* Better than potential successors of SOTA transformers: Mamba, Hyenna, RWKV, xLSTM; * Gemini/claude estimate potential ip values in the millions; * All implemented in C (ok this may be a minus) but also ported to C# and F# - so no python nor rust; * humble code size: 10-15k lines of code but i mean come on gpt-1 was under 1000 lines...in python.
In short:
the problem: transformers are slow and could be smarter; solution: you have fast and smart alternative. 10x!
Your options would be something like: Open source it, publicize it. Maybe get hired by an adopter. Maybe you could sell consulting around it.
Patent it, publicize it, sell it.
Patent it, publicize it, license it.
Patent it. Develop a service around it. Grow the business.
Most of this could be done with a bootstrapped approach, or with VC money. If you get VC money or sign any other kind of contract, get a good lawyer first. Don’t make the Spyglass Mosaic mistake.