Thread AI, a leader in composable AI infrastructure, has raised $20 million in Series A funding. Despite the rapid adoption of AI, many organizations struggle integrating AI into complex, evolving environments. They often must choose between rigid, pre-built AI tools that don’t fit their workflows, or costly custom solutions requiring extensive engineering. Thread AI addresses this gap with composable infrastructure that connects AI models, data, and automation into adaptable, end-to-end workflows aligned with each organization’s specific needs. Unlike traditional RPA, ETL, or workflow engines that mirror human workflows or require large infrastructure investments, Thread AI’s Lemma platform allows enterprises to rapidly prototype and deploy event-driven, distributed AI workflows and agents. Lemma supports unlimited AI models, APIs, and applications all within a single platform built with enterprise-grade security. This speeds up deployment, reduces operational burden, and simplifies infrastructure, while maintaining governance, observability, and seamless AI model upgrades. As a result, Thread AI equips enterprises with the flexibility to keep up with rapidly changing AI ecosystem, and the cross-functionality to unlock the power of AI across their entire organization. Lemma users report a 70% improvement in process response times, along with significant efficiency gains as AI-powered workflows reduce operational bottlenecks. Early customers have expanded their AI implementations by 250% to 500%, demonstrating Thread AI’s scalability and practical impact.
Anysphere’s Cursor code editor can spot more subtle bugs that don’t render a code snippet unusable but lead to unexpected behavior or slow performance
Code editor startup Anysphere has closed a $900 million funding round. Anysphere is now worth $9.9 billion, nearly four times what it was worth after its previous funding round in December. Anysphere offers a popular code editor called Cursor that uses artificial intelligence to automate programming tasks. An embedded chatbot allows developers to generate code, ask for technical explainers and perform related tasks. The software processes user requests using more than a half dozen large language models. Cursor is based on VS Code, one of the most popular open-source code editors on the market. As a result, developers can bring over keybindings from their existing VS Code environments. Keybindings are user-defined keyboard shortcuts that speed up tasks such as jumping to the start of a code file. Cursor also works with VS Code extensions. Using Cursor, developers can describe the task they wish to perform in natural language and have an AI model generate the corresponding terminal command. The code editor also functions as a kind of spell checker. It can automatically spot and correct mistyped characters, which removes the need for developers to interrupt their workflow in order to perform troubleshooting. Mistyped characters render the code file that contains them unusable, which makes them fairly easy to detect. Developers spot the issue as soon as they attempt to run the file. According to Anysphere, Cursor can also spot more subtle bugs that don’t render a code snippet unusable but lead to unexpected behavior or slow performance.
Theta Lake’s solution can detect and report the presence of AI assistants in meetings, identify confidential data exposed in AI content and detect missing disclaimers and disclosures in AI content
Theta Lake’s AI Governance and Inspection Suite is designed to address these challenges and go beyond the scoping and access control guardrails built into the leading AI tools. Specifically, detecting and reporting the presence of AI assistants in meetings; providing a flexible way to decide which summaries are captured and how they are retained; and the ability to specifically inspect AI generated content for conduct, compliance, or data protection risks that may require supervision action, user governance, remediation, and / or adjustment of data access, scoping and related guardrails in the AI tools themselves. Theta Lake’s suite of modules are purpose-built into Unified Communications & Collaboration integrations and their AI suites. This makes it seamless to adopt both the UCC suite, its AI capabilities, and Theta Lake’s additional governance and inspection capabilities, all without complex service engagements or lengthy deployment models. The Theta Lake AI Governance and Inspection Suite has three core modules including: AI Assistant & Notetaker Detection Module; Zoom AI Companion Inspection Module and Microsoft Copilot Inspection Module. The Theta Lake AI Governance and Inspection Suite and modules enable organizations to: Identify confidential data exposed in AI content; Detect missing disclaimers and disclosures in AI content; Recognize AI tools used in communication and collaboration interactions; Insert user notifications into conversations around the use of AI tools; Remediate AI content in conversations; Notify, send training info, and document / prove notifications for compliance; Selective capture, analysis, and retention options for AI content; Seamlessly deploy in parallel with their UCC tools.
