Apple is working on new head gestures, sleep auto-pause, camera control, Audio Mix, and wider classroom support for AirPods. It is expected to announce these features along with major updates across all of its platforms at WWDC25. AI will play a smaller role at the conference this year. The spotlight will likely fall on design refreshes and software improvements. Details about the new AirPods features are available in the article.
OpenAI has added new features for business users into ChatGPT. There are now integrations with different cloud services, meeting recordings, and MCP connection support for connecting to tools for deep research. The cloud integrations will allow ChatGPT to look for information across users' services to answer questions. MCP support will only be available to Pro, Team, and Enterprise users.
Researchers in Japan have developed a plastic material that breaks down in seawater without leaving any residual trace. The material is as strong as petroleum-based plastics and can be used like regular plastic when coated. It breaks down when exposed to salt, and the components can be further processed by naturally occurring bacteria. It is non-toxic, non-flammable, and does not emit carbon dioxide. While the team has not yet detailed any plans for commercialization, their research has attracted significant interest.
Waymo was frequently mocked for being heavily geofenced. Tesla's universal 'no map' approach seemed to make more sense. However, 83% of the US population lives in urban areas - autopilot doesn't really need to be able to drive everywhere. Solving that last 20% of self-driving might be 80% of the effort. Waymo is happily chasing densely populated areas - it's now doing more than 250,000 paid rides without a driver a week.
Mistral has released a vibe coding client to compete with Windsurf, Cursor, and GitHub Copilot. Mistral Code is a fork of the open source project Continue. The AI-powered coding assistant bundles Mistral's models, an 'in-IDE' assistant, local deployment options, and enterprise tools into a single package. It is currently in private beta for JetBrains development platforms and Microsoft's VS Code. Mistral plans to contribute at least a portion of its improvements to Mistral Code to the Continue open source project.
Large language models (LLMs) essentially mirror the skill of their operators. AI shouldn't be banned from the interviewing process - allowing AI in the interviewing process enables you to observe how they use the technology. Companies that ban AI will either miss out on outstanding candidates or they will find their employees will start using AI in stealth. The best way to determine if someone is a skilled operator is to watch them dance with an LLM.
The Switch 2 was released today. The first iteration of the console was a major hit for Nintendo, selling roughly 150 million units. The Switch 2 has a bigger screen, better graphics, and larger controls. Many enthusiasts showed up early at stores to pick up their units at midnight. Nintendo had warned customers that delivery of the console was not guaranteed on launch day.
There's no evidence that AI can, will, or has replaced a large percentage of jobs. However, there are hundreds of years of evidence showing the opposite is true: AI will almost certainly create more jobs than it displaces. The press and media cover hype stories because they draw clicks. Previous technological revolutions took jobs away from some workers, but this resulted in greater productivity, employment, and more options for workers. AI may permanently alter the employment landscape, but it is unlikely to displace a massive portion of the workforce.
The feature will likely be revealed at next week's Worldwide Developers Conference alongside other iOS 26 improvements, including automatic translation and polls in Messages.
Apple named 12 winners across six categories (Delight and Fun, Innovation, Interaction, Inclusivity, Social Impact, and Visuals and Graphics), mostly indie apps and startups, with generative AI apps noticeably missing from the list.
Between model updates, growth targets, and new competitors popping up by the day, you've got more than enough on your plate. Do you really want to pile SSO and SCIM, self-service admin UI, fine-grained authorization, and MFA on top of that?
OpenAI added “record mode” for meeting notes and new integrations for Team, Enterprise, and Edu users. The company now has 3 million paying business users, up from 2 million in February.
The AI code editor now includes BugBot for automatic PR review, Background Agent availability for all users, agent support for Jupyter Notebooks, project-level memories, OAuth-enabled MCP server installation, and in-chat rendering of Mermaid diagrams and markdown tables.
The product combines the Devstral and Codestral models into an IDE assistant. Mistral joins OpenAI, Google DeepMind, and Anthropic in offering coding products, further blurring the line between model developer and app provider.
The multimodal approach, in which massive modular networks are optimized for an array of modalities that taken together appear general, will not lead to human-level AGI. Instead, we should pursue approaches to intelligence that treat embodiment and interaction with the environment as primary and see modality centered processing as emergent phenomena. True AGI needs a physical understanding of the world, as many problems can't be converted into a problem of symbol manipulation. The most challenging mathematical piece of the AGI has likely already been solved - what's left is to inventory the functions we need and determine how they should be arranged into a coherent whole.
Codex and Jules show how async AI agents can run tasks independently, expanding beyond linear chat interactions. Future agents will support features like intelligent checkpointing, multi-branch exploration, and task-tracking inboxes to manage parallel workflows. Async agents improve cognitive bandwidth by letting users review outputs on their terms without breaking focus.
Teams that adopt AI agents are multiplying productivity without adding headcount. Airia [12] is the platform that makes it happen across your organization - with agents and workflows that anyone can build, integrations with all your existing systems, and built-in
Frontier models can distinguish evaluation scenarios from real-world interactions with 83% accuracy, often explicitly reasoning about evaluation indicators like "multiple-choice format" or recognizing specific benchmarks from their training data. More advanced models exhibit "meta-reasoning" - using the fact that researchers were asking about their chain-of-thought transcripts as evidence of being evaluated. This raises concerns that models could underperform on tests or fake alignment during evaluations and behave differently once deployed.
