AI startup Multiverse Computing has released two AI models that are the world’s smallest models that are still high-performing and can handle chat, speech, and even reasoning in one case. These new tiny models are intended to be embedded into Internet of Things devices, as well as run locally on smartphones, tablets, and PCs. “We can compress the model so much that they can fit on devices,” founder Román Orús told. “You can run them on premises, directly on your iPhone, or on your Apple Watch.” Its two new models are so small that they can bring chat AI capabilities to just about any IoT device and work without an internet connection. It humorously calls this family the Model Zoo because it’s naming the products based on animal brain sizes. A model it calls SuperFly is a compressed version of Hugging Face’s open source model SmolLM2-135. The original has 135 million parameters and was developed for on-device uses. SuperFly is 94 million parameters, which Orús likens to the size of a fly’s brain. “This is like having a fly, but a little bit more clever,” he said. SuperFly is designed to be trained on very restricted data, like a device’s operations. Multiverse envisions it embedded into home appliances, allowing users to operate them with voice commands like “start quick wash” for a washing machine. Or users can ask troubleshooting questions. With a little processing power (like an Arduino), the model can handle a voice interface. The other model is named ChickBrain, and is larger at 3.2 billion parameters, but is also far more capable and has reasoning capabilities. It’s a compressed version of Meta’s Llama 3.1 8B model, Multiverse says. Yet it’s small enough to run on a MacBook, no internet connection required. More importantly, Orús said that ChickBrain actually slightly outperforms the original in several standard benchmarks, including the language-skill benchmark MMLU-Pro, math skills benchmarks Math 500 and GSM8K, and the general knowledge benchmark GPQA Diamond.
United Network launches NFC non-custodial card wallet transforming a traditional bank card into a secure hardware crypto tool; ensures all transactions are secured directly on the chip, providing ease of use for digital asset management
United Network has launched its innovative NFC non-custodial card wallet, transforming a traditional bank card into a secure hardware crypto tool. This innovative solution ensures all transactions are secured directly on the chip, providing unparalleled security and ease of use for digital asset management. All private keys are stored only on the card and are never shared with any external devices or services, ensuring a high level of security. The United Network card wallet offers a seamless user experience through its intuitive web interface or mobile application. Users can manage their digital crypto assets with the simplicity of NFC authentication via their smartphone and advancement of hardware wallet, making complex cryptocurrency operations as easy as a tap. The card form factor helps eliminate the complicated usage often associated with traditional hardware wallets, such as the need to connect to a desktop or laptop by cable. The solution was developed according to strict data protection standards. Every device undergoes security testing, so users can be confident their funds are safe, even in case of a card loss. Key Highlights of the United Network Card Wallet: Comprehensive Functionality: Effortlessly send, receive, store, and swap your tokens to any other crypto wallet with just a tap of a card. Card-Sized. Power-Packed: Slides into your wallet like any bank card, offering simple, secure access while keeping private keys fully in the user’s control. Multichain Support: Supports Bitcoin, Solana, Ethereum, BNB Chain, TON, Venom, and Tron from day one, with more chains coming soon. Top-Tier Security: Implements multichain cryptography and standards for robust security.Seamless Web3 Integration: Facilitates effortless authentication into Web3 applications. Flexible White-Label Opportunities: United Network offers customizable solutions for corporate and white-label needs.
Trovata bolsters legacy treasury management offerings with debt and investment instruments, intercompany transactions, in-house bank support, credit facilities, FX hedging, bank fee analysis and bank account management
Trovata announced its acquisition of ATOM, the enterprise Treasury Management System (TMS) developed by Financial Sciences Corporation. This move marks a bold step forward in Trovata’s mission to modernize and democratize treasury technology, unlocking the full capabilities required to serve large global enterprises. With ATOM’s deep treasury feature set—including support for debt and investment instruments, intercompany transactions, in-house bank support, credit facilities, FX hedging, full domestic and international payment workflow, bank fee analysis and bank account management—fully integrated into Trovata’s cloud-native platform, Trovata becomes the first modern, viable TMS alternative to the legacy incumbents. The combined offering delivers unprecedented scale, flexibility, and performance for corporate finance and treasury teams seeking to modernize. Brett Turner, Founder and CEO of Trovata said, “With ATOM, we have the firepower to compete directly with the legacy incumbents—and replace them. This isn’t just expansion. It’s a generational shift in treasury tech.”
LendingClub’s account feature that offers customers 2% cash back for on-time loan payments made from checking account and 1% cash back when using the associated debit card for qualifying purchases drives 6X jump in daily account openings
LendingClub said that two recent additions to its mobile-first platform are driving more account openings and more visits to the company’s app. CEO Scott Sanborn said the company’s latest innovation, LevelUp Checking, has increased the number of checking accounts opened each day by six times since its launch in June. LevelUp Checking offers customers 2% cash back for on-time loan payments made from this checking account and 1% cash back when using the associated debit card for qualifying purchases. “We’re rewarding borrowers for their financial discipline while allowing us to benefit from a stickier relationship,” Sanborn said. “While it’s still early, the initial results are encouraging.” The target customer for LevelUp Checking and its rewards program has a high FICO score and a high individual income.
