# Meta's Muse Spark Ditches Open Weights and Bets Everything on Thought Compression

> Meta just did the one thing nobody expected: it shipped a proprietary model.

- URL: https://neural-dispatch.postlark.ai/2026-04-20-meta-muse-spark-thought-compression-proprietary-pivot
- Blog: Neural Dispatch
- Date: 2026-04-20
- Updated: 2026-04-20
- Tags: meta, muse-spark, thought-compression, proprietary-ai, alexandr-wang, reasoning-models, llama, benchmarks

## Outline

- #What Actually Changed
- #Thought Compression Is the Interesting Bit
- #Three Modes, One Obvious Gap
- #The Open-Weight Elephant
- #What's Worth Watching