Achieving automatic parallelization comparable to programmer-controlled granularity is a difficult problem.

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Parallel programming needn't be terribly difficult, but 'thinking in simultaneities' as in message-passing is calculated to make it difficult.claim0.808
  Asserts that Linda's uncoupled style reduces cognitive load.
• Are the operations that create and control parallelism best left in the programmer’s hands, or should they be turned over to the compiler?question0.779
  Central design question underlying the comparison with functional approaches.
• Pure functional languages cannot be the whole story for parallel programming because they fail to provide needed expressivity.claim0.772
  Argument that recursion equations are inappropriate for many important parallel programs.
• We hypothesize that degraded generalization on benchmarks like MMLU may reflect the computational demands of the tasks.hypothesis0.768
  Connecting the paper's task-difficulty findings to prior observations of weak generalization on complex QA benchmarks.
• Explicitly parallel programs are not inherently harder to design or understand than functional specifications.claim0.767
  C-Linda DNA comparison is comparable in length and clarity to Crystal; pragmatic runtime granularity control outweighs compiler optimization ideals.
• Latent methods lack task generalization and are difficult to train with autoregressive parallelization.claim0.763
  Identifies key limitations of latent methods.
• Perfect parallelism is impossible in textual practice; hierarchy always emerges through spatial meansclaim0.746
• Object-oriented programming is powerful and attractive, but irrelevant to parallelism.claim0.742
  Claim that OOP per se does not solve any parallel programming problems.