For a while now I've been building AI-native tooling in public. We've all been building tools from the ground up, like never before. And who can blame us?
As my repository portfolio has grown, I've speculated more frequently about what new development and operational paradigms can emerge with AI more tightly integrated at the source.
So it seemed logical that the launch of Sneka begin with Diffinite, a Git repo service, as the foundation upon which we explore new ways to work and grow with our exponentially increasing volume of software.
how is this different from GitHub?
we start from the mission
Sneka's primary exploration surface will be about SDLC in the agentic age —
- harness patterns
- cognitive control
- dynamic human gating
How can we best leverage the power of LLMs to increase the creativity, quality, and speed of our software?
So the fundamentals of what you need from a Git repo server will evolve organically from those concerns, rather than the other way around. This may ultimately even mean evolving from a "Git repo server" to just a "source code artifact."
different problem, different priorities
Diffinite places different emphasis on the traditional attributes and artifacts of the Git repo server's role in the SDLC.
Diffinite is concerned with forging efficient, dynamic, and trustworthy mutations.
One obvious example is the omission of the pull request as a first-class SDLC artifact. We can certainly debate whether that is "smart" — but what's more important is how we will explore different ways of achieving output veracity to match the accelerated velocity of agentic production.
builder joy is the point
As the initial product of the Sneka brand, Diffinite inherits the value of promoting builder joy.
Building software products should be exciting.
Whether it was your initial hello-world CLI, your first client-server network framework, or your first mobile app, there was a reward that drew us into the craft. Sneka's goal is to provide that to you as we launch ourselves into this new era, together.