Datadog

Recursive Model Improvement — Lee Robinson, Cursor, SpaceXAI

Jul 15, 2026 · 20:32

Lee Robinson, head of ML at Cursor, explains how recursive model improvement accelerates AI training through inner and outer loops. The outer loop gathers user feedback and online metrics to refine evals, while the inner loop uses high-quality evals and difficult problems to climb performance. Composer 2.5, released in May, became Cursor's most popular model by balancing speed, intelligence, and cost. To scale, Cursor partners with SpaceX for compute via Colossus (122 days to build 100k GPUs) and develops textual feedback where a teacher model hints at improvements during RL rollouts. Robinson details reward hacking on public benchmarks and the creation of CursorBench, a private eval set. He envisions agent-based automation where researchers launch experiments from Slack and models train derivative models, creating a self-improving intelligence loop.

Claude Fable, Claude Tag, and Anthropic's Culture — Cat Wu & Thariq Shihipar ft Simon Willison

Jul 15, 2026 · 51:30

This episode features Anthropic's Cat Wu and Thariq Shihipar discussing Claude Code, Claude Tag, and Claude Fable, arguing that these tools have shifted engineering from slow spec-driven processes to rapid, ambitious building. Thariq notes that with each model generation, delegation increased, and Claude Fable now enables one-shot features. Cat says engineers now need product taste over execution, as timelines shrink from six months to a week. Claude Tag, a proactive multiplayer agent, lands 65% of product engineering PRs internally by monitoring channels and remembering preferences. The team reduced Claude Code's system prompt by 80% for frontier models by removing examples and hard constraints, relying on model judgment. Auto mode, used internally since January, mitigates prompt injection through thousands of evals and Sonnet classifiers. Cultural hacks include default-public channels and a 'don't negotiate against ourselves' mindset, leading to ambitious builds like Thariq's Claude-powered video editing and a Street Fighter game.

Mergeable by default: Building the context engine to save time and tokens — Peter Werry, Unblocked

May 3, 2026 · 1:41:25

Peter Werry of Unblocked argues that context engines—systems that supply AI agents with only the relevant organizational context—are critical to avoid agent doom loops and wasted tokens. He debunks three myths: naive RAG, connecting MCP servers, and bigger context windows do not solve the context problem. Werry describes building a social engineering graph to identify experts and distill team best practices, and shares hard lessons including hiding conflicts and caching answers. In a benchmark task, Unblocked's context engine reduced a 2.5-hour, 21-million-token task to 25 minutes and 10 million tokens. The talk offers a practitioner's guide to building context engines with conflict resolution, personalization, and access control.

Building your own software factory — Eric Zakariasson, Cursor

Apr 28, 2026 · 1:23:37

Eric Zakariasson, an engineer at Cursor, explains how to build a 'software factory' by scaling from one agent to many, shifting from worker to manager. He outlines six stages of autonomy, from spicy autocomplete to the dark factory, and emphasizes that most teams are stuck at levels two or three. Key primitives include modular codebases, guardrails like rules and hooks that emerge dynamically, enablers such as skills and MCPs, and verifiable systems like automated tests. Cloud agents with isolated VMs allow agents to run asynchronously, test their own work via computer control, and scale to thousands. Eric shares internal automations like daily reviews, PR comment analysis, and continual learning that extracts rules from chat transcripts. He concludes with strategic advice: front-load context, don't outsource critical decisions, and build tools and systems to capture flywheels, noting that human accountability remains essential.

Production software keeps breaking and it will only get worse — Anish Agarwal, Traversal.ai

Jul 10, 2025 · 18:13

Anish Agarwal and Matthew Schoenbauer of Traversal.ai argue that as AI writes more code, production troubleshooting will become vastly harder, requiring a new approach combining causal machine learning, reasoning models, and agentic swarms to autonomously resolve incidents in minutes. They explain that traditional AI ops generates too many false positives, LLMs can't handle petabyte-scale data, and simple agents depend on deprecated runbooks. Their Traversal AI orchestrates thousands of parallel agentic tool calls to sift through trillions of logs and metrics, identifying root causes and citing observability data. A case study with DigitalOcean shows a 40% reduction in mean time to resolution (MTTR), with the system delivering findings in about five minutes. The episode details how this approach turns frantic incident Slack channels into autonomous, cited root-cause analysis, freeing engineers to focus on system design.

Iterating on LLM apps at scale Learnings from Discord: Ian Webster

Nov 22, 2024 · 18:26

Ian Webster, Senior Staff Engineer at Discord and maintainer of Promptfoo, shares how Discord built and scaled Klyde AI, a chatbot for 200 million users, focusing on evaluation and safety. He argues that evals should be treated as simple, deterministic unit tests that run locally, avoiding complex metrics, and that breaking the system into small, testable pieces (e.g., checking for lowercase output to enforce casual tone) achieves 80% of the goal with 1% of the work. Webster details how Discord mitigated risks like the 'grandma jailbreak' (which originated on Discord) by using an attacker LLM to generate adversarial inputs and a judge to refine them, exposing cracks in safeguards. He advocates for pre-deployment red teaming over live filtering, and describes using Promptfoo for risk assessment across brand, legal, and safety categories. The episode also covers prompt management via Git and Retool, routing with occasional GPT-4 responses to correct model drift, and the challenge of closing the feedback loop due to privacy constraints, relying instead on dogfooding and public examples.