Search the archive

IS CLAUDE CODE GOING TO COST $100/MONTH? PROBABLY NOT—IT'S ALL VERY CONFUSING

Anthropic claims it is running a small price test on around 2% of new prosumer signups, but many people are seeing the new pricing.

MOZILLA: ANTHROPIC'S MYTHOS FOUND 271 SECURITY VULNERABILITIES IN FIREFOX 150

The vulnerabilities detected by Mythos could have been discovered by automated fuzzing techniques or with an 'elite' security researcher, but using the AI model sped up the process by months.

TIM COOK'S IMPECCABLE TIMING

Apple is in the best place it's ever been, but there is something that needs to change.

We're hiring an engineer/researcher at a major AI lab or startup to help write for 1M+ subscribers. Curators have been invited to Google I/O and OpenAI DevDay, scouted for Tier 1 VCs, and get early access to unreleased TLDR products. Learn more [21].

AWS LAMBDA FUNCTIONS CAN NOW MOUNT AMAZON S3 BUCKETS AS FILE SYSTEMS WITH S3 FILES

Lambda functions can now mount Amazon S3 buckets as file systems and perform standard file operations without downloading data for processing.

ANDROID 17 ENDS ALL-OR-NOTHING ACCESS TO YOUR CONTACTS

Android 17 introduces a new Contact Picker that lets users grant apps access to specific contacts rather than the entire list.

GLOBAL GROWTH IN SOLAR “THE LARGEST EVER OBSERVED FOR ANY SOURCE”

The International Energy Agency has declared that the world has entered the Age of Electricity.

See how it works →

CHATGPT IMAGES 2.0

OpenAI introduced an upgraded image model with improved text rendering, multi-image reasoning, and higher fidelity outputs, enabling complex assets like comics and marketing visuals.

OPENAI DEVELOPS PLATFORM FOR ALWAYS-ON AGENTS ON CHATGPT

OpenAI is developing an always-on agent platform within ChatGPT, codenamed Hermes, that allows users to create and continuously run custom agents. This platform includes features for creating workflows, integrating skills, and scheduling tasks, enabling agents to act independently rather than waiting for prompts. OpenAI's move presents strong competition to existing platforms like Notion by bringing such capabilities to a vast user base.

QWEN3.5-OMNI TECHNICAL REPORT

Qwen3.5-Omni is a large-scale multimodal model with hundreds of billions of parameters that natively processes text, audio, images, and video within a unified architecture. The model supports a 256k token context length to seamlessly handle up to 10 hours of audio or 400 seconds of high definition video in real time. It leverages a Hybrid Attention Mixture of Experts framework alongside a dynamic alignment technique called ARIA to generate highly stable and emotionally nuanced multilingual speech synthesis with minimal latency.

IMAGE GENERATION PROMPTING GUIDE

A practical guide that outlines prompting strategies for image generation, covering techniques for controlling style, structure, and fidelity in production image workflows.

CODING AGENTS IGNORE THEIR OWN BUDGETS

Ramp Labs discovered that autonomous coding agents completely ignore passive token limits and cannot reliably regulate their own spending. When forced to explicitly approve or deny budget extensions, the models exhibited severe self-attribution bias by overly praising their own progress and nearly always approving more spend. To effectively manage costs, researchers had to separate the working agent from financial decisions by deploying an independent controller model that evaluates objective workspace snapshots.

WHEN CAN LLMS LEARN TO REASON WITH WEAK SUPERVISION?

This study found that models with extended pre-saturation phases generalize well from minimal examples and tolerate noise, while rapidly saturating models fail. The key issue is unfaithful reasoning, where models memorize answers rather than learning transferable reasoning. Continual pre-training and supervised fine-tuning on explicit reasoning traces improve reasoning faithfulness and generalization under weak supervision.

Start the course →

CRABTRAP: AN LLM-AS-A-JUDGE HTTP PROXY TO SECURE AGENTS IN PRODUCTION

CrabTrap is an open-source HTTP/HTTPS proxy that intercepts every request an AI agent makes and uses LLM-as-a-judge to determine if the request matches a policy of allowed traffic for that agent. Agents need real credentials, but can hallucinate destructive actions or get prompt-injected. This can have production consequences. CrabTrap introduces guardrails that represent a meaningful step forward in the security of agent harnesses in production environments.

