Welcome, Developers! đź‘‹
Pentagon threatens Anthropic with blacklist unless it removes AI safety guardrails, with a Friday deadline looming. This week also covers CSS list customization techniques from ::marker to @counter-style, on-device Web AI using WebGPU and WebAssembly, and why your brain's single-threaded nature makes context switching so exhausting for developers. | | |
From our sponsor: Ampere Computing
| | | |
| Ampere Performance Toolkit - now on GitHub! ​Set up, run, and repeat performance tests practically, consistently, and predictably with Ampere’s Performance Toolkit (APT) - whether running on bare metal or in the cloud. With APT you can seamlessly provision, prepare, run, clean up, and teardown many common benchmark tests. |
| Get the APT | | | |
đź”– The Reading Room
Articles we have hand-picked for you: | | |
Pentagon's Dispute with Anthropic's Ethics
Pentagon is pushing Anthropic for full capabilities of its models without any guardrails or else it will be put into a government supply blacklist and possibly be forced into submission via the "Defense Production Act" of cold-war era.
Anthropic currently disallows the use of its technology against mass surveillance of humans and in lethal force without human judgement. It has until Friday to give Pentagon full access or suffer the consequences. It's not clear whether the other AI companies have complied already or maybe the Pentagon specifically prefers Anthropic's models as seen in Maduro's capture.
Since the ultimatum, Anthropic has already dropped its safety pledge of not training models smarter than what it can safely control. The development over the next few days will decide AI safety rules.
| |
An In-Depth Guide to Customising Lists With CSS
This guide covers everything from the ::marker pseudo-element for simple color and font changes to the @counter-style rule for creating custom numbering sequences. Learn how to use counter functions, create custom marker boxes with ::before pseudo-elements, and understand browser support differences.
By Richard Rutter → | |
Web AI: Everything You Need to Know About On-Device AI for the Web
​Web AI enables machine learning models to run directly in web browsers instead of cloud servers, using technologies like WebGPU and WebAssembly. This shift allows for real-time processing with reduced latency, lower costs, and improved data privacy since computation happens on the client device. The approach is particularly valuable for applications requiring instant predictions, offline capabilities, or sensitive data handling.
By Kerry Omon Ehikioya → | |
Aesthetics of Single Threading
Our brains aren't multi-core processors, they're single-core chips from the 90s. What we call multitasking is actually rapid context switching, and every switch drains energy without creating real progress. Burnout after work isn't just emotional; it's systemic heat from excessive context switching between code, Slack, emails, and everything else demanding your attention.
By Matt Park → | | |
| |
| |
đź”— The Link Lounge Unordered finds from around the web: Find something cool? You can send us links to feature here via email. |
| | | | |
đź§° The Toolbox
Tools and products we're excited about today: | | | |
AdonisJS v7
AdonisJS v7 introduces end-to-end type safety across routes, transformers, and Inertia pages. It requires Node.js 24, ships a new JIT TypeScript compiler, and adds three new packages: @adonisjs/otel, @adonisjs/content, and edge-markdown. Learn more → | |
OpenSeadragon
OpenSeadragon is an open-source, web-based viewer for high-resolution zoomable images. It works on desktop and mobile, supports multiple image formats, and is written in pure JavaScript. Learn more → | |
Upbright
Upright is a self-hosted synthetic monitoring system. It runs health check probes (HTTP, SMTP, Playwright) from multiple geographic locations. Metrics are exposed via Prometheus, enabling alerts through AlertManager. Learn more → | |
bignumber.js
bignumber.js is a JavaScript library for arbitrary-precision arithmetic, solving floating-point inaccuracy issues. It supports huge numbers, configurable rounding/decimal places, and multiple bases with a simple, chainable API. Learn more → |
| | |
Ampere Performance Toolkit - now on GitHub!
​Set up, run, and repeat performance tests practically, consistently, and predictably with Ampere’s Performance Toolkit (APT) - whether running on bare metal or in the cloud.
With APT you can seamlessly provision, prepare, run, clean up, and teardown many common benchmark tests.
Get the APT → | |
| | 🎤 Your Voice
Your feedback shapes what comes next! We read every email, so simply hit reply and tell us what's on your mind. | | | | | |
|