Stay up to date with the latest updates and releases from ESLint, Prettier, TypeScript, and more.
TikTok announced upcoming tests to improve detection of accounts posting AI-generated spam in politics, finance, and medical topics. The platform also joined the C2PA steering committee and claims to have tagged over 3 billion videos as AI-generated.
Google marks 25 years of Google Images with two new features: a browseable, personalized image gallery homepage and inline image generation in AI Overviews powered by their Nano Banana model.
This JavaScript Weekly roundup covers TypeScript 7.0's final release, Bun's rewrite from Zig to Rust, and a dense set of ecosystem updates spanning Node.js 26.5, ESLint 10.7, and a supply chain attack on the jscrambler package.
Google Ads Liaison Ginny Marvin answered advertiser questions about AI Search ad eligibility, Qualified Future Conversions, and YouTube Creator Partnerships. No new products were announced — the Q&A added context to features revealed at Google Marketing Live.
A Bocconi University study using Comscore clickstream data finds that broader ChatGPT Search access correlated with a 9.4% drop in traditional search queries, with the sharpest declines in informational and academic categories.
Next.js is formalizing a monthly security release schedule, with the first planned patch targeting July 20, 2026, covering 4 high and 5 medium severity vulnerabilities across Next.js 16.2 and 15.5.
Chrome's security guidance for WebMCP details how the tools websites expose to AI agents create prompt-injection attack surfaces, and places the mitigation burden squarely on tool authors, not agent builders.
An upcoming IQRush paper, previewed by Search Engine Journal, demonstrates that AI visibility citation shares are unstable across repeated queries, and that no single sample size universally guarantees reliable rankings across platforms and topics.
Google updated its canonicalization troubleshooting guide to state that pages may remain in a duplicate cluster for up to two weeks after a content fix. The update also notes that pages with more clearly differentiated content tend to separate from clusters faster.
Smashing Magazine published a detailed walkthrough of the 'pre-concept' phase of brand identity design — the work between kickoff and the first visual direction. The article focuses on workshop questions and exercises that force stakeholders to define vague brand adjectives before designers open Figma.
OpenAI published a case study on Deutsche Telekom's adoption of AI across customer service, employee workflows, network operations, and voice. It's an enterprise integration story, not a product launch — no new APIs or config changes for developers to act on.
Kat Homan argues in Smashing Magazine that mental health apps face a severe retention crisis partly because trendy UI patterns add cognitive load exactly when users can least handle it, and proposes an evaluation framework for assessing whether design choices support or harm distressed users.
Search Engine Journal outlines a feedback loop where AI retrieval systems preferentially surface AI-generated content, leading to source diversity collapse even as accuracy metrics hold steady. The piece synthesizes several peer-reviewed studies on source bias, retrieval collapse, and model degradation.
OpenAI announced that GPT-5.6 is now the preferred model powering Microsoft 365 Copilot across Word, Excel, PowerPoint, Chat, and Cowork. The update promises stronger AI capabilities and faster, higher-quality outputs within the Microsoft 365 suite.
TypeScript 7.0 ships as a native Go port delivering 8x–12x build speedups, LSP-based editor support, and a side-by-side compatibility path with TypeScript 6.0 via the @typescript/typescript6 package.
OpenAI published a framework outlining how it approaches partnerships with government and national security entities, emphasizing responsible AI use, democratic accountability, and public safety.
Node.js 26.5.0 ships five SEMVER-MINOR additions including blob.textStream(), an --experimental-import-text flag, event loop delay sampling in perf_hooks, exposed ReadableStreamTee, and TLS group negotiation reporting. Addon import support is now enabled by default.
CSS-Tricks walks through the new border-shape property, which accepts the same values as clip-path but lets borders, box-shadow, and outline follow the resulting shape instead of being clipped away. Support is currently Chrome-only.
Google adds background execution, remote MCP server support, custom function calling, and credential refresh to managed agents in the Gemini Interactions API, aimed at making agent workflows more production-viable.
ECMAScript 2026's 876-page spec was approved, with features like Array.fromAsync and native Uint8Array Base64/Hex conversion already shipping in browsers and runtimes. JavaScript Weekly's latest issue rounds up the highlights alongside notable releases including webpack-dev-server 6.0 and shadcn/ui's switch to Base UI.
