Blorp is a new programming language that combines approachable syntax with performance typically found in lower-level languages. The project targets four main goals: confidence through pure functions and explicit effects, speed via native code and structured concurrency, approachability with minimal syntax, and durability through typed failure handling. These design choices aim to make code both readable and reliable.

The language emphasizes static safety with strong typing, checked imports, and exhaustive pattern matching. Pure functions separate deterministic logic from I/O operations, while value semantics and ARC/COW provide efficient memory management. Structured concurrency features scoped tasks, joins, and channels that keep concurrent operations bounded and predictable. The compiler also performs compile-time bounds checking for arrays and matrices.

Benchmark results on an M4 MacBook Air show Blorp performing competitively against C and Go, with significant advantages over Python. Numeric loops run in 0.1242 seconds compared to 0.1215 seconds for C and 5.1754 seconds for Python. The array_sum benchmark completes in just 0.0011 seconds. These results suggest Blorp achieves near-native performance while maintaining higher-level abstractions.

The compiler uses Hindley-Milner-style inference to minimize type annotations while checking imports and fallible values. Code lowers through Core IR to generated C, which then produces native binaries. Perceus-style ownership analysis with reference counting preserves source semantics. This architecture allows developers to write safe, expressive code without sacrificing execution speed.