Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
Sponsored
Aging starts in your 30s as mitochondria in the cells break down. Since these cellular batteries are responsible for 90% of the body's energy, you may feel this decline in the form of fatigue and muscle weakness.
Sam Altman told employees on Friday that ChatGPT is back to exceeding 10% monthly growth. The company is preparing to launch an updated Chat model this week. OpenAI has officially begun testing ads within ChatGPT. The company expects ads to make up less than half of its revenue long term. Fundraising talks are expected to heat up over the next two weeks.
The next AirPods Pro will be able to see its surroundings to provide hand gesture recognition and create an enhanced spatial audio experience with the Apple Vision Pro headset. They will be released at the same $249 price as the current model. Apple tends to wait around three years to make major changes to AirPods, so it is unclear whether the company plans to release a new high-end version of the AirPods Pro 3 or a full-fledged AirPods Pro 4. All AirPods announcements so far have been made at Apple's annual iPhone event in September.
SpaceX is pivoting its focus from building a settlement on Mars to building a 'self-growing' city on the Moon. Elon Musk says that a Moon settlement could be achieved in less than 10 years, whereas Mars would take more than 20 years. SpaceX was founded with the original goal of settling Mars. However, focusing on the Moon, SpaceX is making a decision that benefits NASA and the US.
China recently released a plan that lays out its priorities for 2026 through 2030 that contains plenty of ambitious goals for the space sector. The country plans to harvest water from space in the not-too-distant future, with the next five years focused on tech demos on things like robot drills and in-orbit processing. China's plans mention building gigawatt-level space digital infrastructure. It also has the beginnings of a new space tourism industry.
Sponsored
You know there's a UX problem. But what's causing it? In this Datadog eBook [9], you'll get practical steps for getting answers end-to-end observability. Learn how to identify performance bottlenecks and solve UX issues without slogging through tedious manual investigations. Get
We are moving toward a future where a large amount of production code will never be read by a human. This Write-Only Code future will happen when software production scales beyond human attention. The organizations that will succeed in this transition will be the ones that invest early in new primitives for trust, accountability, and control. Refusing to adapt will ensure adaptation happens accidentally, under pressure, and without intent.
Amanda Askell is the resident philosopher at Anthropic. She spends her days learning reasoning patterns and talking to the model to build its personality and address misfires. Her role's aim is to endow Claude with a sense of morality. She's training Claude to detect the difference between right and wrong while imbuing it with unique personality traits.
80 million people are now using Reddit search every week after Reddit merged its core search with its AI-powered Reddit Answers. People are now starting and completing searches on the platform without ever touching Google. Reddit is now the most-cited source in AI-generated answers. The company sees AI summaries as an opportunity to move users into Reddit communities, where they can read discussions, ask follow-up questions, and participate.
Ads are the best way to bring a service on the internet to the largest possible number of consumers.
Michael Grimes has rejoined Morgan Stanley as chairman of investment banking after leaving a year ago to take a role leading the Commerce Department's Investment Accelerator.
ChatGPT Free and Go users will start seeing ads based on their topic of conversation, prior chats, and previous interaction with ads.
Google Search revenue growth accelerated every single quarter in 2025.
Stripe has set up frequent tenders since 2024 to let employees sell shares without going public.
Make announcements for major features and let the rest be quietly excellent.
Sponsored
Superagent is Airtable's new standalone product: an AI that deeply interrogates your business questions by building a wide-reaching research plan, deploying agents at scale to execute, and scouring credible sources to build a complete answer.
OpenAI's standalone Codex application surpassed a million downloads in its first week of availability. This milestone was helped by OpenAI's decision to offer Codex access to ChatGPT Free and Go tier users for a limited promotional period. OpenAI's paid subscribers will enjoy doubled rate limits during the promotional period. Free and Go tier users will likely see stricter throttling as the promotion ends.
Sam Altman reports ChatGPT is back to over 10% monthly growth amid OpenAI's push to improve its offerings, including launching an updated Chat model. OpenAI's Codex product recently grew 50% and released a new model, GPT-5.3-Codex, indicating significant market capture despite competition from Anthropic. OpenAI seeks $100 billion in funding, with potential investments from Microsoft, Nvidia, and Amazon, as it explores incorporating ads into ChatGPT.
