README.md: future work on this has long stalled and isn't going to re…

…start Reduceron is still an eye-opening project, but I lost interest for a number of reasons, including: 1. The current implementation is not very easy to modify. Many assumptions are implicit and it's very painful to debug. 2. The reduceron model isn't easy to scale. Like all pipeline processors we'd need to speculate on the instruction stream, but by nature, Reduceron instructions are very dynamic and it would be difficult and expensive. 3. Rust appeared. Rust gives me 95% of what I enjoyed from Haskell, but with much much better performance, tools, learning resouces, community, and even job market. Oh and no space leaks, how about that!