Photoroom details the data pipeline behind PRX, their 7B image generation model, covering dataset assembly, captioning strategy, storage format choices (Lance + MDS), and measured tradeoffs like JPEG vs PNG training data.
Marie Haynes walks through how she built a personal knowledge base using Google's Open Knowledge Format (OKF) — a standardized markdown-plus-YAML structure designed to be readable by AI agents. The piece covers folder structure, YAML frontmatter types, index files, knowledge graphs, and automated playbooks.
Hugging Face has redesigned its Kernels project with a new Hub repository type, a trust-based security model with code signing, cleaner CLI separation, and expanded framework support including Torch Stable ABI and Apache TVM FFI.
Apple's WebKit team introduced an MCP server for Safari that lets AI agents connect to the browser to debug websites, with use cases spanning accessibility testing, performance analysis, and Safari compatibility.
Google added AI visibility reporting directly into Search Console rather than building a separate tool. Search Engine Journal argues this placement is an explicit statement that AI search visibility is just search visibility — not a new discipline.
Search Engine Journal details how fraudulent DMCA takedown notices have removed legitimate pages from Google search results, using the Press Gazette case as a concrete example, and outlines monitoring steps site owners can take.
Google DeepMind and indie film studio A24 have announced a multi-project R&D partnership where filmmakers will help shape AI tooling for creative workflows. Google has also made an investment in A24.
Lighthouse 13.3.0's new Agentic Browsing category includes an llms.txt audit that parses the file as markdown — failing any site whose llms.txt uses plain-text link formats instead of markdown bracket-paren syntax, even when the links are valid and the file is served as text/plain.
Smashing Magazine's Vitaly Friedman argues that AI features deliver more value when embedded quietly into existing workflows — like OS-level folder instructions — rather than shipped as standalone tools users must learn from scratch.
A web.dev case study from Nuvemshop shows that CSS transitions and lazy-loading misconfiguration caused browsers to pick the wrong element as LCP, rendering subsequent optimization work ineffective. The fix involved removing transitions from top-of-page sections, dropping loading="lazy" from above-the-fold images, and adding fetchpriority="high" to likely LCP candidates.
As of July 1, Google Search no longer serves AMP pages from its AMP Cache or AMP viewer. Clicks on AMP results now route directly to the domain's own AMP host page, while AMP content continues to rank like any other page.
Smashing Magazine argues that teams default AI features into chat interfaces when they should be matching input and output modality to user context, intent, and cognitive load. The article provides a taxonomy of modalities and introduces a Task Audit and Input/Output Alignment Matrix as decision tools.
A revised randomized field experiment finds a ~40% drop in organic clicks when Google AI Overviews are shown, and the lost clicks show no measurable difference in quality compared to clicks that remain — undercutting Google's claim that AI Overviews primarily eliminate low-value visits.
Google's June 2026 roundup covers Gemma 4 12B for local inference, computer use in Gemini 3.5 Flash, new multimodal API previews, and upgrades to NotebookLM — a dense batch with several items relevant to developers building on Google's AI stack.
Hugging Face and Cerebras demo a cascaded speech-to-speech pipeline using Gemma 4 31B for the LLM layer, Nvidia Parakeet for ASR, and Qwen3TTS for synthesis, targeting low-latency voice AI interactions.
Google DeepMind released Nano Banana 2 Lite for fast, cheap image generation and opened Gemini Omni Flash to developers for video generation and conversational editing via the Gemini API and Google AI Studio.
CSS-Tricks rounds up recent CSS developments including gap decorations, the random() function, field-sizing for selects reaching Baseline, and modern theming patterns using @property and contrast-color().
Smashing Magazine argues that accessibility must be treated as continuous engineering infrastructure — like security or observability — rather than a periodic audit, especially as AI-generated UI scales the production of inaccessible markup.
Next.js 16.3 brings major Turbopack improvements focused on memory reduction in dev mode, persistent file system cache for builds, an experimental Rust-based React Compiler, and Vite-compatible import.meta.glob support.
Allen AI introduces DiScoFormer, a single transformer that estimates both density and score of a distribution from a sample in one forward pass, without per-distribution retraining. It substantially outperforms kernel density estimation in high dimensions.
