Data

IEEE Spectrum walks through the 2026 AI Index — training compute curves, evaluation saturation, open-weights share, and a sharp rise in domain-specific benchmarks. The least breathless read of the Index so far.

The Dropbox engineering team walks through how they used the DSPy framework to systematically optimize prompt-based relevance judgments for Dash at production scale. Covers the full loop from metric definition through prompt compilation, with real before-and- after numbers. The clearest DSPy-in-production case study to date.

An end-to-end walkthrough of training transformer-based mRNA models across 25 organisms in 55 GPU-hours for under $165. Compares architectures and demonstrates species-conditioned codon modeling.

NVIDIA's pipeline for rapid embedding model fine-tuning using synthetic data generation, with step-by-step instructions for adapting general-purpose embeddings to specialized domains.

Answer.AI analyzed PyPI data to test whether AI tools are actually boosting software production and found no obvious increase in package creation post-ChatGPT. The only measurable effect is a concentrated surge in updates to AI-related packages themselves — likely driven by funding cycles rather than universal productivity gains. It's the most data-grounded skepticism of AI productivity claims published this year.

Forge lets enterprises build frontier-grade models trained on their own data, positioned as an alternative to fine-tuning and RAG for domain-specific performance.

A developer experiment where an AI agent iteratively redesigned its own memory system, improving recall accuracy from 60% to 93% for about two dollars in API costs.

An interactive tool displaying growth projections, compensation, and AI exposure metrics for hundreds of US occupations. Built by Karpathy, it's the kind of side project that reframes how you think about career risk.

The sixth edition of a16z's consumer AI ranking reshuffles the leaderboard. ChatGPT still dominates, but vertical apps are climbing fast and retention patterns are shifting.

New study quantifying AI assistants' share of global search volume — with mobile dominance accelerating the shift away from traditional search engines.

A practitioner had Claude read and categorize every AI safety paper published since 2020, producing a curated, searchable database for navigating the field's rapidly expanding literature.

Fortune reports on Anthropic's landmark research introducing "observed exposure" as a new metric for AI labor impact. Computer programmers show 74.5% theoretical exposure but only 33% actual usage, and employment data shows no unemployment impact yet — though hiring for young workers in exposed occupations is slowing.

Independent analysis showing AI apps have crossed 1B weekly users with ChatGPT holding 900M, but questioning whether usage patterns reflect genuine habit formation or novelty-driven adoption.