Topics
Ten areas where early technical changes tend to matter for developers. Each topic page lists current and historical signals, the areas we track, and how evidence is judged in that field.
Loading topics…
Ten areas where early technical changes tend to matter for developers. Each topic page lists current and historical signals, the areas we track, and how evidence is judged in that field.
Technical developments across machine learning systems: models, training and inference techniques, evaluation, and the tooling developers use to ship AI features.
Systems in which a model plans and calls tools across multiple steps: tool protocols, execution sandboxes, durable workflows, memory, evaluation and observability for agent runs.
Large pretrained models and the techniques around them: architectures, training recipes, fine-tuning methods, open weights, context handling and model evaluation.
Editors, language servers, build systems, testing, debugging and workflow tools — including tools that change how developers work with AI in their codebase.
Serving, inference engines, retrieval and vector storage, accelerators and schedulers — the layers underneath AI applications that determine cost, latency and reliability.
Open-source projects whose momentum, governance or licensing changes affect what developers can build on — tracked through repositories, package registries and maintainer activity.
Robot learning, simulation, control and the software stack for physical systems — with attention to results that are reproduced outside the lab that produced them.
openwam: Official repository for "OpenWAM: An Open, Modular Exploration Towards Systematic World–A…
Quantum hardware milestones, error correction and mitigation, SDKs and algorithms — and the post-quantum cryptography work that affects developers well before large quantum computers exist.
Security developments that change how software is built and operated: supply-chain attacks, sandboxing, memory safety, cryptographic migrations and security tooling.
Runtimes, compilers, instruction-set architectures, databases, networking and operating-system work that shifts performance, portability or cost for software teams.