On this day: Jun 17

Takes an enterprise view, counting observability and governance as part of the harness

Claims up to 10× the speed of speed-optimized frontier models. The first time a diffusion LM entered a commercial track

Extends the 2017 Software 2.0 frame to natural-language programming, and proposes a human review loop, rather than full automation, as a design principle

The theoretical basis of DSPy's optimizers

Measures policy compliance over multiple turns rather than one-off tool calls

The data-generation pipeline itself became the product

An open-weight code model that DeepSeek claimed matched GPT-4 Turbo on code-specific tasks

Freezes the weights and trains only low-rank matrices, cutting trainable parameters to 1/10,000 for GPT-3 175B, with no increase in inference latency