Microsoft

The optimization target of harness engineering has widened from "what to fix" to "what to test". Abstracting failures into patterns to compose the curriculum automatically takes over the place where people used to pick s

A commercial implementation that moves prompt-injection defense down from model alignment to the network boundary. Starting from the premise that agents cannot be kept from being fooled, it keeps data from getting out ev

The operational foundation for agents moving from "running when a person calls them" to "waking up on their own" has become a GA cloud product. It matters that the option of giving agents their own identity came out at t

Shows that as ASR moves to LLM-based architectures, latency and customization of domain vocabulary, not just accuracy, have become competitive axes.

An event that revealed the definition of AGI had really been a contractual and governance device

Made local inference, free of cloud lock-in and per-token billing, an official product line

The finished form of a three-way structure of cross-investment and purchasing

Moves any declaration of AGI to confirmation by an independent expert panel, and lifts constraints on raising capital

Adoption reached down to the OS level, and governance began to become multilateral

The first official adoption by a large platform; the turning point from a single vendor's spec to an industry spec

The first proposal in the prefill-decode disaggregation line

An early demonstration that FP8 training actually works for large models

An improvement that made ZeRO work even on low-bandwidth clusters

Early work on the idea of textual gradients. Led to TextGrad

The starting point of the open VLM lineage

Publicly demonstrated that in a live service the system prompt does not work as a line of defense

A rebuttal to the challenge BEIR had raised. The starting point of the open-weight embedding race

The moment a training-infrastructure team moved down into inference. The standard for serving large models in the generation before vLLM

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

Got around the GPU memory wall by extending into the storage tier. Established the framing that the constraint is bandwidth, not memory

The start of the lineage that lowered the hardware barrier to entry for training large models

Removed the memory barrier to 100B-class training

Separate industry groups confirmed the same shift (HMM→DNN), and deep learning became the industry standard for speech recognition