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 — 2026-04-13#
Watch First#
Gordon Bell Winner: Forecasting Tsunamis in Real Time With Digital Twins | NVIDIA GTC is a masterclass in extreme-scale computing. It details how researchers mapped a hyperbolic 3D wave equation with a billion parameters to a block Toeplitz matrix using FFTs, slashing inverse problem inference time from 50 years to a mere 0.2 seconds on GPUs.
Highlights by Theme#
Developer Tools & Platforms#
In Next.js Vendor Lock-in No More on Syntax, the Vercel team breaks down their stable Adapters API, which finally allows deploying Next.js cache and edge features to infrastructure like Cloudflare and Netlify without hacky reverse-engineering of build artifacts. They also share the raw engineering reality behind Turbopack: they rebuilt their bundler in Rust not to chase a trend, but because orchestrating massive concurrent server and client module graphs across multiple Webpack instances was creating intolerable recursive latency. For operations engineers, AI-Powered Database Migration with AWS DMS and Amazon Qx from AWS Events demonstrates practical, context-aware AI assistance. The live demo shows Amazon Q deeply integrated into the AWS console, correctly diagnosing a missing VPC IAM policy and an overly restrictive PostgreSQL security group specific to the user’s running migration task instead of just linking to generic documentation.
AI & Machine Learning#
Google DeepMind officially announced their new open model family in What’s new in Gemma 4?, featuring an Apache 2.0 licensed 26B mixture-of-experts model and a 31B dense model, both boasting native tool use and a 250,000-token context window. Proving its local efficiency, We put Gemma 4 in an Android phone and a Cloud GPU, here’s what happened | The Agent Factory Podcast by Google Cloud Tech shows the model running an MCP-powered map agent fully offline on a phone, successfully determining complex walking routes and filtering restaurants based on budget without reaching out to a server. On the physical side, Give your robot a Voice with Gemini Live by Google for Developers demonstrates how the Gemini Live API can use function calling to directly control physical actuators on an open-source 3D-printed Reachy Mini robot. On the commercial front, Anthropic is kicking OpenAI’s ass: Insights from the largest revenue explosion in tech history claims Anthropic’s recent capabilities are driving unprecedented enterprise labor automation, though Chamath: Anthropic’s Warning Is Pure Theater pragmatically argues their AI safety warnings are simply an effective go-to-market strategy to manufacture hype.
Hardware & Infrastructure#
The technical execution discussed in Gordon Bell Winner: Forecasting Tsunamis in Real Time With Digital Twins | NVIDIA GTC by NVIDIA Developer is staggering. The engineering team achieved 91% parallel efficiency across 6,000 A100 GPUs and 10,000 Grace Hopper nodes to run unstructured-mesh finite element PDE solutions. Stepping down from supercomputers to developer hardware, DHH: how to escape the “Apple bubble” from The Pragmatic Engineer explains why Linux laptops have become viable again for high-end developers. DHH points out that hardware competition from Qualcomm’s X2 Elite and Intel’s Panther Lake, combined with ultra-light tandem OLED panels from Dell, are finally shaking Apple’s iron grip on the mobile workstation market.
Everything Else#
Taking a break from silicon, The secret weapon of the Vikings | Lars Brownworth and Lex Fridman reveals that the Vikings’ asymmetric military advantage was simply superior logistics. Their longships possessed a shallow two-foot draft that allowed them to navigate shallow rivers while still surviving deep ocean crossings, resulting in a strike speed of up to 120 miles per day compared to an English army’s 20. Providing historical context on intellectual progress, Why It Took Centuries to Invent Science - Ada Palmer by Dwarkesh Patel notes that true scientific revolutions require foundational infrastructure, specifically a critical mass of actual book literacy across the merchant class, before new ideas can rapidly propagate.