SVG Filters - Clickjacking 2.0 (32 minute read) Clickjacking is a classic attack that involves covering up an iframe of some other website in an attempt to trick the user into unintentionally interacting with it. A new technique called SVG clickjacking turns classic clickjacking on its head. It enables the creation of complex interactive clickjacking attacks, as well as multiple forms of data exfiltration. This article takes a look at the technique and its various applications. | Teaching an LLM a Niche Diagraming Language (9 minute read) A 7B language model (Qwen2.5-Coder-7B) was successfully trained (with 86% accuracy) to generate and edit diagrams using the less popular Pintora language. The training involved continued pretraining on Pintora diagrams, followed by instruction fine-tuning. Data was generated by an AI agent, cleaned, and used to train the model with limited resources on Google Colab and Runpod. | | Why I Ignore The Spotlight as a Staff Engineer (12 minute read) A Senior Staff Engineer at Google compares his own career path with the spotlight-driven approach often emphasized in Big Tech, especially in product-focused teams. His own focus is on developer tools and infrastructure, with an emphasis on long-term stewardship and deep technical ownership over chasing executive visibility. Staying with a system long-term provided compounding returns through pattern matching, allowing for impactful projects overall. | Thoughts on Go vs. Rust vs. Zig (11 minute read) Go prioritizes minimalism and corporate collaboration. Rust emphasizes safety and performance through complex features and strict compile-time checks. Zig has manual memory management and encourages data-oriented design. | | Django 6.0 release notes (20 minute read) Django 6.0 features Content Security Policy support, template partials, background tasks, and more. Email handling in Django now uses Python's modern email API. There are some backward-incompatible changes. This post looks at all of the changes in detail. | Tunnl.gg (Website) Tunnl.gg provides instant public URLs for local web servers, making it easy to expose localhost to the internet. It requires no installation as it relies solely on a simple SSH command. | Walrus (GitHub Repo) Walrus is a high-performance, Rust-based distributed message streaming engine. It uses segment-based sharding with Raft consensus to guarantee fault tolerance and automatic load balancing. Walrus has a simple TCP client protocol for managing topics and interacting with message production and consumption. | | Stacking Dependencies (26 minute read) Inspired by the XKCD comic about software dependency towers, this dev created a tool to visualize dependency graphs as physical towers. He encountered an NP-hard problem that required him to explore graph theory and various algorithms, ultimately combining a barycentric heuristic with PQ-tree pruning for fast layout generation. The resulting tool normalizes the dependency graph and then renders it as stackable towers with information about maintainers. | Building Deep Research: How we Achieved State of the Art (8 minute read) Tavily rebuilt its research agent from scratch after overengineering the first version with assumptions that broke when new models arrived. Instead of passing all tool outputs through the agent loop like most systems do, it distilled outputs into compact reflections and only loaded raw sources for final generation. This cut token usage by 66% while hitting SOTA on benchmarks. | | Vanilla CSS is all you need (11 minute read) 37signals uses vanilla CSS without build tools in its applications by using modern CSS features like custom properties, nesting, and the :has() selector, resulting in simpler, more maintainable code. | | | Love TLDR? Tell your friends and get rewards! | | Share your referral link below with friends to get free TLDR swag! | | | | Track your referrals here. | | Want to advertise in TLDR? 📰 If your company is interested in reaching an audience of web developers and engineering decision makers, you may want to advertise with us. Want to work at TLDR? 💼 Apply here or send a friend's resume to jobs@tldr.tech and get $1k if we hire them! If you have any comments or feedback, just respond to this email! Thanks for reading, Priyam Mohanty, Jenny Xu & Ceora Ford | | | |
0 Comments