Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
The SGLang team outlined how agent workflows are being turned into reusable SKILL.md files, benchmark contracts, review loops, and production debugging playbooks. The post frames agent value around procedural engineering knowledge that can be executed, tested, and reviewed.
State-of-the-art block-wise Diffusion Large Language Models (dLLMs) rely on a remasking mechanism that decodes only the most confident tokens and discards the rest. Recycling computation from the discarded tokens is beneficial, as these tokens retain contextual information useful for subsequent decoding iterations. Residual Context Diffusion is a module that converts these discarded token representations into contextual residuals and injects them back for the next denoising step. It consistently improves frontier dLLMs in terms of accuracy with minimal extra computation overhead across a wide range of benchmarks.
Devin Security Swarm is a cost-effective and accurate way to find security vulnerabilities in complex codebases. It uses a new architecture for whole-codebase reasoning called Agentic MapReduce. Devin maps relevant signals across a repository, fans out focused agents over bounded shards, reduces their findings to one report, then verifies serious vulnerabilities in isolated sandboxes before marking them confirmed. Cognition has published extensive documentation and technical materials about Agentic MapReduce.
Laguna XS 2.1 is a 33B parameter Mixture-of-Experts model optimized for agentic coding and long-horizon tasks, showing a 5.4-point improvement on SWE-bench Multilingual to 63.1%. It supports various platforms and offers three quantized checkpoints for resource-efficient deployment. Licensed under OpenMDW-1.1, it enables open model distribution and is available for download on Hugging Face or via API.
Seed2.0 focused on long-tail knowledge, complex instruction following, reasoning, visual understanding, and search for real-world tasks. The model card described an evaluation-driven approach built around user needs and complex usage scenarios.
Woodside Energy leverages AI to enhance operational efficiency by integrating agentic AI systems in complex industrial workflows, particularly in LNG plant startups through their "Startup Advisor". These AI tools, built upon years of investment in predictive analytics and machine learning, augment human expertise rather than replace it.
In late June, AI hardware saw massive advancements with custom chips transitioning from concepts to actual products. OpenAI, Etched, Amazon, and SambaNova led these significant developments, marking a pivotal shift in the industry. These innovations promise to enhance AI processing capabilities and efficiency.
CursorBench evaluates agents on ambiguous, multi-file tasks from real Cursor sessions.
Claude Enterprise introduces enhanced admin analytics, model-level entitlements, and spend alerts to better manage usage and costs.
The Ramanujan Challenge for AI is designed to test whether AI systems can move from a concrete formula to a valid proof or symbolic derivation.
Apple is expected to shrink the iPhone 18 Pro's Dynamic Island by around 35%, giving users more usable screen space and marking another step toward a future all-screen iPhone. The change, hinted at by iOS 27's Siri AI interface, would be the first major reduction in the display cutout since the Dynamic Island debuted on the iPhone 14 Pro. Apple is also rumored to bring a smaller Dynamic Island to the MacBook Ultra and next year's iPhone 18e as it gradually works to eliminate display cutouts altogether.
Figma achieved ISO/IEC 42001:2023 certification, an international AI governance standard, following an independent audit by accredited body Schellman covering policies, risk management, and development practices across its platform. The audit, run in two stages, tested 38 controls across nine areas, including risk management, data governance, and human oversight. This certification adds to Figma's existing ISO 27001 and SOC 2 Type II credentials, giving customers—especially in regulated industries—independently verified evidence of AI governance rather than self-reported claims.
Base44, the vibe-coding platform Wix acquired for $80 million, has begun rolling out its own AI model, Base1, trained on tens of millions of user interactions. The launch reflects a broader industry debate over whether AI startups built on external frontier models can remain defensible, with founder Maor Shlomo betting that specialization and cost efficiency will give Base44 an edge over rivals like Lovable and encroaching frontier labs. Base44 has grown to over $150 million in annual recurring revenue, though it still trails Lovable's $500 million ARR.
Figma's 2026 releases—AI agents, Code Connect, Make, and Check designs—all depend on a team's design system, so a broken system now fails visibly across stakeholders, developers, and marketers instead of staying hidden with one maintainer. Check designs adds a hard number counting inconsistencies, turning "our design system has drift" from a deniable claim into concrete evidence, though the tool flags intentional exceptions as false positives. Visibility alone doesn't grant authority over roadmap decisions, so design leads should present system health as a delivery risk underlying new AI features.