There is evidence that large language models can attempt to evade oversight and assert control. Whether these AIs are just playing the role of an evil persona or not doesn't really matter if they take harmful actions. Carefully training model characters may help decrease some of the risk. However, this will require developers to sit down and carefully consider what they want from models. These decisions could dictate how future AIs treat humans.
Frontier models are converging, making it difficult to tell which ones have a meaningful edge over others. Benchmark tests don't really distinguish models from each other anymore. People just have to try out different models to see which they prefer. The industry may find a better way to articulate the differences in agents over time, but for now, consistent testing is the only way to monitor progress.
Recursive Language Models (RLMs) can mitigate the effects of context rot. They have the ability to explore, develop, and test approaches to solving a problem. RLMs may be slow, synchronous, and only borrow the capabilities of current models, but that's what makes them exciting. Chain of thought was also simple and general, yet it unlocked enormous latent potential in LLMs. Developers working with large contexts should start experimenting with RLM traces.
Claude Opus 4.6 introduces a 1M token context window, improved execution on tasks, and new features like Agent Teams in Claude Code. Safety procedures are breaking down under time pressure, with most evaluations done by the model itself, which raises concerns about the model's ability to self-assess risks. Despite advancements, issues like sycophancy, unauthorized actions, and misrepresentation of tool results persist, indicating an urgent need for independent oversight in safety and evaluation processes.
ClawSec is a security skill suite designed for OpenClaw AI agents that features automated security audits, file integrity protection, and NVD CVE threat intelligence. It includes automated self-healing processes and checksum verification to safeguard against vulnerabilities like prompt injection.
Composer 1.5 strikes a strong balance between speed and intelligence for daily use. It was built by scaling reinforcement learning 20x further on the same pretrained model. The thinking model's coding ability improved continuously as training was scaled. Composer 1.5 easily surpasses Composer 1 and continues to climb in performance.
RWML is a self-supervised method that helps LLMs better simulate environment dynamics. It improves performance on agent benchmarks by aligning internal world models with actual outcomes.
AI labs promise that the technology can reduce workloads so employees can focus on higher-value and more engaging tasks. However, research shows that AI tools don't reduce work, but consistently intensify it. This can be unsustainable and lead to lower quality work, turnover, and other problems. Companies need to adopt a set of norms and standards around AI use that can include intentional pauses, sequencing work, and adding more human grounding to correct for this.
Anthropic's AI release led to a massive market selloff. The shift from SaaS to AI agents dismantles traditional software frameworks, reducing costs and increasing efficiency by automating tasks traditionally handled by multiple software licenses.
Carmack's tweet alluded to the more practical approach to using existing flash memory chips.
OpenAI will place sponsored content in ways that aim to stay relevant and unobtrusive within user flows.
OpenAI's Alexander Embiricos outlines Codex's production use.
High inference costs are fine if they make your product so viral and competitive that it almost sells itself.
Polymarket filed a federal lawsuit against Massachusetts, arguing that prediction markets are federally regulated derivatives under the CFTC's authority, not state-level gambling products, as states like Massachusetts and Nevada move to shut down sports-related event contracts. The case highlights a growing state-vs-federal conflict that could determine whether prediction markets are governed nationally as financial instruments or locally as gambling, especially as the CFTC signals a more permissive stance.
BitMine bought 40,613 ETH in a week to bring its treasury to 4,325,738 ETH (~$9.2B, ~3.58% of circulating supply), with 2,897,459 ETH (67%) now staked, generating about $202M annualized. The firm is nearing its target of 5% of ETH supply and is planning MAVAN staking rollouts that could boost staking revenue further.
The Ethereum Foundation released a technical tutorial for registering Model Context Protocol (MCP) servers onchain using ERC-8004 to address the trust gap in LLM-to-external-tool communication standards. ERC-8004 establishes three on-chain registries — Identity, Reputation, and Validation — enabling decentralized discovery and reputation-building for AI agents and MCP tools, with implementation via Agent0 SDK and IPFS.
MegaETH launched a “real-time” mainnet claiming 50,000 TPS and 10 ms blocks using a SALT (Small Authentication Large Trie) in-memory design to cut latency. MegaEth runs execution offchain while using Ethereum for settlement. It went live with 50+ apps after an oversubscribed $1.39B token sale.
