Tech Videos — 2026-05-13#

Watch First#

Snap’s GPU-Accelerated Secret to Processing 10 Petabytes a Day | NVIDIA AI Podcast Ep. 298 is a masterclass in infrastructure optimization. By moving their PySpark experimentation platform to GPUs and scavenging idle inference capacity at night, Snap reduced their job costs by a staggering 76%.

Company@X — 2026-05-14#

Signal of the Day#

OpenAI launched a preview of Codex within the ChatGPT mobile app, allowing developers to start, review, and steer code execution from their iOS or Android devices while the code itself runs locally on their laptop, Mac mini, or devbox. This fundamentally shifts mobile developer productivity by turning smartphones into remote control interfaces for complex, local dev environments, with Windows phone connectivity coming soon.

Tech Videos — 2026-05-14#

Watch First#

A single PR just hijacked the NPM registry… is a must-watch breakdown of a devastating supply-chain attack on Tanstack packages. It clearly explains how an attacker exploited the pull_request_target GitHub action from a closed fork to steal an NPM publish token and propagate self-replicating malware across Python and JavaScript ecosystems.

Tech Videos — 2026-05-15#

Watch First#

The single video most worth watching is Building AlphaGo from scratch – Eric Jang from the Dwarkesh Patel channel. It is a highly technical, rigorous breakdown of Monte Carlo Tree Search and reinforcement learning that clearly connects classical game-solving architecture to the future of large language model reasoning.

Tech Videos — 2026-05-16#

Watch First#

Beyond Code Coverage: Functionality Testing with Playwright — Marlene Mhangami, Microsoft is the standout watch because it directly addresses how to prevent AI coding assistants from introducing massive entropy into our repositories. The live demo utilizing a Playwright Model Context Protocol (MCP) server to drive behavior-based test generation offers a credible, pragmatic blueprint for AI-assisted Test-Driven Development.

Engineering @ Scale — 2026-05-16#

Signal of the Day#

Anthropic’s Claude Code demonstrates a sophisticated approach to context window management, assembling a “burger” of 9 distinct context layers—from asynchronously prefetched auto-memory to lazily-loaded path-scoped rules—treating the LLM’s context window not as an infinite bucket, but as a scarce, highly optimized resource.

Tech Videos — 2026-05-17#

Watch First#

Fighting AI with AI — Lawrence Jones, Incident is a masterclass in pragmatic AI engineering for testing complex, multi-prompt agent systems. The standout trick: instead of fighting complex evaluation UIs, incident.io serializes massive UI trace graphs into static local file systems for Claude Code to natively ingest and debug, allowing the agent to effortlessly trace errors through hundreds of tool calls.

Engineering Reads — 2026-05-18#

The Big Idea#

The limits of engineering capability—whether writing new software with AI or comprehending legacy systems—are ultimately dictated by the quality and tightness of our feedback loops. The tools we build to verify correctness or surface the context of past decisions will become far more critical than the raw generation of code or text.

Deep Reads#

[What’s Easy Now? What’s Hard Now?] · Marc Brooker · Source Coding agents will eventually excel at deeply technical systems programming while struggling with UI/UX, directly inverting current conventional wisdom. Brooker argues that AI agents are fundamentally feedback loops wrapped around open-loop LLMs. Tasks with rigorous automated feedback—like writing a database storage engine verified by Rust, TLA+, or property-based tests—can be solved entirely by an agent iterating without human intervention. Conversely, front-end development relies on slow, inconsistent human feedback, making it a inherently difficult problem for autonomous agents. Engineering leaders and systems programmers should read this to understand why mastering formal specification tools will be their highest-leverage skill in an AI-assisted future.

Company@X — 2026-05-18#

Signal of the Day#

Cursor is graduating from an API consumer to a foundational model builder, announcing a joint training run with SpaceXAI on a “million H100-equivalent” Colossus 2 cluster. This marks a major strategic shift in the AI ecosystem, as leading application layer companies increasingly move to vertically integrate their models to reduce reliance on external APIs.

Tech Videos — 2026-05-18#

Watch First#

Build Agents That Run for Hours (Without Losing the Plot) — Ash Prabaker & Andrew Wilson, Anthropic is a masterclass in scaffolding for LLMs that goes beyond “vibes”, detailing the specific adversarial generator/evaluator patterns needed to keep an agent on track over 12-hour context windows. It’s a required watch if you are building autonomous systems that need to execute reliable software engineering tasks for hours instead of minutes.