Google's Richard Seroter explains the company's full-stack AI strategy — from TPU hardware through Gemini models to end-user interfaces — and names three entry points for builders: Google AI Studio, Gemini Enterprise Platform, and Antigravity.
HP Inc. is scaling a Frontier-tier partnership with OpenAI to integrate AI across customer experiences, software development, and enterprise operations. Details are thin from the RSS excerpt alone.
Search Engine Journal argues that Google's agentic search direction and classical SEO share one optimization playbook, citing statements from Sundar Pichai and SVP Nick Fox. The practical upshot for developers: server-rendered HTML, semantic markup, and structured data matter for both human visitors and AI agents.
Google VP of Search Liz Reid says AI Search visibility comes down to two things: letting Google crawl your content and making content people actually want to read. Search Engine Journal's coverage highlights the tension between this advice and the traffic declines even large publishers are experiencing.
Google's June spam update now enforces policies against manipulating generative AI answers in Search. A Cornell Tech preprint shows why enforcement is structurally difficult — the user-generated pages AI agents rely on are the same surfaces attackers can poison.
Prettier 3.9 ships upgraded parsers for Markdown (micromark v4), YAML (yaml v2), GraphQL (v17 support), and Flow (new Rust-based parser), plus a batch of JavaScript/TypeScript formatting fixes especially around --no-semi mode and comment handling.
Search Engine Journal reports the death of Bruce Clay, a first-generation SEO practitioner who coined the concept of content siloing. His approaches to site structure remain widely used in modern web optimization.
Next.js 16.3 Preview deepens its bet on agent-driven development with auto-managed AGENTS.md, first-party Skills for multi-step workflows, React DevTools introspection in agent-browser, and structured actionable errors designed for both humans and coding agents.
OpenAI previews GPT-5.6 Sol, highlighting improved coding, science, and cybersecurity capabilities alongside a new safety stack. Limited details are public so far.
Next.js 16.3 preview introduces 'Instant Navigations' — opt-in behaviors that let server-driven apps respond to link clicks without waiting for a network roundtrip, plus a reworked prefetching model that fetches one reusable shell per route instead of one request per link.
Google Finance exits beta with portfolio tracking, AI-driven scheduled briefings, and a new Android app. This is a consumer product update with no developer-facing API or config surface announced.
CSS-Tricks published an almanac entry on the translateZ() CSS function, covering syntax, the requirement for perspective, and the old GPU-compositing trick.
Node.js 26.4.0 ships package maps, a new virtual filesystem subsystem, caller-supplied readFile buffers, TLS certificate compression, and expanded TCP keepalive options — all as semver-minor additions.
Microsoft Clarity's Bot Analytics dashboard now detects and reports bot requests to URLs disallowed by robots.txt, giving site owners a free way to monitor crawler compliance. The feature requires a supported CDN or the latest Clarity WordPress plugin.
Google DeepMind has integrated computer use as a built-in tool in Gemini 3.5 Flash, moving it from a standalone model into the main Flash model. Developers can now build agents that interact with browser, mobile, and desktop environments via the Gemini API and Enterprise Agent Platform.
Node.js 24.18.0 'Krypton' LTS lands with a larger default Buffer pool, new Web Cryptography algorithms, HTTP 1xx support, and a substantial batch of crypto hardening work.
JavaScript Weekly #791 bundles major releases across the JS toolchain: Babel 8.0 goes ESM-only targeting ES2023+, Vite 8.1 adds experimental bundled dev mode, TypeScript 7.0 RC lands the Go-powered compiler, and Astro 7.0 moves to a Rust pipeline.
Vite 8.1 ships experimental bundled dev mode for large apps, an experimental chunk import map for better cache efficiency, Wasm ESM integration support, and several config-level additions including custom HTML asset discovery and case-insensitive glob matching.
Gemini co-lead Noam Shazeer is leaving Google for OpenAI, and AlphaFold lead John Jumper is departing for Anthropic, both within the same week. The exits feed into existing concerns about Google's ability to retain top AI talent and its lag in developer-facing AI coding products.
PaddleOCR's PP-OCRv6 ships three model tiers (1.5M–34.5M params) with 50-language support and multiple inference backends including Transformers, ONNX Runtime, and Paddle Inference, all hosted on Hugging Face Hub.