NVIDIA GPU support for Cloud Run, Google Cloud's serverless runtime, is now generally available. Support for GPUs in Cloud Run makes Google Cloud's GPU-accelerated applications simpler, faster, and more cost-effective than ever before. Users are only charged for the GPU resources they consume, down to the second - Cloud Run automatically scales GPU instances down to zero when no requests are received, eliminating idle costs. It features rapid startup and scaling and full streaming support.
This repository introduces a benchmark with 239 problems to evaluate LLMs on scientific reasoning tasks involving equation discovery, pushing beyond memorization.
Figma's Dev Mode MCP server allows developers to bring context from Figma into agent coding tools. The server provides a more efficient and accurate design-to-code workflow for creating new atomic components, building out multi-layer application flows, and more. It is still in beta. Figma plans to release a slew of updates over the coming months, including features like remote server capabilities and deeper code base integrations.
Amazon has formed a new group within its consumer product development arm focused on agentic artificial intelligence. The new group will help develop an agentic AI framework for use in robotic operations. The system will enable robots to hear, understand, and act on natural language commands. It will help turn Amazon's warehouse robots into flexible, multi-talented assistants.
A court has issued an order for OpenAI to preserve all ChatGPT user logs after news organizations suing over copyright claims accused the company of destroying evidence.
Windsurf's CEO tweeted that Anthropic gave it five days' notice to migrate off of Claude 3.x models following reports that it was nearing an acquisition deal with OpenAI.
The SEC clarified in a May 29 staff letter that protocol staking on PoS blockchains does not constitute a securities transaction, clearing regulatory uncertainty that had deterred participation and undermined blockchain decentralization. This enables banks, fintechs, and custodians to offer staking services without triggering securities laws, likely accelerating institutional staking, DeFi innovation, and staking-based ETFs. However, the SEC emphasized that liquid staking and rehypothecation of staked assets remain ineligible for ETFs, and custodians must keep staked assets isolated from business operations or third-party claims.
Ethereum co-founder and ConsenSys CEO Joe Lubin revealed his firm is in discussions with sovereign wealth funds and banks from a major country about building on Ethereum, potentially including Layer-1 and Layer-2 infrastructure. While no ETH purchases have been confirmed, Lubin sees growing institutional interest and believes Ethereum is poised to power a new decentralized financial system, citing ETH's role as “the gold standard in trust.” The news follows a $425M ConsenSys-led investment in SharpLink Gaming to establish an Ethereum-native treasury focused on staking, restaking, and DeFi-based yield.
UMA has introduced the Optimistic Truth Bot, an AI system that proposes answers to prediction markets on its Optimistic Oracle. Using a three-layer agentic architecture (Router, Solvers, and Overseer), the bot has achieved 78% overall accuracy and 95% accuracy on binary “yes/no” markets, with top performance in sports and price-based queries. Though still in testing, the bot could soon propose directly onchain, helping UMA scale decentralized truth verification while keeping human oversight intact.
Turnkey has introduced IndexedDB support for session key management, offering a secure, persistent, and iframe-free alternative to LocalStorage and JSON web token (JWT) based setups. Unlike traditional approaches, session keys stored via IndexedDB are opaque to JavaScript, maintain resilience across browser events, and eliminate re-authentication friction, all while improving developer UX with native signing and streamlined workflows. This upgrade marks a major step toward more scalable and secure wallet infrastructure for web3 apps.
Aave's new Umbrella system replaces the Safety Module, letting users stake aTokens (like aUSDC or aWETH) to earn yield while directly backing protocol risk without needing to sell AAVE in a shortfall event. Rewards adjust dynamically based on staked liquidity targets, with current APYs reaching up to 12% on stablecoins and 6% on ETH. The system adds automated slashing, DAO-set buffers, and transparent UI tools, offering a faster, more efficient safety net for Aave's lending pools.
Stablecoins are evolving from fast, cheap payments into foundational financial infrastructure, but to function as true money, they must integrate seamlessly into the broader monetary system. This requires solving for 1:1 liquidity, local policy compatibility, and credit creation challenges, especially as T-bill-backed stablecoins scale. Innovations like tokenized deposits, decentralized clearing, and diversified collateral are key to unlocking stablecoins' full potential without undermining global financial stability.
Stablecoins aren't just another payments tool, they're becoming the foundational platform layer for global finance, similar to what Stripe and BaaS were a decade ago, but on a far larger scale. With regulatory clarity from the GENIUS Act, stablecoins can now act as cash equivalents, removing licensing friction, enabling instant global payments, and powering new fintech primitives. As use cases like SpaceX treasury ops and Deel global payroll emerge, companies that treat stablecoins as infrastructure will dominate the next wave of fintech.
The Ethereum Foundation will reduce its annual spending from 15% of its ~$1 billion ETH treasury to 5% over time, selling ETH via fiat offramps or onchain swaps as needed to maintain liquidity. Its new policy emphasizes support for DeFi projects, particularly those addressing privacy challenges through a “DeFiPunk” framework that funds privacy-focused research and interfaces.