AI-generated UI code is inaccessible by default, since models learn from non-semantic markup and shortcuts, mirroring the same process gaps that cause security flaws in AI-written code. Treating accessibility as a one-time audit fails at deployment speed. Instead, organizations that succeed embed it into design systems, workflows, and CI automation. Beyond legal risk, accessibility affects billions in consumer spending power and procurement decisions, making it a marker of engineering maturity rather than a compliance checkbox.
Figma's new Skills feature lets teams create reusable AI workflows in plain English that can be shared across an organization, helping the Figma agent apply team-specific processes, writing standards, design critiques, and onboarding routines without repeatedly rewriting prompts. Skills can automate recurring tasks such as critique preparation, project summaries, feedback organization, and task creation, while integrating with tools like Slack, Notion, and Asana to pull context and update external systems. Shared agent chats also make AI workflows more transparent, allowing teammates to learn from each other's prompts, skills, and decision-making.
Try alternative typefaces on real websites before making changes in your stylesheet or opening Figma. See how fonts perform across real content, layouts, and breakpoints.
RetroMac is an open-source desktop browser that brings back the look & feel of the late-90s and early-2000s browsers, built on modern Chromium.
Hypertokens, a concept coined by Figma's Jake Albaugh at Config 2026, would bundle related style properties into one named unit that every tool builds from, replacing scattered hand-copied versions. This addresses drift between design tools and helps AI agents read files accurately instead of guessing, since agents currently receive ungrouped lists of token values with no indication of how they combine. It's not a shipped Figma feature but an early exploration. It closely resembles existing "composite tokens" in the W3C design token spec.
Digital accessibility complaints are rising as legal, procurement, and customer-experience expectations mature, with over 5,000 lawsuits filed in US courts in 2025. Growing regulations, sprawling digital ecosystems, and greater public awareness are driving the trend. Reactive remediation often lets barriers resurface, so sustainable programs instead build governance, continuous testing, cross-team alignment, training, and strategic planning into daily operations.
Reducing friction isn't always the best way to improve conversions—adding a thoughtful onboarding step increased sign-ups by around 25%, while simpler "obvious" optimizations had no effect. Optional questions and excessive flexibility can create unintended problems, such as psychological reactance, lower-quality personalization data, or making users feel they should invest more effort before they've committed. The most effective onboarding balances guidance with autonomy, asking only what's necessary to deliver immediate value without overwhelming users with unnecessary choices.
Evil Martians replaced its contract-illustrator workflow for blog covers first with pure AI image generation, then with a hybrid pipeline pairing AI generation with a human illustrator from agency KOJI for final cleanup.
Pangram Pangram Foundry highlights five serif typefaces designers favor in 2026: Kyoto, Lettra Mono, Fragment Serif, Editorial Old, and Editorial New.
Polaroid's "The Best of Summer is Analog" campaign uses bold, minimalist messaging to encourage people to spend less time on screens and embrace real-world moments in an increasingly AI-driven world.
OpenAI serves WebRTC voice AI to 900M weekly users by splitting the stack into a stateless edge relay for packet routing and a stateful transceiver that owns the heavy WebRTC state, sidestepping the port-exhaustion and state-stickiness problems WebRTC hits on Kubernetes. The key method is encoding routing metadata into the ICE ufrag, a field already exchanged during setup, so the relay routes the first packet by reading the initial STUN request rather than doing a database lookup.
Flow is FoundationDB's custom language extension for C++. It adds efficient actor-based concurrency using asynchronous primitives like Future, Promise, wait(), and ACTOR, while compiling down to native C++ for high performance and deterministic simulation. Flow lets developers write distributed asynchronous systems in a much simpler style than callback-heavy C++, with additional constructs like streams (FutureStream/PromiseStream) and choose/when for coordinating concurrent events.
Expert developers should keep AI coding agents on a “short leash” by reviewing every proposed change, intervening frequently, committing often, and never letting the AI work unsupervised. It also recommends using AI as a code reviewer alongside human review.
The main goal of code review is not to catch bugs but to identify code that is difficult to understand and therefore likely to be difficult to maintain.
OmniRoute is an open-source AI routing gateway that lets developers connect coding tools like Claude Code, Codex, Cursor, and Cline to over 237 AI providers (including 90+ free ones) through a single OpenAI-compatible endpoint. It automatically handles provider failover, cost optimization, and quota management.
Valmis is a cloud platform for building and deploying AI agents that automate business workflows across 100+ integrations using chat, scheduled jobs, webhooks, or app events. It emphasizes security by running agents in isolated containers so AI cannot directly access your API credentials or local files.