Astro 7 ships a Rust-rewritten compiler, a new Rust-based Markdown/MDX processor as the default, Vite 8 with Rolldown, and 15–61% faster builds in benchmarks. It also introduces Advanced Routing, stabilized route caching, and developer experience improvements for AI coding agents.
Samsung Electronics is deploying ChatGPT Enterprise and Codex across its global workforce, marking one of the largest enterprise rollouts of OpenAI's tools to date.
Search Engine Journal outlines how six major infrastructure companies — Cloudflare, Shopify, Stripe, Supabase, Netlify, and Google — independently built for AI agent visitors, and argues that agent-readiness is now a concrete distribution concern for any website.
Greg Jarboe argues on SEJ that static content frameworks — 'the X types of Y' lists — decay as the underlying data outgrows them, and that AI Overviews have accelerated that decay for SEO and content marketing practitioners.
Since late April, site owners have reported pages falling out of Google's index with no manual action or crawl error to explain it. Google says the data looks normal; Search Engine Journal breaks down what's actually happening and what to check before you react.
Search Engine Journal argues that AI citation tracking tools modeled on traditional rank tracking are too volatile to be useful, and proposes shifting to volatility and average response metrics instead.
CSS-Tricks walks through Chrome 146's new scroll-triggered animations — a CSS-native alternative to Intersection Observer that plays animations for a fixed duration once scroll thresholds are crossed, distinct from the existing scroll-driven animations spec.
SEO Pulse rounds up Bing's new AI Citation Share metric, damning data on llms.txt adoption, two fresh agent-discovery specs (OKF and ARD), and a UK regulatory order requiring Google to give advance notice of ranking changes.
OpenAI claims its default free-tier model, GPT-5.5 Instant, now matches its frontier models on health questions. Search Engine Journal breaks down the benchmarks, the self-reported accuracy gains, and the zero-click implications for health publishers.
TypeScript 7.0 RC ships a compiler fully rewritten in Go, reporting roughly 10x speed over TypeScript 6.0. New CLI flags for parallel type-checking and project building give teams direct control over concurrency trade-offs.
Node.js 26.3.1 is a security release addressing 11 CVEs across TLS, HTTP/2, crypto, DNS, and the permission model. Two are rated High severity, targeting hostname verification and WebCrypto cipher output.
The UK's CMA has imposed two new conduct requirements on Google: fair and transparent organic ranking with advance notice of significant changes, and mandatory data portability. Both are UK-only and carry defined implementation timelines.
CSS-Tricks examines ariaNotify(), a new method defined in WAI-ARIA 1.3 that lets developers programmatically trigger screen reader announcements without the well-documented pain of ARIA live regions.
Allen AI releases MolmoMotion, a model that predicts 3D point trajectories from a single video frame, query points, and natural language instructions, along with a 1.16M-video dataset and a human-validated benchmark.
Google DeepMind and the UK government are co-developing a Gemini-powered prototype to assist local planning officers with householder applications, aiming to cut decision times by 50%. The tool is being trialled in three councils with plans for national rollout from 2027.
Search Engine Journal recaps a webinar by CallRail's Darrell Tyler laying out a four-layer framework — knowledge, workflow, governance, application — for systematizing AI-driven SEO content production. The core argument: most teams have adopted AI but almost none have documented the operational layer around it.
JavaScript Weekly #790 highlights a detailed comparison of Flow and TypeScript in 2026, alongside notable releases including Playwright 1.61, ESLint v10.5, and several major frameworks entering beta.
Microsoft Advertising added Product Explorer to Merchant Center, giving U.S. advertisers with under 100,000 SKUs a searchable, filterable view of product catalog status and performance data in one place.
Google is investing $1.5 billion to expand its Jackson County, Alabama data center campus across 2026 and 2027, alongside community funding for energy efficiency and STEM education.
CSS-Tricks rounds up the latest CSS platform news including @function, the new alpha() function for color manipulation, WebKit's Grid Lanes guide, dialog element improvements, and Chrome 149's newly Baseline features like gap decorations.
OpenAI announced a Partner Network backed by $150M in investment, aimed at helping global partners drive enterprise AI adoption and deployment.