DMind.ai’s open-source Web3-native expert AI model excels at complex crypto topics such as DeFi protocols, tokenomics and smart contracts, and outperforms general-purpose at just 10-30% of their token cost
DMind.ai is an open-source AGI research lab aiming to create a robust, transparent AI infrastructure for digital finance, starting with Web3, which includes cryptocurrency, DeFi, and real-world asset tokenization. The goal is to create open AI models, benchmarks, and datasets rooted in deep, expert-curated knowledge, supported by a global open-source community. The goal is to empower developers, builders, and researchers across the digital financial ecosystem by sharing tools and research. DMind-1. Fully open-source and built on the robust Qwen3-32B model, DMind-1 is fine-tuned specifically for the nuanced world of Web3: Expert-Level Understanding: DMind-1 excels at complex crypto topics such as DeFi protocols, tokenomics, smart contracts, and strategic investment planning. Enhanced Accuracy and Reasoning: It demonstrates significant improvements in Web3 task accuracy, expert-level reasoning, and provides reliable, precise interactions. Cost-Effective Excellence: Impressively, DMind-1 outperforms larger, general-purpose, general-purpose models on Web3-specific tasks—all at just 10-30% of their token cost. DMind.ai also launched the “DMind Web3 Benchmark,” the first comprehensive evaluation standard designed exclusively for Web3-specific reasoning and domain expertise. The benchmark quickly made waves, skyrocketing to the #1 spot on Huggingface’s dataset rankings shortly after release. This open-source evaluation suite contains thousands of expert-reviewed questions spanning nine core areas – from blockchain fundamentals and infrastructure to DeFi, NFTs, DAOs, token economics, security, and more. The benchmark provides precise domain-level scoring, empowering precise model comparisons and fostering deeper community insights.
Endless’s on-chain AI platform combines Stability’s image generation tech that lets users transform simple sketches into high-quality, stylistically consistent images for building Web3 apps
Endless Web3 Genesis Cloud aims to accelerate the adoption of decentralized artificial intelligence (AI) by integrating Stability AI’s image generation capabilities with its robust Web3 infrastructure. Supported by academic resources from the University of Surrey, Endless bridges Web2 and Web3 ecosystems, offering a one-stop platform for building Web3 applications with a Web2-level user experience. Its modular framework enables efficient AI application development, facilitating the rise of hyper-intelligent AI agent systems. Endless’s custom Sketch-to-Image workflow enhances its product offering. Utilizing Stability AI’s technology will initially target high-demand, pain-point-intensive use cases, showcasing the enhanced user experiences enabled by technical synergies. Key applications include: Streamlined Content Creation: By integrating Stability AI’s “sketch-to-image” functionality with Endless’ on-chain AI infrastructure, users can transform simple sketches into high-quality, stylistically consistent images. This lowers the creative barrier, encouraging broader participation while boosting efficiency for professional creators through AI-driven style learning and automated complex draft generation. On-Chain Assetization and Trading: Creators can leverage Stability AI’s AI-generated content (AIGC) tools to produce works efficiently, then use Endless’ to mint these creations as on-chain assets, such as non-fungible tokens (NFTs), ensuring verifiable ownership and seamless trading. Endless’ community-driven token incentives further motivate creators to contribute high-value content to the platform.
Google Gemini Live is now adding captions so you can read Gemini’s responses in real-time
Google is adding real-time captions to Gemini Live. After ending a Gemini Live session, you already get a text transcript of the conversation. Google is now adding captions so you can read Gemini’s responses in real-time. The top-right corner of the fullscreen Gemini Live interface is gaining a transcript button. (Google is using the same rectangular icon as Android’s Live Caption capability.) Tapping presents a translucent overlay at the center of the screen. At the moment, Gemini Live does not let you start a conversation if the volume is muted or too low. There might be occasions where you can briefly speak to your phone, but can’t have audio playing out loud (like when you are without headphones). The new transcript allows you to go Live in those situations, while some prefer reading instead of listening to responses. For other people, this will be faster. The upcoming Search Live, which is also powered by Project Astra, has a prominent “Transcript” button in the UI to read responses as they come in. Captions started rolling out to Gemini Live on Android earlier this week with more reports today, but it’s not yet widely available.
Fidelity’s new Managed Futures ETF to offer a liquid, alternative strategy by capitalizing on price trends in a broad set of markets through disciplined, systematic long-short investing
Fidelity Investments announced the launch of Fidelity Managed Futures ETF (FFUT), a liquid alternative strategy that aims to capitalize on market trends through disciplined, systematic long-short investing. FFUT is listed on The Nasdaq Stock Market LLC and available commission-free for individual investors and financial advisors through Fidelity’s online brokerage platforms. “The new managed futures strategy is designed to provide clients with an investment option that can help diversify their portfolios with the ease of an ETF wrapper,” said Roberto Croce, portfolio manager of Fidelity Managed Futures ETF at Fidelity Investments. “Fidelity’s robust quantitative research, sophisticated investment capabilities, and disciplined investment process help us provide a differentiated strategy.” FFUT seeks capital appreciation across market regimes, while aiming to provide especially attractive risk adjusted returns during periods of equity market drawdowns. In an effort to achieve its investment objective, the fund pursues a strategy intended to capture the persistence of price trends (up and/or down) in a broad set of markets — including equities, fixed income, currencies and commodities — utilizing futures, forwards and other derivatives. The ETF is competitively priced with an estimated gross expense ratio of 0.83% and estimated net expense ratio of 0.80%.
