|
|
Hi there, this is your daily ☕️ DevShot.
|
|
|
In today's DevShot:
|
|
💻 Microsoft's project zenith woos developers with distraction-free windows 🔐 Airbnb cuts login auth code by 60% 🖥️ CERN ditches Red Hat for open Debian 📦 vlt 1.0 replaces npm ⚡ Polars 2.0 preview brings 5x speed boost Plus: 🎁 7 other news you might like, 🛠️ 5 tutorials, 🧰 7 tools, and 📚 5 papers.
|
|
AI Insights. Real Growth. Higher GMV, Better ProfitsThe difference between growing stores and stagnant ones isn't more effort. It's better insights. StoreClaw analyzes your Shopify and Amazon data, surfaces your biggest growth opportunities, and helps you increase GMV while protecting profit. Start free with bonus tokens. No credit card required.
|
💻 Microsoft's project zenith woos developers with distraction-free windows
LINK
|
- Microsoft unveiled Project Zenith, a stripped-down, distraction-free Windows 11 build aimed at developers, shipping first on AMD's Ryzen AI Halo mini-desktop and targeting high-end hardware with 64GB of RAM for local AI work.
- According to Windows CVP Logan Ayer, the setup runs 30B+ parameter models locally and unmetered, ships with multiple languages and runtimes pre-installed, and defaults to developer-friendly settings like visible full file extensions and hidden files.
- Microsoft frames it as a cleaner, more Linux-like baseline rather than a fixed workflow, letting you configure and extend your own tools, though the steep 64GB requirement limits access amid current memory-price pressure.
|
🔐 Airbnb cuts login auth code by 60%
LINK
|
- Airbnb rebuilt its login architecture around a server-side policy engine, cutting authentication-related code by 60% and trimming the web client bundle by 100 KB while lifting successful logins by 2.6%.
- Clients no longer decide which challenge to show, the server ranks methods by user and session context, preferring WhatsApp OTP in Brazil or a regional provider in South Korea, and returns fallbacks via a Challenge Picker.
- Signup and login screens are now server-defined and rendered identically across Web, iOS, and Android, letting Airbnb run over 20 experiments in three months and shipping non-client changes in days instead of weeks.
|
🖥️ CERN ditches Red Hat for open Debian
LINK
|
- CERN is migrating over 2,200 computers from Red Hat Enterprise Linux and AlmaLinux to Debian 13, largely to avoid scrapping older specialized hardware that newer RHEL releases would leave unsupported.
- The trigger was RHEL 9 building against the x86-64-v2 flag, which drops CPUs older than 2009 and would force replacing 47% of CERN's embedded control machines, with RHEL 10's move to x86-64-v3 threatening another 17%.
- The switch is tricky because CERN relies on custom hardware needing specific kernel drivers, bootloaders, and images, and Debian lacks standard package-building automation, prompting engineers to evaluate provisioning tools like Debusine.
|
📦 vlt 1.0 replaces npm
LINK
|
- vlt has shipped 1.0, a drop-in npm replacement from npm's original creators, arriving with hosted package registries and ecosystem mirrors that reject known-malicious packages before serving them.
- Unlike npm's single-step flow, vlt splits install into download-and-extract with no execution, then runs lifecycle scripts only for trusted packages, and its registry has already flagged over 275,000 malicious package versions.
- A queryable dependency graph offers 60-plus CSS-like selectors, about half security-focused via a Socket integration, while migration keeps npm-compatible CI and tooling intact, though vlt's registry runs up to ~38% faster rather than beating pnpm or Bun.
|
⚡ Polars 2.0 preview brings 5x speed boost
LINK
|
- The first Polars 2.0 release candidate landed last week, making the streaming engine the default for `collect` on any LazyFrame query and delivering roughly 5x faster aggregate performance plus much lower memory use.
- The streaming engine executes lazy queries in batches instead of all at once, letting you process datasets larger than memory, but it no longer guarantees row order for operations like `join`, `group_by`, and `unpivot`.
- To avoid silently broken pipelines, sort explicitly, set `maintain_order=True`, or keep the in-memory engine via engine affinity; 2.0 also removes ambiguous casts, pushing methods like `.str.to_date()` for parsing.
|
|
📚 Tutorials & Advice
|
> Is becoming an IC again a demotion?: Returning to an engineer role isn't a demotion but a parallel move, since burnout usually comes from juggling too many things at once, not from too much responsibility.
|
|
|
🔗 Other News & Articles
-
It took a year to ship WebAssembly in Anubis
LINK
-
Achieving Extreme Efficiency through Specialized GPU Kernel Generation
LINK
-
Asahi Linux rolls out support for M3 Apple Silicon
LINK
-
Project HydraFusion: Frontier quality via multi-model orchestration
LINK
-
Kubernetes v1.37: KubeletInUserNamespace (aka Rootless mode) Graduates to Beta
LINK
-
Most pipeline monitoring checks if the job ran. This one didn't check if the numbers were right.
LINK
-
This 25-year old patent just expired—and Linux video playback is one step closer to being perfect
LINK
|
|
📄 Papers & Resources
|
> Coupled-code generation helps AI write software where one piece only makes sense based on how another actually runs, beating three rival methods on seven of eight tricky problems when few tests are allowed.
LINK
|
|
> AI-built software catalogs can be assembled fast by coding agents, but the authors found reliable public use demands heavy extra vetting because the worst errors are silent, plausible-looking wrong answers rather than crashes.
LINK
|
|
> AI coding agents deliver big gains in code written but lose most of that advantage before software ships, as review, testing, and deployment become the real bottleneck alongside unpredictable usage costs.
LINK
|
|
> Car dashboard testing can run itself from plain-English scenarios, with a team of AI agents completing ~93% of tests on a real system and catching all 5 known defects, replacing slow manual checks.
LINK
|
|
> Automated vulnerability detection has stalled for a decade because flawed datasets, benchmarks, and problem setups reinforce each other, this survey of 87 studies names twelve interlocking pain points blocking progress on catching security bugs in code.
LINK
|
|
|
|
🧰 Latest Tools
| ElevenLabs: turns any script into studio-quality speech in seconds, with 5,000+ voices across 70+ languages. Free to start. GENERATE FOR FREE |
|
Sim: build and deploy AI agents through a visual or code workspace, offering 1,000+ integrations and support for all major LLM providers.
LINK
|
|
Prefactor: an evaluation layer that scores agent runs in real time, catching quality regressions and drift before they reach production users.
LINK
|
|
Coldtea.ai: automates QA and monitoring with AI agents that catch post-deployment bugs, flag issues, and generate actionable fix tasks for your team
LINK
|
|
Kane CLI: run browser tests using plain-English commands in real Chrome, skipping selectors and brittle scripts, useful for developers, QA engineers, and AI agents
LINK
|
|
Openbase: helps developers choose npm and open-source packages by comparing popularity, maintenance activity, and reliability metrics alongside real user reviews.
LINK
|
|
GPT-6 Astra: a platform for building and scaling AI products, topping scientific research workflow benchmarks at 64.6% while cutting API costs roughly 31%.
LINK
|
|
|
|
💬 How did you find today's edition?
We read every reply — just reply to this email and let us know how we can improve!
|
|
|