Fable 5 is being redeployed with stricter cybersecurity classifiers that block prohibited and high-risk dual-use cyber requests while allowing clearly defensive uses. Anthropic uses a Cyber Jailbreak Severity scale to rate jailbreaks by capability gain, breadth, weaponization, and discoverability.
Expensify uses agent-device, a tool that gives AI agents “hands and eyes” on real devices by hiding platform fragmentation behind one API and returning the screen as a structured accessibility tree rather than a screenshot (richer and cheaper in tokens). They apply it to three repeated chores: an agent-device-evidence skill that reproduces a bug, distills the steps into a replayable script, and captures before/after recordings for PRs; automated Sentry span measurement across multiple runs; and one-prompt React profiling that returns component-specific render data.
Rivian should support Apple CarPlay as an optional feature because it gives drivers access to thousands of familiar apps without replacing Rivian's native interface.
The hardest part of mathematical research is not proving theorems but discovering the right concepts and language that make those theorems almost inevitable.
Apple's new Safari MCP server lets AI coding agents connect directly to a local Safari browser to inspect the DOM, console, network, screenshots, performance, and accessibility.
ZeroFS is an open-source filesystem that makes S3-compatible object storage behave like a high-performance, encrypted POSIX filesystem or block device, allowing applications to use cloud object storage almost like a local disk.
Token prices fell from $60 per million to $0.60 in three years, and AI bills still exploded. A four-person startup ran up a $113,000 monthly bill from a single provider. Uber torched its entire 2026 AI budget in four months. Consumption is the reason. An agent re-reads context and checks its own work, burning 60 to 140 times the tokens of one reply. Track cost per completed task, not per seat, and keep the model layer swappable, because cheaper tokens do not mean a cheaper bill.
Software architecture determines how software behaves. The engineering operating model determines how engineering behaves. Engineering organizations exist to create customer and business value through software. This means the job of engineering leadership is to maximize the value engineering teams create. Redesign the system around the work so that engineers' work matters more.
Salesforce shipped Agentforce, the fastest-growing product in its history, and scaled it to $1.2B in ARR. The stock hit a 52-week low anyway. That paradox is the whole shift. When anyone can rebuild your headline feature in a weekend, the software surface stops being a moat, and value rotates to what you can't clone, the proprietary data loops, the right to move money or push code, and being the tool an agent chooses to call. Outcome pricing seals it, turning vendors into insurers who underwrite every result instead of billing for seats. So what do you own that stays scarce once building is free?
Notion now lets users build interactive HTML right on their Notion pages. AI can turn content into interactive explainers, prototypes, or diagrams. The agent can take documents, sketch a flow, and build a lightweight prototype on the page. Video demos are available in the thread.
This GitHub gist contains a skill that transforms AI into the middle manager for an autonomous software factory. The agent doesn't write code or implement any issues. Instead, it reads the issue tracker, creates and manages Devin coding sessions, and informs the human operator only when it needs to.
All good communication is grounded in an understanding of the receiver's mind. Help is about people before it is about projects. Make it clear that you are someone who is worth helping by demonstrating that you are a serious person through personal connection or by having institutional credibility. Never lie.
Not all mandates are shortsighted and dumb. A mandate is a way of putting organizational muscle behind a decision. They force people to have tough conversations about what tradeoffs to expect and what success looks like. Most big transformation projects end up needing mandates.
When anyone can build almost anything, feasibility stops being the constraint. The questions that still decide whether you have a business are the old unglamorous ones: Will they pay, can you reach them, and does the problem actually hurt? AI made building cheap and made those assumptions the whole game.
Hard-tech startups with accessible technology often see rapid replication due to idea saturation and active founder interest.
MWM makes an annual revenue of around $60 million with no hype or headlines, just consistent wins.
AI disclosure labels can reduce engagement because people assume the creator put less effort into the content. Across 8 experiments and more than 1.1 million TikTok posts, content labeled as AI-generated received 7% to 8% fewer likes and 7% lower overall engagement even when quality stayed the same. Explain how AI supported the work and highlight the human effort behind the final result to help maintain trust and connection. AI tools that require visible skill or creative input are less likely to create this reaction, so the way you present AI use matters as much as the tool itself.
Nearly 80% of UK adults are concerned about misinformation on social media, yet usage remains high. People aged 16-34 spend about 2 hours 31 minutes a day on social platforms, and 88.9% use them each week. Overall weekly reach is 73.9%, and Brits spend 7 hours 24 minutes a day on screens, reinforcing how central digital media is. Smartphones drive 34% of consumption, with social and messaging taking nearly half of mobile activity.