to spend the next 10 years of your life getting all those developers to do pretty much the same thing on your product?"
My conviction meter in Casa as a business hit empty.
He left me with this, and I hope future and current founders reflect on it too.
Founders dedicate their life for a half decade or more to something before realizing – oh shit, this is very cool but not at all a viable business.
So here are two points of advice I can offer:
1. Have a team
This is the fifth startup I've been a part of. Three failed as businesses, and two, at the time of writing this post, are worth more than $250M combined. Want to know the difference between the ones that failed and the ones still doing well? Team.
Every time I founded a company without a team, the company failed. I might have made it out ok financially, grew personally, sharpened my skills, grew my network, and so on, but I failed to create something bigger than myself, and that's really unlikely to happen on your own at venture scale.
I now advise against solo-founding after this experience. If you must solo-found and find yourself in a waiting period because a friend or co-founder isn't quite ready to join for whatever reason; do some consulting first, make some money if you need to, and learn about the problems of your users before going all-in by yourself. The building, selling, and collecting overhead as a solo founder isn't sustainable for a venture scale business, if that's what you're out to achieve. Investors understand this too. I think most investors are smart. They know it makes more sense to invest in a team rather than one person on average.
For non venture-scale businesses, I would ease up on solo-founding. You can make impressive revenue as a solo technical founder shipping SaaS and consulting here and there. I did it, so can you, but that's not what I wanted with Casa.
With a team, you're making meaning that's bigger than yourself. Making meaning before making money is a strong signal that you're onto something unique and truly valuable. Which brings me to my second point.
2. Know your financial goals and growth opportunities
Should you choose to fundraise for your venture scale company, remember this: do you have conviction in 1000x-ing investor money? Is your business tuned for that? Are you? Your circumstances? What do you really want out of this financially?
It's important to understand your financial goals because investors have them too. Additionally, you should be aware of venture capital incentives and how they're changing. You might have heard about the "2 (% fees on AUM) and 20 (carry - % incentive fees from profits above a certain threshold)" rule, which is a common structure that VC and hedge funds use. This is a great podcast that talks about how this incentive structure is evolving and affecting the behavior of some larger VC firms. Give it a listen. You might get a sense of how likely an investor is to invest at a particular stage based on their financial return strategy and the state of their fund and/or firm.
Even though I was a solo-founder, it takes a village. Lots of people have helped me during this experience with advising, networking, and product testing. A special thanks to my wife, love you!
If you're reading this, and you were one of those people, maybe you simply joined Casa's waitlist, were an advisor, someone I interviewed, a friend, user, or customer – I sincerely appreciate you. Thank you.
So that's all folks. I'm on to figuring out what comes next. Maybe it's founding again (with a team for sure), or joining a team on a meaningful mission. More on that soon.
If this resonated with you and you'd like to chat more about what's next, it would be great to hear from you.
Onwards.