retire(blog): drop scripts/{freeze,setup-freeze-env}.sh + freeze-era docs
The freeze model is fully retired on both CI lanes:
- .github/workflows/gh-pages.yml renders Julia LIVE (julia-actions/setup-julia
+ xvfb-run quarto/hugo); see git history at fb831a1.
- .tangled/workflows/render.yml renders Julia LIVE off the R2-distributed
depot + render-env closure (see .tangled/workflows/build-depot.yml).
Neither lane reads scripts/freeze.sh or scripts/setup-freeze-env.sh any more,
and the referenced .github/workflows/freeze.yml is gone. Drop the dead scripts
+ rewrite scripts/README.md to document the actual two-lane flow; update the
stale 'tangled renders Julia-free' comment in _quarto.yml.