The wiki

Developer and founder of Command Code whose AI work focuses on learning coding preferences and composing agents from infrastructure primitives.

A practical account of coding agents: how they use tools, recover from errors, reuse workflows, and turn human feedback into completed work.

Video generation, segmentation with memory, faster detectors, and the challenge of teaching small models to see precise details.

Statsbot’s early Slack interface exposed two problems: understanding a question and understanding the data. Cube’s semantic layer separated them—and foundation models reopened the conversational side.

Part 3 of 3: the closing exchange of Latent Space’s conversation on NVIDIA’s AI engineering.

Andreessen Horowitz combines technology investment with operating support, experimental developer tools, and analysis of AI businesses across models, applications, and physical infrastructure.

The Agentic AI Foundation is a Linux Foundation directed fund hosting open protocols, tools, and infrastructure for AI agents.

Box’s CEO and cofounder has focused on enterprise AI agents, the information they need, and the permissions and organizational changes required to put them to work.

dbt Labs security leader whose AI governance work emphasizes enforceable constraints, recoverable mistakes, and informed human oversight.

Product manager who helped create Gemini Deep Research, with an emphasis on editable plans, visible sources and research grounded in users’ needs.

Research scientist and language-model architect who led MPT-7B training at MosaicML and DBRX development at Databricks.

Systems engineer who built Arrakis, microVM-based sandboxes for AI-agent computer use and code execution.

Healthcare AI company whose products evolved from recording patient conversations to clinical documentation, decision support, and revenue cycle workflows.

Ambient documentation gives Abridge a starting point for a harder problem: delivering accurate, timely assistance while clinicians can still act on it.

Heroku and Ink & Switch cofounder whose work spans AI interfaces, personalized email triage, local-first software, and portable application architecture.

Engineering leader working on Claude Code, publisher of workflows for AI coding agents, and advocate for human responsibility in automated development.

An AI company whose 2022–2024 work combined software-action models and enterprise workflows; in June 2024 it narrowed its focus as co-founders and some staff joined Amazon.

Part 4 of 4 moves from an open-source community gathering to Julius AI’s practical priorities: faster answers, multiple tool calls, visual feedback, and building useful products alongside a competing platform.

From broader task coverage to persistent codebases, coding evaluations are confronting what it means for an agent to do useful development.

Pydantic’s approach to agents connects data contracts, explicit control flow, and visibility into running applications.

Osmantic’s founder and CEO advocates local, open AI built around usable tools, specialized models and control of compute.

Nathaniel Whittemore’s AI news podcast and publication, combining news analysis with conversations about models, tools and developer practice.

A conference and technical education organization focused on building with foundation models, with public recordings, open-source conference tools, and an agent-accessible knowledge library.

A June 2024 conversation about turning foundation models into products, learning from deployment, and planning for changing inference costs and capabilities.

Unexpected polymer designs, active learning and quantum chemistry reveal what it takes to turn a promising prediction into a useful material.

As assistants take over discovery and reading, businesses need content that survives search selection, answers concrete questions, and reaches the model intact.

Closer model competition widened the opportunity for startups. Conviction’s NeurIPS 2024 keynote examined where that opportunity depended on better models—and where it depended on data, distribution and execution.

The dbt–Fivetran merger, oversized seed rounds, and the product case for memory and continual learning.

Generating a clip is becoming easier. Sustaining a character, controlling the result, and turning attention into a business still take work.

The final part follows AI infrastructure from financing and chip supply to memory budgets, specialized accelerators and CPU demand, before closing on writing habits and the value of stepping away.