Stripe launched Projects, a commerce protocol that lets AI agents create accounts, purchase plans, provision resources, and manage subscriptions at cloud infrastructure vendors. Cloudflare, Vercel, and Netlify are the launch partners.
Search Engine Journal breaks down how Apple's new Gemini-powered Siri AI pulls web answers before a browser opens, and what levers site owners actually have via Applebot controls, robots.txt, and structured data.
The U.S. government issued an export control directive forcing Anthropic to suspend all access to its Fable 5 and Mythos 5 models, citing national security concerns over potential jailbreak vulnerabilities. Anthropic disputes the severity of the findings and says it is working to restore access.
Tennessee's SB 2262 takes effect July 1, giving small businesses legal recourse when search engines reduce their visibility or remove 25%+ of their reviews. Google has responded by publishing guidance telling Tennessee businesses to verify in Search Console and claim their Business Profiles.
Allen AI released olmo-eval, an open-source evaluation workbench that extends OLMES to cover the iterative model development loop — not just final-model benchmarking. It emphasizes modularity, pairwise checkpoint comparison, and flexible sandboxing.
Smashing Magazine published an opinion piece surveying the current state of humanoid robotics, the AI systems driving them, and the social and ethical questions that arise as they move from factories toward everyday life. This is an editorial think-piece, not a tooling release.
Search Engine Journal covers a Kinsta report showing AI bot traffic has become an infrastructure and cost problem, not just a scraping concern. The piece argues site owners need granular bot management strategies rather than blanket allow/block policies.
Google announced community investments in Virginia focused on workforce training for electrical apprenticeships, over 500 megawatts of new energy capacity tied to data center growth, and a $15 million Energy Impact Fund for local energy affordability.
OpenAI announced plans to acquire Ona, aiming to bring secure, persistent cloud environments to Codex for long-running AI agent workflows in enterprise settings.
Google is wiring Google Business Profile directly into the Gemini app, letting business owners query performance data, draft review replies, and edit profile details from a chat interface. Rollout starts this month globally, excluding the EEA and UK.
OpenAI announced that its models and Codex are now accessible through Oracle Cloud, letting enterprise customers apply existing Oracle cloud commitments toward OpenAI usage with enterprise-grade security and governance.
A Smashing Magazine study found that participants with cognitive disabilities surfaced 1.8× more usability issues and suggestions than general-population testers across three test websites. The article details recruitment methods, study design, and quantified findings relevant to anyone building user-facing tools.
ServiceNow AI published a benchmark and dataset evaluating seven ASR models on code-switched (bilingual) speech across four language pairs, released through their AU-Harness evaluation tool. ElevenLabs Scribe V2, Gemini 3 Flash, and AssemblyAI Universal 3-Pro came out on top, while Whisper Large V3 Turbo performed poorly due to defaulting to translation mode on mixed-language audio.
Google DeepMind announced Gemini 3.5 Live Translate, a speech-to-speech translation model supporting 70+ languages, available via the Gemini Live API, Google Meet, and the Google Translate app. Developers get public preview access through the Live API and Google AI Studio.
VoidZero, the company behind Vite, Vitest, Rolldown, and Oxc, is joining Cloudflare. This JavaScript Weekly roundup also covers Safari 27 beta, Electron 43 beta, the Rust port of React Compiler, and several notable releases.
Google's John Mueller confirmed on Bluesky that hyphenated domain names carry no SEO penalty, noting the technical limit is 61 hyphens. Search Engine Journal provides historical context on why the SEO community has long treated them as spammy.
OpenAI has filed a confidential draft S-1 registration statement with the SEC, signaling movement toward a potential public offering. No timeline for next steps has been disclosed.
CSS-Tricks published a concise taxonomy distinguishing four commonly conflated CSS concepts: scroll-driven animations, scroll-triggered animations, container query scroll states, and view transitions.
A Hugging Face hackathon participant documents their failed attempt to build an LLM-powered game generator using Nemotron 30b, detailing the prompt engineering and RAG strategies that didn't work and the scaled-back HTML toy maker that did.
Google published new Search Central guidance positioning itself as the canonical authority on SEO, AEO, and GEO best practices, while explicitly distancing itself from third-party SEO tools and services.
