PRIMA is an eye prosthesis that can restore functional sight to patients with advanced macular degeneration. It gives wearers the ability to perceive shapes and patterns, with most wearers regaining the ability to read a year after receiving the device. Wearers can adjust contrast and brightness and magnify their vision up to 12 times. The device currently only provides black-and-white vision, but researchers are working to produce a full range of grayscale.
An experimental drug called valiltramiprosate (ALZ-801) slowed cognitive decline by 52% in people with mild cognitive impairment in a study of more than 300 people aged 50 or older who were genetically predisposed to Alzheimer's disease. Participants who received the drug experienced about 18% less atrophy in the hippocampus compared to those who received a placebo. ALZ-801 is a twice-daily pill that can be taken at home. It works by preventing amyloid proteins from clumping and forming plaques.
One in three employees uses AI tools without approval, risking data leaks and compliance violations. We love the Enterprise AI Governance Kit from You.com [14]. Its ready-to-use templates help protect your organization, including security checklists, usage policies, and
Chrome's Writer API helps people create content that conforms to a specified writing task. It can be used to write any kind of content, improve content, or generate ideas for more content. People looking to use the API will have to join the Writer API origin trial running in Chrome 137 to 148.
uv revolutionizes how easy installing and using Python can be. The tool can install any Python version, install packages, manage virtual environments, and solve dependency conflicts extremely quickly. It is very fast and works on almost any operating system or platform. This post includes a quick guide on how to get started with uv.
Project Rainier, a 1,200-acre site near Lake Michigan, is one of the largest operational AI data centers in the world. Amazon is using the site to train frontier artificial intelligence models entirely on its own chips. Amazon and its competitors have pledged more than $1 trillion towards AI data center projects. They are all racing to build supercomputing sites to meet an expected explosion in demand.
Tens of thousands of newly laid-off white-collar workers in the US are entering a stagnant job market with seemingly no place for them. Investors are pressuring companies to streamline operations. Nearly two million people in the US have been without a job for 27 weeks or more. Meanwhile, opportunities for front-line, blue-collar, and specialized workers are growing.
Three new physical attacks that target Trusted Execution Environments (TEEs) have surfaced, raising questions about the true security offered by these TEEs and the exaggerated promises and misconceptions coming from the big and small players using them.
OpenAI is sharing any prompt that requires a Google Search with Google, whoever is doing its scraping, and also with whoever's site shows up in the search results.
Your agent needs to call APIs, read documents, and make decisions. But without proper controls, it can access data it shouldn't, call services without approval, or leak sensitive information.
Cursor 2.0 features Cursor's first coding model, Composer, and a multi-agent interface. Composer can complete most tasks in under 30 seconds. Its training allows it to accurately navigate and understand large and complex codebases. Cursor 2.0's new interface allows multiple agents to operate in parallel, managed via git worktrees or remote machines. The new release is aimed at addressing workflow challenges seen in modern development teams.
Nvidia's 2025 revenue more than doubled to $130.5 billion, up 114% year-over-year. Jensen Huang announced $500 billion in AI chip orders and plans to build seven government supercomputers. President Trump is expected to discuss Nvidia's chips with Xi Jinping as the high-end processors remain a key sticking point in US-China trade tensions due to Washington's export controls.
Optimization for the wrong target produces excellent results at that target while failing at the actual problem. Pathology needs optimization for clinical utility, generalization across institutions, and robustness to real-world conditions. The next generation of pathology AI will be smarter about what size and architectural approach actually fit the problem. It will be smaller, more interpretable, more rigorously validated, and more aware of its own limitations.
Anthropic researchers used concept injection, artificially inserting neural activity patterns into Claude's processing, to test whether models could identify the change to their own thinking. Claude Opus correctly recognized injected concepts about 20% of the time, sometimes noticing that something was “off” before mentioning the inserted concept.
The gpt-oss-safeguard models (120b and 20b parameters) let developers apply custom safety policies at inference time rather than training their own classifiers on thousands of labeled examples, using chain-of-thought reasoning to explain decisions.
This post discusses how to make Claude Skills available globally. The trick is to treat the skills and an enumerator as global assets. Put the skills in Codex's installation folder and drop a script that enumerates skills into the PATH. This allows Codex to discover skills in any repository by running that script as instructed in AGENTS.md.
SWE-1.5 is a frontier-sized model with hundreds of billions of parameters optimized for software engineering. It achieves near state-of-the-art coding performance and sets a new standard for speed. The model is now available in Windsurf, serving at up to 950 tokens per second. It can be used to deeply explore and understand large codebases, build end-to-end full-stack apps, and easily edit configurations without needing to memorize field names.
Agent labs ship product first and build infrastructure later. They turn AI models into goal-directed systems that deliver outcomes, capturing the real value in the AI stack. Founders only need a deep understanding of a domain and the ability to build reliable workflows on top of existing models. The most valuable developer skills are shifting from model architecture to system design, evaluation engineering, and domain-specific workflow optimization. Investors should look for companies that capture workflow data and have clear evaluation metrics. The moat is in the data and feedback loops, not the models.
