The BLAS have long supported portable high-performance computing. We owe the BLAS a lot, but have they become an impediment for scientific innovation? Imagine that• Optimizing the BLAS for a new CPU architecture can be done over the weekend. • Mixing operands of any precision and domain is supported. • Implementing tensor contraction and other BLAS-like operations is easy. • Understanding the underlying science is accessible to experts and novices alike. •.Interacting with a vibrant community is available. • Utilizing an open-source, sustainable, extensible, well-documented, and well-supported code base with backward and forward compatibility is the new norm. As the Science of High-Performance Computing group and Matthews Research Group wrap up their CSSI project, these will soon become reality, yielding the BLAS-like Library Instantiation Software (BLIS), a user community, and a collection of educational materials.We further imagine a future project for higher level linear and multi-linear algebra functionality that builds on this progress.