Make Yourself AI is now a member of NVIDIA Inception, an accelerator program for technology startups.
We're grateful to be included. This is more than marketing, compute, and support. We're moving down the stack: beyond context management and into model fine-tuning, to deliver better outcomes for our customers.
From steering the model to tuning it
Most AI products today work by instruction. You write a system prompt, you retrieve some documents, you hand the model a manual and a library and ask it to consult them before it speaks. It works, and we do a lot of it. But the model is always looking things up. There's a cognitive distance between a person's judgment and the agent's execution.
We think the next step is lower in the stack. Not just giving a model your context, but changing how it reasons through that context, so the way a specific person weighs a decision becomes part of how the model thinks, not a note it reads at runtime.
That's the whole idea behind a mirrored agent: an extension of a particular human, not a generic assistant wearing their name tag. Context is an external map; what we're building toward is an internal compass.
We're less interested in teaching an AI what we know, and more in teaching it how we judge.
Here's a simple example. Someone comes to myai wanting to build an app. The reflex of most AI tools is to start generating React code. myai does something different first: it asks what problem you're actually trying to solve, whether you even need an app instead of a workflow or an email integration, and how you plan to handle adoption and change management once it exists. The whole crux is asking better questions before rushing to an answer.
Today we get that behavior with careful context and prompting. Moving down the stack is how we make it native, not an instruction it has to be reminded of. We'll keep most of the how to ourselves for now. But the direction is the point: lower in the stack, closer to the person, so working with an agent feels less like delegating to a tool and more like working with someone who gets it.
© 2026 NVIDIA, the NVIDIA logo, and NVIDIA Inception are trademarks and/or registered trademarks of NVIDIA Corporation in the U.S. and other countries.