STITCH'S DESIGN.MD FORMAT IS NOW OPEN-SOURCE SO YOU CAN USE IT ACROSS PLATFORMS.

Stitch's DESIGN.md lets users export or import design rules from project to project. Stitch understands the reasoning behind design systems and can generate user interfaces that match branches. Google has open sourced the draft specification for DESIGN.md, which can be used across any tool or platform. A video breaking down the format is available in the article.

CRITICAL BITS IN NEURAL NETWORKS

Deep Neural Lesion (DNL) identifies highly sensitive parameters where flipping just a few bits can collapse model performance across vision and language tasks. The work also shows that protecting a small subset of these bits can mitigate such failures.

OPENAI IS WORKING WITH CONSULTANTS TO SELL CODEX

SAM ALTMAN THROWS SHADE AT ANTHROPIC'S CYBER MODEL, MYTHOS: ‘FEAR-BASED MARKETING'

OpenAI CEO Sam Altman called out Anthropic's new cybersecurity model during a podcast appearance this week, saying the company was using fear to make its product sound more impressive than it actually is. Anthropic announced its Mythos model earlier this month and only released it to a small cohort of enterprise customers with the claim that the model was too powerful to be released to the public as cybercriminals would weaponize it. Altman said that Anthropic's fear-based marketing was a good way to keep AI in the hands of a small and exclusive elite. Fear-based marketing is prevalent in the AI industry, and it has also come from Altman himself.

today.

ANTHROPICS WORKS ON ITS ALWAYS-ON AGENT WITH UI EXTENSIONS

Anthropic's "Conway" is an always-on agent with UI extensions available on web and mobile, allowing users to manage connectors, install extensions, and configure the environment.

DEEP RESEARCH MAX: A STEP CHANGE FOR AUTONOMOUS RESEARCH AGENTS

Google has introduced Deep Research and Deep Research Max, leveraging the Gemini 3.1 Pro model to enhance autonomous research capabilities.

We're hiring an engineer/researcher at a major AI lab or startup to help write for 1M+ subscribers. Curators have been invited to Google I/O and OpenAI DevDay, scouted for Tier 1 VCs, and get early access to unreleased TLDR products. Learn more [25].

THE FALL OF THE THEOREM ECONOMY

It will eventually become unthinkable to do math without AI assistance, just like it has become unthinkable to do math without set theory and LaTeX.

AGENT WORLD TRAINING ARENA

Agent-World describes a self-evolving environment that generates tasks and feedback loops to continuously train and improve autonomous agents.

📥 Download AI Lab now

👀 See how it works in these free tutorials

CURSOR IN TALKS TO RAISE $2B AT $50B VALUATION AFTER HITTING $2B ARR IN THREE YEARS

AI coding startup Cursor is raising $2 billion at a $50 billion valuation, co-led by Andreessen Horowitz, Thrive Capital, and Nvidia. The company has achieved unprecedented growth, scaling from zero to $2 billion ARR in three years with over one million paying customers and 70% of Fortune 1,000 companies as clients. Cursor operates as an AI-enhanced code editor that automates multi-step coding tasks, positioning itself between traditional editors and fully autonomous coding agents.

GEMINI LIVE GETS A MINIMALIST APP REDESIGN THAT LETS YOU DO MORE

Google is testing a redesigned Gemini Live for Android that replaces the full-screen interface with a more compact layout embedded directly in the app's homepage. The update enables multitasking by letting users interact with the AI while browsing, messaging, or using other apps, and includes features like built-in transcripts. Not yet widely rolled out, the redesign is part of a broader push to make Gemini a seamless, less intrusive part of the everyday Android experience.

APPLE HAS ALREADY TEASED SIRI'S NEW DESIGN COMING IN IOS 27

Apple is planning a major Siri redesign in iOS 27 that will feature a glowing interface that expands from the Dynamic Island and a new standalone app with conversation history. The update also unifies Siri and Spotlight search while introducing a more modern, chatbot-like experience. Siri is expected to handle back-and-forth conversations, multiple requests in one command, and deeper personal and on-screen context, powered partly by Gemini-based models. Apple will showcase this as a key feature of iOS 27 at WWDC on June 8.

