19 August 2026

Two open-source AI development tools released

• Miles, a reinforcement learning framework developed with 72 contributors over nine months, became available for training language models like Kimi K3 and DeepSeek V4.
• Mojo, a programming language for GPU computing, released version 1.0 and open-sourced its compiler under Apache 2 license after shifting away from full Python compatibility.
• Both projects moved from closed to open development, though Mojo represents a strategic pivot from its original goal of being a complete Python alternative.

Where they differ

Latent Space covered Miles, a machine learning training tool, while Simon Willison reported on Mojo, a programming language for GPU tasks. These are separate projects with no overlap.

Latent Spaceswyx & Alessio

Miles is a new open-source reinforcement learning framework developed over nine months with 72 contributors, battle-tested on models including Kimi K3 and DeepSeek V4.

Simon WillisonDaily notes and links

Mojo released version 1.0 and open sourced its compiler and toolchain under Apache 2 license, fulfilling a promise made in May 2023. The language has pivoted from aiming to be a full Python superset to being its own language optimized for GPU programming with Python-inspired syntax, relying on AI tools to help migrate code rather than full compatibility.