A Hugging Face hackathon project runs four different labs' small models as separate agents in an emergent economy simulation, surfacing practical lessons about serving heterogeneous models, information isolation, and bounded memory in multi-agent setups.
The UK's CMA now requires Google to let publishers opt out of AI search features without ranking penalties, and Google has started testing a Search Console toggle for it — but the reporting only includes impressions, not the click data publishers need to make an informed decision.
Search Engine Journal breaks down what Google's Gemini Deep Research Max means for web operators: AI agents now arrive with private user context, and your page competes not just against other public pages but against the user's own data.
A Build Small Hackathon project runs five autonomous trading agents on Qwen2.5-3B via vLLM, demonstrating that a small model can reliably produce structured output for a multi-agent simulation while requiring heavy prompt engineering to compensate for weak reasoning.
Google has documented a native integration between Google Business Profile and Google Analytics, surfacing local engagement metrics like calls, directions, and bookings directly in GA reports. The integration has notable limitations around multi-location segmentation and data retention.
NVIDIA releases Nemotron 3.5 Content Safety, a 4B-parameter model built on Gemma 3 that adds custom policy enforcement, auditable reasoning traces, and a public safety dataset to its existing multimodal and multilingual classification capabilities.
VoidZero is joining Cloudflare, bringing the core Vite team along. Vite remains MIT-licensed and vendor-agnostic, and Cloudflare is launching a $1M open source fund for the Vite ecosystem.
CSS-Tricks updated their Almanac entry for offset-path, the CSS property that defines a movement path for animated elements. The reference covers the rename from motion-path, current value support, rotation control, and integration with the Web Animations API.
OpenAI announced expanded capabilities for GPT-Rosalind, a model focused on life sciences, adding biological reasoning, medicinal chemistry, genomics analysis, and experimental workflow features.
Google highlights five Search features — AI Mode, Google Lens, Circle to Search, Virtual Try-On, and Lens-based resale estimates — positioned as tools for thrift and vintage shoppers. This is a consumer-facing feature roundup, not a developer or API announcement.
Microsoft announced Web IQ, a set of grounding APIs that serve passages and structured evidence from Bing's index directly to AI agents. It's currently accepting expressions of interest with no public availability or pricing yet.
H Company releases the Holo3.1 family of computer-use models in four sizes (0.8B to 35B-A3B) with quantized checkpoints for local inference, expanded mobile support, and native function-calling protocols.
JavaScript Weekly #788 spotlights a practical npm vetting checklist alongside npm 11.16.0's new opt-in install-script policy, plus major releases from Ember, Node.js, and Astro.
OpenAI published a statement outlining its stance on AI policy, political advocacy, regulatory engagement, and clarifying that no outside political group represents the company.
JetBrains released Mellum2, a 12B-parameter MoE model that activates 2.5B parameters per token, aimed at latency-sensitive code and text tasks. It's Apache 2.0 licensed and available on Hugging Face.
Node.js 26.3.0 ships a larger default Buffer pool, a new permission.drop API, configurable HTTP header validation, updated root certificates, and a heads-up about the future of macOS universal binaries.
Search Engine Journal synthesizes statements from Microsoft AI's Mustafa Suleyman, Nvidia's Jensen Huang, and a Boston Globe book review to ask what differentiates human professionals once AI can approximate most white-collar tasks. The piece is aimed at SEO practitioners but the underlying question applies to anyone whose daily output is increasingly reproducible by tooling.
Smashing Magazine published its June 2026 community wallpaper collection, continuing a monthly series that's been running for over 15 years. The batch includes new submissions from international artists plus archived favorites, available in a wide range of resolutions with and without calendar overlays.
Astro's May 2026 roundup covers two minor releases (6.3 and 6.4), the ongoing Astro 7 alpha with Vite 8 and Rust compiler support, TinaCMS defaulting to Astro, and a deep bench of community integrations and migration stories.
Search Engine Journal analyzes how Google I/O's agentic commerce demos — Universal Cart, agentic booking, and background information agents — shift business visibility away from traditional click-based search toward agent-mediated flows that are harder to measure and optimize for.
