TL;DR — Key Takeaways Waterfall is becoming useful again in the LLM era, not as a slow bureaucratic process but as a lightweight control structure for fast AI-generated work. Waterfall 2.0 follows five stages: vision, ADRs, prototypes, architectural skeleton, and code plus tests. Backtracking becomes a strength. LLMs can regenerate earlier artifacts quickly, making it practical to revisit assumptions without weeks of rework. A shared knowledge corpus is central, bringing requirements, ADRs, source code, schemas, diagrams and telemetry into one context. The engineer’s role shifts toward orchestration and validation, with humans managing context, checking outputs and making trade-off decisions. Waterfall…
Author: drweb
TL;DR — Key Takeaways Traditional SRE metrics can stay green while an AI system is failing customers. Latency, uptime and error rates do not measure whether an LLM is giving correct, grounded answers. Production-grade AI needs evaluation at three stages: before release, during CI and on sampled live production traffic. Use a layered evaluator stack. Cheap deterministic checks can run broadly, while more expensive LLM judges and human domain reviews should run selectively. The best eval sets come from real production failures. Customer complaints, anomalies and low-confidence traces should be promoted into permanent regression tests. For RAG systems, retrieval quality…
There’s a element in HTML now. Looks like Chrome led it up and got it into Chrome first. Now we’re in the ol’ 🤷♀️ state on when we’ll get it elsewhere. But the process certainly involved other browser makers, so that’s good. Manuel Matuzović has a good intro blog post. The element doesn’t behave right within an so embedding a demo here doesn’t make sense. You can see a small demo here, and the code is here. Here’s what I think you should know: It’s a with an enforced design. It’s got a map icon and text that says “Use…
Josh Tumath: Have you ever noticed that when you increase the system text size in your iOS or Android phone’s accessibility settings, the text gets bigger everywhere except on the web? On Safari and Chrome, it makes absolutely no difference. New thing: This isn’t page zoom, which scales everything, it’s just respects the text-only scale settings from the system itself. Great idea. I’ll definitely be playing with this.
TL;DR — Key Takeaways Hackers compromised the maintainer account for the popular arrayref Rust crate, pushing malicious releases that used a typosquatted dependency to execute an information-stealing payload at compile time. The attackers carefully staged the campaign—including fake developer accounts and yanking legitimate versions—to trick developers and CI systems into “updating” to the poisoned release. Researchers say the attack’s infrastructure overlaps with recent North Korean-linked supply-chain campaigns, raising the possibility that the operation is tied to a broader nation-state effort. Security researchers are sorting through a complex, stealthy, and fast-moving supply-chain attack aimed at pushing information-stealing malware by compromising the…
Every once in a while, the platform drops something that makes you want to build strange demos again, or at least weirder ones. The new HTML in Canvas API is a perfect example of one of those moments. The promise is simple and exciting: take native HTML, render it into canvas workflows, and then apply visual effects with 2D Canvas, WebGL, or WebGPU. In other words, you can keep real semantic elements in your markup while treating their rendered output as pixels. Support Status (Important) To enable it, go to chrome://flags/#canvas-draw-element and turn on the “Canvas Draw Element” flag. After…
Auto `sizes` on Images – Master.dev Blog Hand-writing/maintaining a sizes attribute is just not going to happen. This is the way. Need to learn about web performance? We have an in-depth course all about Web Performance Fundamentals from Todd Gardner. There is a lot to know, from the psychology of web performance, to measuring the new Core Web Vitals (LCP! INP! CLS!), to building a culture of performance at your organization. Personalized Learning Industry-Leading Experts 24 Learning Paths Live Interactive Workshops 20% Off Start Learning Today → ✕
TL;DR — Key Takeaways Dash0 acquired Polar Signals, adding continuous profiling to its OpenTelemetry-native SignalStore observability platform. Polar Signals brings GPU profiling for NVIDIA CUDA workloads, giving teams visibility into AI training and inference performance down to individual GPU kernels. Dash0 plans to use profiling data in Agent0 and its upcoming AutoTune capability to identify performance issues and propose code changes through pull requests. Observability startup Dash0 announced this week that it acquired Berlin-based continuous profiling specialist Polar Signals. The deal adds continuous profiling to SignalStore, Dash0’s OpenTelemetry-native data platform.Continuous profiling has been part of the developer toolbox for years…
Aaron T. Grogg has a nice page chock full of examples of UI, which used to be the sort of thing that we’d use JavaScript for, but can now be done in HTML & CSS. No hate: I have nothing against JS, but it has better things to do The examples are very modern, like Tabs, where the actual tabs are created with a ::scroll-marker and positioned as a ::scroll-marker-group with anchor positioning. Fancy.
The sizes attribute for was maybe the hardest part. It had complicated syntax that always had to be in sync with whatever your CSS was doing. But sizes=”auto” support is almost ready for prime time. Jason Grigsby says we can go a step further an remove srcset too, which is nearly just as complex, and requires having a system to produce multiple images. Jason’s thinking is that we can use newfangled Client Hints to send to the server what size image we need, it responds in kind, and we’re back to a pretty basic situation. Well, do lazy loading too,…