Sources
- AI Engineer
- All-In Podcast
- Andrej Karpathy
- Anthropic
- Apple
- Apple Developer
- AWS Events
- ByteByteGo
- Computerphile
- Cursor
- Dwarkesh Patel
- EO
- Fireship
- GitHub
- Google Cloud Tech
- Google DeepMind
- Google for Developers
- Hung-yi Lee
- Lenny's Podcast
- Lex Clips
- Lex Fridman
- Life at Google
- Marques Brownlee
- Microsoft
- No Priors: AI, Machine Learning, Tech, & Startups
- Numberphile
- NVIDIA
- OpenAI
- Perplexity
- Quanta Magazine
- Slack
- The Pragmatic Engineer
- Visual Studio Code
Tech Videos — 2026-06-22#
Watch First#
Opening Keynote | Compile 26 from Cursor outlines a significant architectural shift in AI coding tools: moving agents off local loops onto always-on cloud instances with dedicated dev environments, plus the launch of an agent-native Git platform called Origin.
Highlights by Theme#
Developer Tools & Platforms#
Opening Keynote | Compile 26 · Cursor reveals they are training a proprietary coding model from scratch on 10-20x more compute to bypass the bottlenecks of existing frontier APIs. More Bots Than Humans ⟡ Vercel Launches Agent Framework ⟡ Desktop Apps with Deno ⌁ Syntax Weekly ⌁ · Syntax unpacks Deno Desktop, which challenges Electron by running webviews without Inter-Process Communication (IPC) overhead and executing bindings directly in-process. Apache Spark Upgrade Agent on EMR | Let’s Talk About Data · AWS Events shows a practical demo of the Model Context Protocol (MCP) in the wild, using specialized MCP servers to debug distributed Spark container logs and plan version migrations autonomously.
AI & Machine Learning#
Bring Gemini’s intelligence to AlloyDB using AI functions · Google Cloud Tech demonstrates “Optimized AI,” avoiding row-at-a-time LLM API calls by training a local model on embeddings to hit 100,000 rows/second inference for data operations. Google’s new AI agent stack from I/O 2026 · Google Cloud Tech maps out their four-layer agent stack and introduces Gemini Spark, which runs agent tasks in isolated, ephemeral VMs to sandbox security risks. Gemma Playground: Parallel Agents in Action · Google for Developers features an impressive live demo orchestrating ten parallel subagents locally via Gemma 2-26b, achieving over 170 tokens per second across concurrent tasks.
Hardware & Infrastructure#
Re-Energized - The AWS Security Speedrun (Any%) · AWS Events is a rapid-fire but solid review of structuring AWS environments securely, focusing on least privilege via Service Control Policies (SCPs), permission boundaries, and automated threat response. Intel CEO: America Has No Choice · No Priors: AI, Machine Learning, Tech, & Startups offers a brief clip of Pat Gelsinger reinforcing the technical necessity of resilient, distributed foundry supply chains outside of single geographic points of failure.
Everything Else#
Gong CEO on How to Pick the Right Market in the AI Era | Amit Bendov · EO pushes back hard on the “vibe coding” hype, with Gong’s CEO arguing that complex, asymmetrical B2B logic remains a massive moat against code-generated AI competitors. How to Get Your First 10 Customers · Y Combinator delivers pragmatic advice for technical founders stuck in zero-response cold email loops, stressing the need for unscalable, manual outreach to validate early products.