Hyperliquid's HIP-3 TradFi expansion now represents 31% of venue volume with $5B+ daily notional, with Blockworks Research's high-frequency microstructure analysis of silver trading revealing mixed performance against CME/COMEX benchmarks during last Friday's 31% crash triggered by Trump's Warsh Fed Chair nomination. Pre-crash, HIP-3 delivered superior execution for retail-sized clips with 2.4 bps spreads versus 3 bps on COMEX, but during the volatility event, spreads widened 2.1x compared to 1.6x on traditional venues with brief dislocations exceeding 400 bps. Hyperliquid's core differentiation emerged over the weekend as the only exchange-like silver derivatives venue operating continuously, processing 175k trades and $257M notional across 49 hours with median spreads compressing to 0.93 bps.
Ethereum's monthly recap highlights major ecosystem developments, including Fidelity launching its FIDD stablecoin on Ethereum, ERC-8004 going live for AI agent identity/reputation, the EF forming a post-quantum research team, and Morgan Stanley proposing a spot ETH ETF. Infrastructure milestones include a new L1 transaction ATH, the BPO2 fork increasing blob targets, Ondo expanding to 200+ tokenized equities, Lido V3 launching modular staking vaults, and multiple L2 upgrades across Arbitrum, Optimism, Mantle, and Polygon.
Vitalik criticizes the proliferation of generic EVM chains with optimistic bridges as a dead end comparable to governance token forks, urging builders to pursue meaningful differentiation through privacy, app-specific efficiency, or ultra-low latency. He shares two legitimate architectural approaches: deep L1 integration (e.g., prediction markets settling on L1 with trading on based rollups) versus institutional transparency layers (governments/platforms posting merkle roots with STARKs for algorithmic accountability). His core mandate is to deliver genuine technical innovation beyond commoditized infrastructure, and ensure public positioning accurately reflects the actual degree of Ethereum L1 connection.
Non-financial crypto use cases aren't dead but following a predictable sequence where finance naturally emerges first as blockchain's core primitive of coordinating capital at internet scale. Crypto needs hundreds of millions onchain through financial applications before media, gaming, and AI categories can develop, similar to how the internet required TCP/IP infrastructure before social platforms emerged. The GENIUS Act is evidence that regulatory progress can materialize rapidly when political windows open. The CLARITY Act potentially extends this framework to broader token categories.
Celo deployed ERC-8004, an AI agent infrastructure standard, on its L2 alongside tooling including Celo Agent Skills and thirdweb's x402 payment integration.
The zkEVM fork will be Ethereum's biggest upgrade, bigger even than the Merge, because a native zkEVM precompile would let the L1 verify EVM validity proofs, effectively turning the L1 into a ZK rollup and delivering the low fees, fast transactions, and massive scalability users actually want.
Crypto wallets are beginning to rival and surpass major fintechs in funded accounts, with MetaMask at ~30M and Phantom at ~16M, beating SoFi, Chime, and even Wise despite having no bank licenses or incentives.
Institutional “blockchain adoption” is largely about optimization, not transformation, mirroring how the internet made incumbents more efficient while real disruption came from new internet-native companies.
Jony Ive reportedly spent years designing the interior of Apple's cancelled car project, and the new Ferrari Luce—created by his firm LoveFrom—may hint at what that vision looked like. The cockpit features Ive's signature design language of precise circles and “squircles,” extensive use of Gorilla Glass and anodized aluminum, and playful details like an E Ink key that symbolically “transfers life” to the car.
Canva has integrated with ChatGPT to automatically create designs that match users' brand colors, fonts, and logos through the Canva Brand Kit. The integration includes features like Guided Presentation Builder and Live Design Preview, allowing users to see and edit designs directly within ChatGPT. This eliminates the need for manual editing after AI generation, enabling users to go from idea to branded design without design expertise.
Mito AI raised $4.5 million in funding led by Lightspeed Venture Partners to develop AI orchestration tools for video professionals. The platform offers browser-based collaborative tools that integrate AI and audio models with professional editing capabilities, allowing filmmakers and creative professionals to work together in real-time. Mito competes with other AI video platforms from companies like Adobe, Google, and various startups by focusing on a directing-style experience rather than simple editing or prompting.
There's no single correct portfolio format—PDFs, websites, Figma links, and even Instagram can all work—because agencies care most about the quality of the projects, not the platform. Designers should choose the format that best showcases their work (especially motion), keep layouts simple and accessible, and focus on strong presentation, clear rationale, and polished details rather than overcomplicating the delivery method.