Search Engine Journal highlights why scaling AI-generated content isn't translating to better rankings, pointing to a mismatch between AI outputs and evolving long-tail search behavior. The piece promotes a structured 4-layer AI Ops framework from CallRail's Darrell Tyler.
Google published a trivia quiz about I/O 2026 announcements, built entirely through vibe coding in Google AI Studio by a non-developer. It doubles as a showcase for the Antigravity coding agent and the studio's expanded capabilities announced at I/O.
CSS-Tricks' latest roundup covers cross-platform Safari testing, the new ::checkmark pseudo-element, anchor positioning workarounds using data attributes and advanced attr(), and notes Firefox 151 bringing container style queries to Baseline status.
Smashing Magazine covers CSS contrast-color(), a native browser function now shipping in all three major engines that returns black or white based on which has more contrast against a given background color. The article details the spec split between Level 5 and Level 6, progressive enhancement patterns, and the uncertain future of APCA.
OpenAI published a case study on how Endava, a global IT services company, uses Codex to restructure software delivery workflows, reportedly cutting requirements analysis from weeks to hours.
Astro 6.4 makes the Markdown pipeline pluggable, ships a Rust-based Markdown processor called Sätteri, and adds Cloudflare helpers for advanced routing.
Google is extending its Preferred Sources labels into AI Overviews and AI Mode, while expanding the Highly Cited badge and introducing new article carousels. The feature has grown from roughly 90K to 345K selected sources since its global expansion.
Artificial Analysis and IBM launch ITBench-AA, a benchmark testing frontier AI models on agentic SRE tasks like Kubernetes incident diagnosis. No model breaks 50%, making it one of the least saturated agentic benchmarks available.
CSS-Tricks demonstrates several text-reveal techniques using negative letter-spacing, transparent color, and transitions — no JavaScript required. The patterns combine existing CSS features like ::first-letter and the ch unit to simulate per-character targeting without a ::nth-letter selector.
iPullRank tested how Google's Personal Intelligence feature affects brand mentions in AI Mode, finding that Gmail-seeded brands saw significantly higher appearance rates. The study is small but is one of the first controlled attempts to measure this opt-in personalization layer.
CSS-Tricks editor Geoff Graham reflects on sharply declining traffic across front-end publications and makes a case for why human-written technical content still matters — and how it should adapt.
Part 2 of CSS-Tricks' series on cross-document view transitions tackles the real scaling problem: going from one named element to hundreds without your CSS collapsing into an unmaintainable wall of selectors. The key is view-transition-class, which decouples animation styling from per-element identity.
Google's Liz Reid says AI Mode's multilingual model architecture lets the feature expand across countries and languages faster than previous Search features. The comments came in a post-I/O interview but included no specific rollout timelines or verifiable data.
Search Engine Journal breaks down Cloudflare's new Agent Readiness Score scanner — a 16-check, 5-category audit tool at isitagentready.com that scores how prepared a website is for AI agent interaction. The article details every check, flags where the composite score is structurally misleading, and walks through a real scan result.
Search Engine Journal argues the real post-I/O risk isn't that SEO is dead but that Google's AI features — especially information agents — reduce the need for users to click through to websites at all.
Search Engine Journal synthesizes three independent data points — an MIT lecturer's classroom confession session about AI-written essays, a Graphite study showing AI content plateauing near 50% of new web content, and freelancer stress data from The Accountancy Partnership — into an argument that the content market is splitting into commodity and non-commodity tiers.
CSS-Tricks revisits CSS centering with a comprehensive look at modern approaches, cataloging roughly 100 ways to center elements and arguing that understanding alignment mechanics matters more than memorizing snippets. The article also highlights newer CSS features like text-box for trimming font-related whitespace.
Google started rolling out its May 2026 core update on May 21, the second broad core update this year. The rollout may take up to two weeks; no specific goals or new guidance accompanied the announcement.
The first LTS point release for Node.js 24 'Krypton' lands with UUIDv7 generation, abortable fs.stat(), test order randomization, and a hardened HTTP client request options merge.
Search Engine Journal published an on-demand webinar recap covering Answer Engine Optimization (AEO) strategies for 2026, presented by Conductor's Shannon Vize and Pat Reinhart. The session focuses on content types that earn AI citations, KPI reframing for AI-first search, and agentic workflows for scaling authority content.