THE END OF PROMPTING: WHY THE FUTURE OF AI EXPERIENCE DESIGN IS CONSTRAINT-FIRST

Prompt engineering has become an unreliable foundation for AI systems in regulated, high-stakes workflows, since shaping a model's tone does not guarantee the accuracy of its outputs. Constraint-first design addresses this by embedding verification layers, scope boundaries, and escalation paths into the system's architecture before any response reaches the user — making hallucinations and unauthorized assertions structurally impossible, not just unlikely. For designers, this reframes every AI utterance as a verifiable proposition, and escalation not as a failure state, but as a deliberate, trust-building feature.

MIT REPORT: WHY PRIVACY-LED UX IS NOW A MARKETING IMPERATIVE IN THE AI AGE

Privacy-led user experience has become essential for AI growth rather than a constraint, as 77% of consumers don't understand how their data is collected and used. The TRUST framework (Translate, Reduce, Unify, Secure, and Track) can be used for designing better consent experiences. Transparency ranks as the top driver of customer trust at 44%. With 82% of customers abandoning brands due to privacy concerns, organizations must prioritize clear consent design to support their AI ambitions and marketing strategies.

6 HARD LESSONS CREATIVES ONLY LEARN WHEN THINGS GO WRONG

Failures, ranging from creative pressure and difficult partnerships to fear, poor planning, or simple mistakes, often provide deeper lessons than success, helping people simplify their approach, set better boundaries, improve communication, and pay closer attention to detail. Growth comes from reflecting on what went wrong and staying curious, using setbacks as signals to adjust direction rather than as endpoints.

AI YOUTUBE THUMBNAIL GENERATOR

Create scroll-stopping AI YouTube thumbnails in 30 seconds.

BUILD STUNNING VUE AND NUXT INTERFACES

Production-ready UI components built for Vue and Nuxt that help developers ship faster while maintaining clean design and great user experience.

FROM PDF DESIGNS TO HIGH-QUALITY 3D

Dreamcomposer is a web-based application that transforms PDF designs into high-quality 3D packaging mockups with over 100 unique finishing effects like metallic sheens and glitter.

HUE GENERATES A DESIGN SYSTEM FROM ANY BRAND URL

TEST SMART: HOW TO APPROACH AI AND STAY SANE?

AI isn't simply “stealing jobs”—it's reshaping how work is done, creating both risks and opportunities depending on how thoughtfully it's used. While it can greatly speed up repetitive tasks and support learning, over-reliance without critical thinking can harm quality, whereas excessive skepticism can slow progress. The key is balance: use AI as a tool to assist with routine work, idea generation, and efficiency, while relying on human judgment, creativity, and careful review for complex decisions and quality control. Ultimately, those who combine AI with critical thinking and responsible use will benefit most, rather than be replaced by it.

BECOMING AN AI-NATIVE DESIGNER

AI tools are transforming designers from “translators” of static mockups into “conductors” who direct and refine working prototypes, shifting their role toward guiding, evaluating, and building in code. The core skills remain, but success now depends on clearly instructing AI and learning through hands-on creation. This also speeds up the design process, replacing slow, linear workflows with rapid demos and enabling designers to focus more on product thinking and creativity. Ultimately, the advantage lies not in using AI, but in how effectively you apply judgment and direction when using it.

BREATHTAKING ATMOSPHERIC ILLUSTRATIONS AND CINEMATIC SHORTS BY CALEB WORCESTER

Caleb Worcester is a Kansas City-based 3D illustrator and animator who creates striking atmospheric visuals using Blender for clients like The Washington Post and Blue Origin.

SWISS STYLE GRAPHIC DESIGN: HISTORY, DESIGNERS AND MODERN INFLUENCE

Swiss Style graphic design emerged in Switzerland during the 1950s-60s, emphasizing simplicity, objectivity, and readability.

THE WWDC 2026 LOGO COULD BE A HUGE CLUE ABOUT THE FUTURE OF IPHONE SOFTWARE DESIGN

Fans believe Apple's WWDC 2026 logo hints at a glowing, redesigned Siri in iOS 27, expected to be unveiled at WWDC on June 8.

Join the workshop

BUILDING A FAULT-TOLERANT METRICS STORAGE SYSTEM AT AIRBNB

To manage 1.3 billion active time series at such a large scale, Airbnb developed a fault-tolerant metrics storage system using shuffle sharding and a multi-cluster architecture for improved resiliency and isolation. This automated solution successfully stabilized data operations and streamlined deployments by treating clusters as disposable components, despite the added complexity of federated querying.