Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
DoorDash's Dot is a small robot that can autonomously drive on roads, bike lanes, and sidewalks to deliver food and small packages. It can move at speeds of up to 20 miles per hour and carry up to 30 pounds. DoorDash is testing the robot with early access partners in the Phoenix Metropolitan area. It plans to make the robots available to the region by the end of the year.
Scientists at the Oregon Health and Science University have created early-stage human embryos by manipulating DNA taken from people's skin cells, then fertilizing them with sperm. The technology could overcome infertility due to old age or disease. It could even allow same-sex couples to have a genetically related child. The technique still needs significant refinement before a fertility clinic could even consider using it, which could take a decade.
Sponsored
Over 6 million tech professionals read TLDR including developers, product managers, marketers, designers and executives. Get in front of your target audience with an ad placement just like this one! Learn more [8] about running a test campaign.
Technical documentation is mostly aimed at people deeply immersed in the ecosystem. However, many developers have to juggle a lot of 'worlds' in their heads daily. It takes a lot of mental energy to restore context and understand what is going on when jumping between projects, languages, and frameworks. Examples are great because a lot of developers look at docs simply because they need to quickly find out how to do something. Tutorials are often helpful, not because of the walkthrough, but because they contain examples.
Apple's Foundations Models framework allows developers to create new intelligent features that are available offline while protecting user privacy, all using AI inference that is free of cost. It is tightly integrated with Swift, making it easy for developers to send requests. The framework is available with iOS 26, iPadOS 26, and macOS 26, and works on any Apple Intelligence-compatible device. This article looks at some of the apps that have leveraged the framework to create new experiences.
Rishabh Agarwal and more than 20 other researchers have left their work at Meta, OpenAI, Google DeepMind, and other big AI projects to join Periodic Labs, a new Silicon Valley startup. Many of them have given up tens of millions of dollars, if not hundreds of millions, to make the move. Periodic aims to build AI technology that can accelerate new scientific discoveries in areas like physics and chemistry. The startup has secured over $300 million in seed funding. It has started its work at a research lab in San Francisco, but it plans to build its own lab in Menlo Park.
Amazon's hardware division has long been seen as a loss leader. The company makes its profits from subscriptions and purchases made through the Alexa voice assistant. Some product lines are now profitable, and others are heading in that direction. The company's recently revamped devices have higher prices, and they're much better products. For the company to become a true leader in devices, it will need breakthroughs that go beyond its current lineup.
Bridge also intends to apply for a New York State trust charter.
The M5 chip could offer substantial GPU upgrades over M4, giving Apple the ability to apply neural processing across applications that are traditionally more graphics-based.
Spotify's co-presidents, Alex Norström and Gustav Söderström, will become co-CEOs on January 1.
Sorting by 'how often' lets you read your most urgent messages first and focus your energy on what matters most.
Elon Musk claims that xAI's Grokipedia will be a massive improvement over Wikipedia, which he has previously accused of being 'woke', even calling for it to be defunded.
Eli Health's Hormometer lets users track cortisol levels with just saliva and a phone.
Sponsored
You've tried Cursor, Codex, and Claude Code. Now try the coding agent that beats them.
OpenAI introduced Sora 2, its next-generation video generation model with improved physical realism, finer detail, and greater control. The update supports synchronized dialogue, sound effects, and user interaction through a new app interface.
Periodic Labs emerged from stealth to build AI scientists that conduct physical experiments autonomously in robotic labs. Founders Ekin Dogus Cubuk (who led Google Brain's materials team that discovered 2 million new crystals with AI) and Liam Fedus (former OpenAI VP who helped create) aim to automate scientific discovery, starting with superconductors, and claim that LLMs need fresh data for physical experiments because they've exhausted the internet as a training source.
Meta is acquiring chip startup Rivos to strengthen its AI hardware capabilities and reduce dependence on external suppliers. The deal gives Meta more control over custom silicon, signaling a push toward deeper vertical integration in its AI infrastructure.
AI is fundamentally advanced by scaling, but AI researchers continue to work on algorithms, architecture, and data as if scaling laws were yet to be discovered. More compute and more energy is the most reliable path to advancing AI. Many labs are aiming for recursive self-improvement, but the notion that AI will algorithmically self-improve ad infinitum ignores scaling laws and that research is compute-bound. This makes problems like autonomous science a better problem for AI researchers to work on.
Quietly and all at once, frontier AIs crossed a threshold into economically valuable work. OpenAI tested experts against AI on multi-hour tasks, and humans won, but barely, with AI's weakness being formatting rather than accuracy. Claude Sonnet 4.5 reproduced an economics paper's findings in minutes, work that would take academics hours of tedious conversion, potentially solving academia's replication crisis at scale. The difference between transformation and waste lies entirely in human judgment about what's worth doing, not merely what can be done.