CSS-Tricks examines Stack Overflow's dramatic decline from over 200,000 questions per month at its 2014 peak to roughly 3,000 in 2026, attributing it to aggressive moderation culture and, more recently, the rise of LLMs.
A sponsored study from Victorious, published via Search Engine Journal, tested 177 brands across eight AI platforms and found roughly 90% had zero AI search mentions in Q1 2026, with visibility patterns varying sharply by vertical.
JavaScript Weekly #786 covers a dense week: an RFC to make npm install scripts opt-in, Dr. Axel Rauschmayer pulling his blog offline due to AI crawlers, Bun's Rust rewrite merge, Deno 2.8 teased, Angular 22 RC, and TanStack's post-compromise hardening efforts.
CSS-Tricks walks through the real, current way to enable cross-document view transitions on multi-page sites — replacing a deprecated meta tag with a CSS at-rule — and digs into silent failure modes like the 4-second timeout and aspect ratio warping.
Lily Ray tracked 220+ sites using AI content platforms and found that the majority lost significant organic traffic after an initial peak, with 54% losing 30%+ of peak traffic. The piece documents a recurring boom-bust pattern across industries and vendors.
Google added Google-Agent to its official fetcher list — a user-triggered AI agent that browses on behalf of humans and explicitly ignores robots.txt. The article covers the robots.txt implications and Google's experimental adoption of the web-bot-auth cryptographic identity protocol.
Search Engine Journal examines Similarweb traffic data alongside Meta's Q1 2026 earnings to argue that Meta's serial pivots — metaverse, then AI — reflect a company that has never settled on a coherent business definition, with user decline and ballooning costs as the visible consequences.
Google ended FAQ rich results and an Ahrefs study found no meaningful AI citation lift from adding JSON-LD schema to already-visible pages. The SEJ article examines what this means for schema strategies and the growing GEO advisory space.
Google published official documentation stating that AEO and GEO are just SEO, and explicitly lists tactics like llms.txt files, content chunking, and special AI markup as unnecessary for its generative AI search features.
CSS-Tricks' latest roundup covers Heerich.js for 3D voxel scenes, the of <selector> syntax reaching Baseline, range syntax nuances for container queries, scroll-driven animations, and platform updates from Chrome 148 and Safari 26.5.
CSS-Tricks walks through calculating and displaying discounted prices entirely in CSS using the upgraded attr() function, CSS math functions like mod() and round(), and custom counters — no JavaScript required.
Liquid Web consolidated its WordPress plugin brands — including Kadence, SolidWP, Iconic, and others formerly under StellarWP — into four core products, triggering user confusion over licenses, disappeared product pages, and broken redirects. Existing customers are reportedly being grandfathered in.
Node.js 22.22.3 'Jod' is a maintenance LTS release with OpenSSL 3.5.6, updated root certificates, a batch of V8 backports, and several fixes for use-after-free and null-pointer bugs across crypto, zlib, and http2.
Search Engine Journal published an on-demand webinar covering how to build local landing pages that surface in AI-generated search answers, with a focus on structured data, listings, and review signals for multi-location brands.
CSS-Tricks breaks down the TC39 ShadowRealm proposal, which introduces a new kind of JavaScript realm designed for code isolation without spawning a separate thread. The article explains what realms are, how they differ from Web Workers, and why a same-thread sandbox is useful.
Ahrefs ran a controlled test on 1,885 pages and found that adding JSON-LD schema markup produced no meaningful increase in AI citations across Google AI Overviews, AI Mode, or ChatGPT. The correlation between schema and AI visibility appears to reflect overall site quality, not a direct causal link.
Search Engine Journal argues that Google's April 2026 AI product launches matter less than the behavioral shift happening underneath: mainstream users are adopting conversational, multimodal search patterns, and that trend should drive content strategy more than any single infrastructure announcement.
Google's Universal Commerce Protocol added cart management, live catalog queries, and OAuth 2.0-based loyalty account linking in its March 2026 update, alongside a simplified Merchant Center onboarding path targeting retailers without dedicated engineering teams.
Google has fully deprecated FAQ rich results, completing a rollback that started in 2023. The markup won't break anything, but it no longer produces visible results — and Search Console tooling support is being removed on a staggered timeline through August 2026.