> introduce reproducible image builds This is a great idea in concept, but in practice very challenging. RUN curl https://www.random.org/integers/?num=1&min=1

> introduce reproducible image builds This is a great idea in concept, but in practice very challenging. RUN curl https://www.random.org/integers/?num=1&min=1&max=99999 https://www.random.org/integers/?num=1&min=1&max=99999 Docker will cache this after the first invocation. The build is not reproducible. Now what? Replace "curl random.org" with "nondeterministic and really expensive code build/model training/etc operation". > extending docker push and pull to allow discovery from different sources that use different protocols like IPFS, TahoeLAFS, or filesharing hosts This is great, if you can solve the image integrity/trust issues therein, which should be just some signing/merkle tree work.