Commit the typechecker as an actual lambda

As of now, we're committing the typechecker as a datum and calling `eval` on it in Lurk.

We'll probably have a faster typechecker in Lurk if we commit the direct lambda expression.

No files, tests, or entry points are named. Start by locating where the typechecker is committed as a datum and evaluated in Lurk, then trace how a direct lambda expression would be represented and invoked. Done means the typechecker is committed as an actual lambda rather than a datum passed to eval, with the expected faster behavior verified.

Written by the indexing model from the issue text.