SOLVE’s AI-driven pricing tool for bond markets blends crowd-sourced bids and offers from market participants and over 300 feature inputs derived from reference, trade and quote data to offer predictive, real-time trade-level pricing
SOLVE has launched SOLVE Px for the Corporate Bond Market, a new AI-driven pricing tool for the high-yield (HY) and investment-grade (IG) corporate bond markets. Covering over 100,000 corporate bonds with industry-leading precision, SOLVE Px provides predictive trade-level pricing that helps traders and portfolio managers act with greater confidence. SOLVE Px is designed to support fixed income professionals who need to make quick, well-informed decisions in often illiquid markets. Built on AI models trained with over 300 feature inputs, including SOLVE’s proprietary quote data, the tool delivers pricing that dynamically adjusts to reflect evolving market conditions, enabling buy-side and sell-side users to uncover trade opportunities, manage risk, and refine execution strategies. Delivered via the SOLVE Quotes web application, FTP and FIX feeds, and soon through APIs and Excel add-in, the solution enhances transparency and pricing accuracy by integrating two core components: 1) SOLVE Quotes™ – crowd-sourced bids and offers from market participants, providing rich real-time pricing inputs: Real-time bid, mid, and offer levels delivered across 50,000+ daily securities and 20 million+ daily quotes. 2) AI-Generated Predictive Pricing – a fully machine learning-driven model trained on hundreds of inputs derived from reference data, trade data, and quote data. Together, these components provide a more comprehensive and actionable view of corporate bond pricing. Eugene Grinberg, CEO of SOLVE said “This offering provides real-time insight into the next likely trade level for both investment-grade and high-yield corporate bonds. With predictive pricing, we’re helping traders and PMs combine their market judgment with data-driven support to improve outcomes and capture opportunity faster.”
Bank of England tests AI to spot real-time payment fraud and uncovered 12% more illicit accounts than they would otherwise have found
The bank of England’s Project Hertha tested the application of modern AI techniques to help spot complex and coordinated criminal activity in payment system data. The experiments were conducted using a state-of-the-art simulated synthetic transaction dataset, developed as part of the project. It includes data for 1.8 million bank accounts and 308 million transactions. The dataset was built by using an advanced AI model trained to simulate realistic transaction patterns. It found that payment system analytics could be a valuable “supplementary tool” to help banks and payment service providers (PSPs) spot suspicious activity. Banks and PSPs participating in the project uncovered 12% more illicit accounts than they would otherwise have found. The experiment also proved particularly valuable for spotting novel financial crime patterns. When trying to spot previously unseen behaviours, it helped achieve a 26% improvement. “The results demonstrate promise but also show there are limits to the application and effectiveness of system analytics,” states the BofE. “It is just one piece of the puzzle. The introduction of a similar solution would also raise complex practical, legal and regulatory issues. Analysing these was beyond the scope of Project Hertha.” The central bank says the results also highlight the importance of labelled training data, robust model feedback loop and explainable AI algorithms to maximise effectiveness.
Report: Stablecoin growth could increase volatility of US treasuries and may trigger less demand for U.S. Treasuries from banks
The rising adoption of stablecoins could reportedly increase the volatility of U.S. Treasury Securities with short-term maturities. Some analysts say that as these dollar-pegged cryptocurrencies grow, their volatility could spread to the bills market. Any disruption in the stablecoin market could trigger liquidations that could drive down Treasury prices, they say. In addition, if money moves from bank deposits to the stablecoin market, there could be less demand for U.S. Treasuries from banks. Other analysts counter that an increase in stablecoin activity would increase the number of buyers of T-bills, which are considered to be cash-equivalent securities, around the world. The stablecoin bill that is making its way through Congress would require stablecoins to be backed by liquid assets like T-bills. Already, two stablecoin issuers — Tether and Circle — hold a collective total of $166 billion in U.S. Treasuries. U.S. Treasury Secretary Scott Bessent has said that a codification of federal rules for stablecoins could boost demand for U.S. debt. The reserves that help maintain the peg of the coin are often a mix of assets that can be exposed to shocks. If there are fluctuations, stablecoin issuers must sell or rebalance those holdings to keep the peg or meet redemptions if and when they are demanded by holders. If an issuer has to sell assets in response to a swell in redemptions, losses may ensue.