Decomposition
Decomposition at Beyond Market Intelligence is a file of 2 stories. The newest of them: “How Netflix Rebuilt Its Commerce Platform for Global Scale” and “Ternary decomposition frees LLM compression from fixed-size limits”. Netflix didn't become a global streaming giant by accident, and Kasia Trapszo's talk digs into the messy, deliberate work behind that evolution. Fixed-size ternary matrices were the wall. Acme AI is the next-generation, AI-powered spreadsheet platform built to replace Excel and redefine how analysts, data scientists, and enterprise teams work… The list below is every Decomposition story on Beyond Market Intelligence, newest first.

How Netflix Rebuilt Its Commerce Platform for Global Scale
Netflix didn't become a global streaming giant by accident, and Kasia Trapszo's talk digs into the messy, deliberate work behind that evolution. She walks through the shift from a U.S. DVD service to worldwide infrastructure, tackling real-world friction like international payments and strict regulations. What stands out is her focus on decomposing monoliths along domain lines, then rebuilding for massive live events. It's a grounded look at how systems survive, not through clever hacks, but through constant, practical reinvention.
Ternary decomposition frees LLM compression from fixed-size limits
Fixed-size ternary matrices were the wall. ExTernD walks around it. The approach decomposes a weight matrix into two ternary matrices with a diagonal scaling matrix between them, and because that inner rank can grow, accuracy stops being a ceiling. It is a direct answer to a limitation most PTQ work treats as a given. The trade-off is a modest VRAM increase over current quantization methods. For the accuracy retained, that cost is easy to justify. This is practical progress, not a promise.