The wiki

How GPU efficiency, completion design and workload-specific evaluation turn model capability into everyday utility.

Long-running coding agents depend on familiar tool interfaces, context management, and working habits developers can inspect and evaluate.

An AI developer-tools company that grew from IDE test generation into code review and enterprise governance, adopting the name Qodo in 2024.

The applied AI company behind Devin, DeepWiki and Devin Desktop, combining autonomous software engineering agents, an IDE and coding-model research.

An AI podcast and newsletter combining interviews with researchers and builders, commentary, and Nathan Labenz’s educational AI Scouting Reports.

AI/ML engineer and consultant who created OpenInspect, an open-source system for shared background coding agents, and advocates company control over agent infrastructure.

The pseudonymous creator of ComfyUI, whose experiments with diffusion models led to a node-based interface and a focus on local execution.

An open-source generative AI workflow platform that grew from diffusion experiments into local and cloud tools for creative production.

A second image-generation pass became a model-composition tool—and made local inference, memory management and extensibility central to creative control.

Ahmad Awais’s company, formerly Langbase, builds a coding agent that learns developers’ preferences from their interactions and carries them across projects.

Dagger’s approach to coding agents starts with their working environment: isolate parallel work, preserve reusable setup, and let humans inspect what happened.

A research agent’s growing tool history turns context into an engineering problem—one that reaches into retrieval, multi-agent coordination, memory, and the architecture around the model.

CoreWeave combines GPU infrastructure, model-serving tools, shared-cluster scheduling and agent execution in a capital-intensive AI cloud.

A British AI company whose work evolved from codebase retrieval and the Genie software-engineering agent to secure deployments and the planned Lumen Sovereign model.

From Statsbot’s metric-to-SQL infrastructure to an open-source semantic layer and first-party agentic analytics.

Cursor is Anysphere's AI software development platform, combining an editor, coding models and agent execution environments. Its evolution shows how retrieval, tool use and reliable environments support coding agents.

From CAD operations to code edits, Cursor’s early development showed why useful generation needs context, tools, and a workflow people can inspect and steer. This historical snapshot accompanies the original recording published August 22, 2023.

Configured cloud computers let agents run applications, demonstrate changes, and work in parallel. The harder handoff is turning that output into software people can confidently ship.

CuspAI develops generative AI and molecular-simulation tools for materials discovery, combining computational design with industrial expertise and experimental feedback.

A materials search engine needs generators, progressively tougher evaluations, experimental feedback, and chemists who can recognize missing requirements.

AI and computational neuroscience researcher whose collaborative work examines learning and forgetting, efficient local–cloud inference, and models that learn from experience.

Machine learning and systems researcher working on hardware-aware GPU kernels and efficient model architectures.

Machine learning researcher and rekursiv.ai co-founder whose work spans autoregressive video generation, efficient video recognition and multilingual parsing.

Prediction Guard founder and CEO, Practical AI co-host, and practitioner of low-resource NLP and distributed machine-learning research.

AI researcher at Noetik whose work spans self-supervised visual models and neuroscience.

Cognitive scientist working on human–agent collaboration, interaction training data, and evaluation of AI’s effects on human thinking.

Anthropic CEO whose collaborative research spans AI safety, learning from human preferences, and language-model scaling, and whose public writing addresses powerful AI’s benefits, risks, and governance.

Arize AI’s Head of Agent Strategy, whose technical writing addresses prompt lifecycle management and practical evaluation of retrieval-augmented applications.

Databricks combines data engineering, analytics, model development, and agent infrastructure in a shared platform, building on its origins in Berkeley research and Apache Spark.

How tokenization, filtering, data mixtures, and training rights shape a model’s knowledge, deployment costs, and benchmark results.