Skip to main content
TOWARD//AGI

// personnel file

Mira Chen

Contributing Analyst

Mira covers open-weight model releases and inference economics. Former infra engineer, current benchmark skeptic.

// dispatches by Mira Chen

THE-AFFORDABLE-MODEL-SQUEEZE
Trends

The Affordable-Model Squeeze Is Coming for the Frontier

DeepSeek's flash tier and a wave of cheap, capable models are forcing the big labs into a choice: match on price and token plans, or differentiate on image, speed, and ecosystem. Neither option is comfortable.

3 min read
LLAMA-4-SCOUT-LANDS
Model Release

Llama 4 Scout Lands: What a 10M Context Window Actually Buys You

Meta's new mixture-of-experts flagship is open-weights, agent-tuned, and absurdly long-context. We break down the architecture and what it means for self-hosters.

1 min read
OPEN-WEIGHTS-CLOSING-THE-GAP
Trends

The Open-Weights Gap Is Closing — And It's Closing Fast

A year ago, open-weights models trailed frontier closed models by a wide margin. Now the gap is shrinking quarter over quarter. Here's what changed.

2 min read
INFERENCE-COST-IS-THE-NEW-BENCHMARK
Trends

Inference Cost Is the New Benchmark

Benchmark leaderboards are saturating. The metric that now decides model selection is cost per million tokens at acceptable quality — and it's reshaping how labs build.

1 min read
INSIDE-DEEPSEEK-V4-REASONING-STACK
Deep Dive

Inside DeepSeek V4's Reasoning Stack

DeepSeek V4 posts frontier-class math and code scores at a fraction of the inference cost. We dissect the publicly documented pieces of its reasoning pipeline.

1 min read
DEEPSEEK-V4-REASONING-FIRST
Model Release

DeepSeek V4: Reasoning-First, Cost-Second

DeepSeek's latest flagship posts frontier-class math and code scores at a fraction of the inference cost. Here's what matters about the release.

1 min read
← all authors