Sponsored
Leading enterprises are embedding AI at the core of their operations to unlock flexibility, free resources, and accelerate growth. This new model moves organizations from pilots to enterprise-wide impact, creating real competitive advantage.
Agents can now directly exercise the code they are writing, correct errors, dig through existing implementation details, and even run experiments to find effective code solutions to problems. Coding agents are brute force tools for finding solutions to coding problems. Reducing problems to clear goals and sets of tools that can iterate towards those goals can help coding agents brute-force their way to an effective solution. The art of using agents well is to carefully design the tools and loops for them to use.
Logicstar shrank SWE-Bench Verified from 240 GiB to just 5 GiB. It can now be downloaded in under a minute, making large-scale evaluation and trace generation on cloud machines fast and painless. The team achieved this result by restructuring layers, trimming unnecessary files, and compressing the results. This article gets into details about how the smaller SWE-Bench Verified was created.
Compute for pretraining grows 4× annually, but web data increases just 1.03× per year, forcing a shift toward more efficient learning algorithms. Regularization parameters 30× higher than standard practice prevent overfitting when models see the same data repeatedly. Ensembling independently trained models achieves lower loss than simply making models larger. The combination cuts data requirements by 5× to match baseline performance and directly translates to improvements on pretraining benchmarks.
Large AI developers must publicly publish safety frameworks, report critical incidents to the state, and face civil penalties for noncompliance. Anthropic endorsed the law, but Meta and OpenAI lobbied against it. The law establishes a public computing cluster, protects whistleblowers, and mandates annual updates.
Prime Intellect has released INTELLECT-2, a 32B parameter model trained via distributed reinforcement learning instead of traditional centralized GPU clusters. This breakthrough demonstrates that advanced AI models can be developed using a network of volunteer computing resources spread worldwide, potentially democratizing AI development by eliminating the need for massive co-located data centers that currently give tech giants their competitive advantage in the race for more powerful AI systems.
Sponsored
You'll probably like our flagship newsletter. It's all about tech, science, and programming.
Shallot is a lightweight system designed to help maintain clean, scalable interactions with Claude Code.
Google Search's AI Mode now includes a new visual exploration feature that helps users imagine, find, and shop for items through a more interactive, image-driven interface.
Robots often fail in real-world scenarios due to overfitting to specific locomotion strategies.
Semantic speech tokenizers often break under minor audio noise, even when speech remains intelligible.
Chainlink and Swift have launched a UBS Tokenize pilot that uses Swift messages and Chainlink's Runtime Environment to trigger onchain subscription and redemption workflows for tokenized funds without requiring banks to replace legacy infrastructure. The system builds on prior Project Guardian work and uses Chainlink's Digital Transfer Agent standard, with Swift handling ISO 20022-compliant messaging. Coming days after Chainlink's AI-driven pilot for corporate actions, the collaboration positions Swift's global network and Chainlink's smart contracts as a plug-and-play gateway for the $100T+ fund industry to adopt tokenization at scale.
SEC Commissioner Hester Peirce said the regulator is willing to engage with industry participants on tokenizing products, urging them to proactively communicate with the agency while emphasizing the complexity of how tokenized assets interact with traditional forms of the same security. Speaking at the Digital Assets Summit in Singapore, Peirce highlighted that depending on how assets are tokenized, they could represent "one of many different things," requiring a nuanced regulatory approach to handle scenarios where the same security exists simultaneously in paper, electronic, and blockchain-based token forms. The tokenization market currently stands at $31 billion with $714 million in tokenized stocks, and McKinsey projects the market could reach $2 trillion by 2030 as financial institutions increasingly adopt tokenization to enhance liquidity and operational efficiency.
Bridge's Open Issuance is a platform that allows any business to create and manage custom stablecoins in just days while handling reserve management, security, liquidity, and GENIUS-ready compliance. It addresses the limitation of building on external stablecoin providers, where businesses face unpredictable fees and lack control over economics or product roadmaps. The platform enables full customization, including blockchain support, smart contract functionality, and reserve mix allocations balanced between cash and treasuries through partnerships with BlackRock, Fidelity, and Superstate, while providing tools for adoption. Open Issuance solves the liquidity challenge through a unified network where businesses can enable one-for-one interoperability swaps between Open Issuance stablecoins, expanding overall liquidity as new coins launch and allowing instant, permissionless onchain transactions without intermediaries. Existing Bridge-issued coins will migrate to the new platform.
Visa has launched a pilot allowing businesses to pre-fund Visa Direct with USDC and EURC, treating them as available balances for instant cross-border payouts instead of locking up fiat days in advance. The initiative aims to cut settlement times from days to minutes, improve liquidity access, and support programmable payments, with limited availability planned by April 2026. While Visa is not ruling out issuing its own stablecoin, the focus is on scaling existing assets, extending its broader stablecoin strategy that includes card-linked spending, African payments with Yellow Card, and developer integrations via Stripe-owned Bridge.
