Kite Says Farewell
By Adam Smith, Founder
November 16, 2022
From 2014 to 2021, Kite was a startup using AI to help developers write code. Unfortunately, we have stopped working on Kite and are no longer supporting the Kite software. Thank you to everyone who used our product, and a heartfelt thank you to our team members and investors who made this journey possible.
Our Journey at Kite
While we built next-generation experiences for developers, our business faced significant challenges. Firstly, we failed to deliver our vision of AI-assisted programming because we were over a decade too early to market. The technology simply wasn't ready yet. Although we developed the most advanced AI for assisting developers at the time, it fell short of the necessary 10× improvement required to break through, primarily because the state of the art for machine learning on code was not adequate. This is evident in GitHub Copilot, which was developed by GitHub in collaboration with OpenAI. As of late 2022, Copilot shows promise but still has a long way to go.
The biggest issue is that state-of-the-art models struggle to understand the structure of code, particularly non-local context. We made some progress towards better models for code, but the problem is highly engineering-intensive. Building a production-quality tool capable of synthesizing code reliably may cost over $100 million, and no one has attempted that yet.
Despite this, we could have built a successful business without achieving a 10× increase in developer productivity using AI, but we did not manage to do so. Our product failed to monetize, and it took us too long to realize that. We sequenced our business development in the following order: first, we built our team, then the product, followed by distribution, and finally monetization.
Because our product was challenging to build, we focused on assembling a world-class engineering team first, which we accomplished successfully. Next, we concentrated on product development, but we did not achieve product-market fit until 2019, five years after starting the company. It took numerous iterations and significant engineering efforts to reach that point. We then grew our user base to 500,000 monthly active developers with almost zero marketing spend. However, our product failed to generate revenue, as our 500k developers were unwilling to pay for it. Individual developers typically do not pay for tools; their managers might, but engineering managers only want to pay for discrete new capabilities. For instance, making their developers 18% faster when writing code did not resonate strongly enough.
We explored pivoting the business and conducted extensive customer discovery, which led us to a new direction—code search—that could leverage our AI technology and our developer footprint. However, after seven years of intense work and startup stress, our team was too exhausted to pursue that pivot, leading us to seek a soft landing.
The Human Side
After outlining our business trajectory, I want to acknowledge the courage exhibited by our users, team, and investors throughout this journey. To our investors: thank you for your belief in us, and I apologize that it did not work out. I appreciate living in a world where investors take risks, as this risk-taking is essential for progress.
To the team: it’s remarkable how much energy goes into propelling early-stage startups forward. Every day is an act of optimism and selfless contribution to the world. We dedicated seven years to Kite, making countless sacrifices, including taking below-market salaries to extend our runway and working long hours, especially in the early days. Our team came from top backgrounds, and we were frequently approached by recruiters with larger roles and offers elsewhere. However, we wanted to contribute to the world in a unique way, which we did through our efforts at Kite, despite the risk that our project might not succeed.
I regret that it did not work out. We took a chance to dramatically accelerate software development globally. We ran the experiment, and despite everyone's great efforts, it did not yield the desired results. While I believe in taking responsibility and learning from failure, I do not believe in regret. It’s too easy to use hindsight to second-guess past decisions. Given what little we knew when we started Kite in 2014, I can only look back with pride at our courage to take the risk. It is this courage that drives rapid progress.
I hope you take immense pride in what we accomplished together and continue to be courageous in your careers moving forward. I am optimistic about the future, especially considering the impact we are making post-Kite, including alumni who have founded startups like Silo, Zippy, Pipekit, Skipper, StandardCode, Firezone, and many more to come.
The Future is Bright
We eagerly await the day when AI revolutionizes programming. Computers are magical, and it will be incredible to witness a significant increase in their capabilities. While Kite was a pioneering startup in this field, we were not the company to achieve that breakthrough. Although this was not the outcome we hoped for, we celebrate the courage and contributions of everyone who made this experiment possible.
Thank you, and keep building.
P.S. Most of our code has been open-sourced on GitHub . It includes our data-driven Python type inference engine, Python public-package analyzer, desktop software, editor integrations, GitHub crawler and analyzer, and much more.