Mercor's marketplace connects former employees of investment banks, consulting houses, and law firms with the AI labs looking to automate those industries. Its customers include OpenAI, Anthropic, and Meta. The startup pays industry experts up to $200 an hour to fill out forms and write reports for AI training. It has tens of thousands of contractors that it pays more than $1.5 million to daily. Mercor's annual recurring revenue has grown to roughly $500 million in just under three years.
Saddled with lawsuits over teen suicides and heightened legal scrutiny, the AI roleplay startup will start phasing out teen chat access by November 25.
Bitcoin held around $113,000 as traders rotated into stablecoins ahead of the Federal Reserve's expected 25 basis point rate cut, with stablecoin supply rising for the first time since September, while exchange order book depth collapsed to 40% of pre-liquidation levels. BTC futures open interest remained at $26.8 billion with highly divergent funding rates amid $514 million in 24-hour liquidations split 69-31 between longs and shorts. Spot BTC ETFs recorded $202.4 million in daily inflows while ETH ETFs saw $246 million, as Grayscale's Solana Trust converted to a spot ETF (GSOL) on NYSE Arca and World Liberty Financial prepared to airdrop 8.4 million WLFI tokens to early USD1 users following a points program that drove USD1 to become the sixth-largest stablecoin.
Consensys has reportedly hired JPMorgan and Goldman Sachs to lead its initial public offering. The move follows other successful crypto IPOs this year, including Circle and Bullish, as firms capitalize on a renewed wave of institutional and regulatory openness under the current administration's pro-crypto stance. Consensys also operates Infura, Linea, and backs ETH treasury firm SharpLink, positioning itself as a core pillar of Ethereum's infrastructure ahead of the IPO.
Western Union announced plans to launch the US Dollar Payment Token (USDPT), a Solana-based stablecoin issued by Anchorage Digital Bank, in the first half of 2026. The token will allow Western Union's 100 million customers to send money internationally without relying on local currencies, expanding its 175-year legacy of global payments into the digital asset era.
Brale has launched a powerful new API that bridges traditional banking rails with onchain payments, enabling users to ACH debit via Plaid and mint stablecoins instantly. The API supports custodial and non-custodial wallets, ~20 blockchains, and ~50 stablecoins, all at bank-level pricing (as low as $0.01 per transaction). It also includes RTP support, customizable transfer branding, and the ability to fund any off-chain payment with stablecoins, creating a low-cost link between legacy finance and crypto-native systems.
Verify to trade, verify again to stake — crypto companies keep paying to verify the same users, adding friction that tanks conversion rates. Persona Connect [8] lets you share verified KYC and KYB data with partners you trust. Users verify once, data travels safely, and you can stay compliant and onboard faster. See how reusable KYC works
Crypto is repeating the 1980s-90s hedge fund era when blowups like LTCM led to the creation of funds of funds as a middle layer for allocation and risk management. Vaults now require the same onchain infrastructure that traditional finance developed after massive failures. Onchain funds of funds would handle full risk management systems tracking positions across vaults, quantitative risk checks, manager due diligence, protocol-level fundamental research, and dynamic rebalancing to maintain concentration limits and risk-adjusted returns. Crypto users lack the expertise to properly size vault positions or track underlying risks, creating demand for professional fund-of-funds services that will reach billions in AUM as the winners filter for best strategies and publish public research to justify allocations.
N1 is a permissioned high-performance L1 launching with sub-1ms latency, 100,000+ TPS per application through dedicated compute environments, and gasless transactions where protocols pay less than a cent in fees instead of users. The chain supports Solidity, Rust, and TypeScript development with in-house collaboration on flagship apps to avoid ecosystem fragmentation, using Jito (Re)staking for validation and NordVM for asynchronous virtual machine execution.
The 2020s mark the start of an 80-year civilizational reinvention cycle with AI, clean energy, and bioengineering converging for the first time as solar and wind costs drop 20% per doubling of output while CRISPR makes biology programmable. Crypto serves as the coordination layer through blockchains and zero-knowledge proofs, enabling a shift from financial capitalism optimized for extraction to sustainable capitalism based on transparent protocols and abundance-driven cooperation. This author's investment thesis expresses this supercycle view by going long crypto (trust layer), AI (intelligence layer), and abundance plays in energy storage and bioengineering. They predict a 25-year boom despite current political division and institutional turbulence that mirrors crisis periods preceding past reinventions.
Hyperliquid launched HIP-3 permissionless perp market deployment requiring 500,000 HYPE stake (~$20 million) and charging double the standard fee rate, with projections of $60 million in trading fees, $16 million in auction revenue, and 5 million HYPE staked across 10 markets in year one at 5% of total Hyperliquid volume. The upgrade responds to $3.3 billion bridging to competing alt-perp DEXs like Lighter, Aster, and EdgeX, with nearly 20 teams announcing HIP-3 markets. Equity perps remain unproven with weak demand – Ostium generates only $1.8 million daily stock perps volume versus $44.3 million in crypto – though the model mirrors Ethereum's L2 strategy with stronger economic alignment through 50% protocol fees and mandatory HYPE staking versus L2s paying minimal fees to Ethereum.
We should question the sustainability of risk curator business models in DeFi (platforms like Aave, SparkLend, and Gauntlet) as they generate poor returns relative to total value locked (TVL).