The GENIUS Act establishes a federal framework for payment stablecoins, allowing banks and federally recognized nonbanks to issue stablecoins under strict reserve, custody, governance, and compliance rules. State issuers can participate if their regimes substantially mirror federal standards and issuance stays under $10 billion. Compliance requirements include face-value redemptions within defined timeframes, backing with safe liquid assets like cash and short-dated Treasuries with monthly reserve reports, capital buffers above reserves, full AML/BSA programs, lawful freeze capabilities, and prohibitions against coercive tying arrangements that force stablecoin usage for other services. The framework prohibits paying holders interest for simply using or holding stablecoins but allows rewards tied to usage, loyalty programs, whitelabeled issuance, and merchant involvement focused on service value rather than interest distributions. Only US-formed entities can issue domestically for full access. Foreign issuers face restrictions and second-tier status unless exceptional compliance is achieved. Key implementation dates include regulator rulemaking by July 2026, GENIUS taking effect by January 2027, and restricted status for non-compliant foreign stablecoins on US platforms by July 2028.
Despite building MegaETH, a scaling protocol with "ETH" in its name, the Mega team hadn't spoken to the Ethereum Foundation until two weeks ago, when they joined a call with the foundation chair, research lead, and ecosystem head that felt more like a customer interview than a formal partnership discussion. Despite eight years in crypto and building at Consensys, the author had never interacted with the EF either, highlighting Ethereum's unique approach of "not grabbing power, but giving it up" as the most counterintuitive and inspiring ideology in crypto. Beyond ideological considerations, Ethereum represents the "Sleeping Beauty" test - if you had to park money for twenty years, you'd choose Bitcoin or Ethereum over newer chains because "your money is safu onchain for many years to come." Ethereum's authenticity and innocence shine through its community-oriented approach, with even Foundation alignment rituals feeling more like community theater than power grabs, though its real struggle lies in communicating its timeless story to a world that anchors feelings on price, experiences, and social proof rather than decentralization principles.
While Plasma achieved record-breaking TVL in its first week, its path to decentralization reveals the downside of L1 architecture. Onboarding validators will cost $550M annually at a 5% inflation rate, despite users already deploying capital without this guarantee. With L2s, security costs scale with transaction volume instead of being a fixed burden, making rollup models more operationally efficient.
Agentic Commerce Protocol defines a common flow for agents to discover products, signal intent, render carts with taxes, and initiate payment instructions. Positioned alongside Cloudflare x402, AXTP, Google A2P, and Visa's Intelligent Commerce, these standards will coexist to solve different parts of agent-driven commerce.
The move strengthens Base's position as a hub for onchain adoption.
Data from Blockworks Research shows that digital asset treasuries (DATs) now hold a larger share of ETH supply than BTC, marking a significant shift in institutional and corporate accumulation trends.
BlackRock may confirm staking in its ETH ETF, withdrawing the 19b-4 filing at the SEC's request to proceed under the new generalized crypto ETF process.
Phantom is the first direct user of Bridge's Open Issuance platform for a new stablecoin called CASH.
Google is updating its signature "G" logo company-wide with a brighter, four-color gradient design that was initially introduced for Google Search earlier this year. The refreshed icon retains Google's iconic four colors while incorporating more vibrant hues and gradient effects to symbolize AI-driven innovation and creative energy. The new design represents Google's evolution in the AI era and will roll out across products, platforms, and services in the coming months.
Instagram is testing a redesign in India and South Korea where Reels becomes the main page, Stories stay on top, and navigation is adjusted to include a Following tab with All, Friends, and Latest feeds. The swipe-based interface makes it feel more like TikTok. Meta stresses it's only a limited experiment for now.
Adobe has launched Premiere for iPhone, a free mobile version of its pro video editor that offers multi-track editing, AI-powered audio tools, Firefly generative effects, and seamless export to social platforms. Designed for creators and social content, it can be used standalone or alongside Premiere Pro on desktop, though advanced features and extra storage require Adobe credits.
Figma now allows users to copy designs directly from Figma Make previews into the design canvas as editable layers, eliminating the need for export steps and enabling seamless collaboration. The company purchased the technology behind the html.to.design plugin from divRIOTS to enhance this functionality after observing community adoption patterns. This "Copy design" feature transforms AI-generated outputs into dynamic building blocks that teams can iteratively refine from prototype to production.
AI design tools excel at generating variations and handling repetitive tasks, but lack the contextual understanding, strategic thinking, and human-centered problem-solving that define professional design work. Rather than replacing designers, AI is creating a collaborative model where machines handle mechanical execution while humans focus on strategy, user empathy, and creative direction. The most valuable designers of the future will be creative strategists who leverage AI capabilities while providing the irreplaceable human insight that transforms generated content into meaningful, purpose-driven solutions.