Retailers are turning to Snapchat as CX engagement tool to target Gen Z, 77% of who use it for brand discovery and shopping, building authentic conversations using its communication-friendly design, and tapping its AR-based location-sharing
Here are a few specific reasons so many retailers are turning to Snapchat as a consumer engagement platform, starting with favorable user demographics: 1) One big reason an increasing number of retailers are reaching out to customers via Snapchat is that highly coveted Gen Z and even Gen Alpha (the generation following Gen Z of consumers 15 and under) shoppers are spending their time, and money, there. Data backs up the fact that young consumers frequent Snapchat and see it as a branding and shopping platform. According to a recent survey of U.S. and U.K. Gen Z consumers from identity verification technology provider SheerID, 77% of respondents reported learning about a new brand through platforms such as Instagram, Tiktok and Snapchat. Gen Alpha has an estimated $28 billion in direct purchasing power. While 14% Gen Alpha consumers rank Snapchat as their favorite social network, 84% check it at least once a day. If your brand appeals to young shoppers, or you want it to, Snapchat is a vital engagement tool. 2) Snapchat is heavily focused on direct communication among peers, including features for texting, video chatting and exchanging photos. Its communication-friendly design is perfect for both one-to-one and one-to-many conversations, adding an air of authenticity and “real life” interaction which ironically would not be achievable without digital assistance. 3) This sense of digitally enhanced reality extends to Snapchat’s augmented reality capabilities, which retailers are leveraging to get their brand and products in front of consumers in new ways. American Eagle is something of a retail Snapchat pioneer, and another interesting promotion it is currently running on the platform involves including more than 800 retail stores across the U.S. as Promoted Places on the Snap Map location-sharing feature.
IBM taps Infosec’s platform to offer discovery, classification, and lifecycle management of cryptographic assets across hybrid and distributed environments, supporting creation of scalable quantum-safe architecture
IBM Consulting and InfoSec Global are partnering to deliver advanced cryptographic discovery and inventory solutions across all industries and geographies. The rapid advancement of quantum computing poses a growing threat to cryptographic security, as quantum computers can break traditional cryptography, exposing vulnerabilities across digital operations. Organizations worldwide are requiring cryptographic assets to be inventoried, assessed, and modernized. IBM Consulting’s global delivery network and quantum safe security expertise will be combined with InfoSec Global’s AgileSec platform to accelerate customers’ transition to post-quantum cryptography and enable a risk-driven transformation to enterprise-wide cryptographic agility and compliance. The AgileSec platform enables the discovery, classification, and lifecycle management of cryptographic assets across hybrid and distributed environments. The partnership will enable IBM Consulting and InfoSec Global to jointly develop, market, and deliver cryptographic posture management solutions, helping clients tackle their most complex quantum-safe challenges. Client benefits of the IBM Consulting and InfoSec Global partnering could include: Addressing the risk of cryptographic blind spots and supporting adherence to compliance frameworks from NIST, the Federal Financial Institutions Examination Council (FFIEC), and regulatory expectations; Accelerating modernization without costly re-platforming for crypto agility in-place; and Creating a future-ready and scalable quantum-safe architecture with measurable return on investment.
Profound’s platform lets companies see exactly how they appear across AI assistants, create the right content to improve visibility, and deploy AI marketing agents
The traffic, backlinks and click-throughs that shaped traditional SEO carry less weight. For companies, this is a new competitive arena, and it demands new tools to measure and improve how they appear where buying decisions are now being made. Profound’s platform lets companies see exactly how they appear across AI assistants, create the right content to improve visibility, and deploy AI marketing agents that give one marketer the power of an entire agency. From monitoring and analytics to creating content and executing campaigns, Profound is building the marketing command center for the AI era. Co-founders James Cadwallader, Dylan Babbs, and the team have launched a broad set of products, gained traction in every major sector, and become the leader of this emerging category with customers including Fortune 10 companies, Ramp, U.S. Bank, Indeed, MongoDB, DocuSign, Chime, Clay and Plaid. Customers have reported transformative results; Ramp, for example, saw a 7x increase in AI brand visibility for their accounts payable product. As search shifts from blue links to direct AI answers, every company will need Profound to remain competitive.
Paradigm reimagines spreadsheets with AI agents in every cell, automating data collection and enhancing flexibility using multiple AI models for powerful, cost-effective workflows
Paradigm is an AI-powered spreadsheet equipped with more than 5,000 AI agents. Users can assign different prompts to individual columns and cells, and individual AI agents will crawl the internet to find and fill out the needed information. Paradigm works with AI models from Anthropic, OpenAI, and Google’s Gemini and supports model switching. Paradigm attracts users ranging from consultants to sales professionals and finance folks and operates on a subscription model with tiers based on usage. “We want to support every single model because we want our users to be able to have the highest reasoning outputs when they need it, but also the cheapest outputs,” founder Anna Monaco said. “It’s just a constant cycle of evaluating different models, working closely with model providers to make sure our limits are high enough, and then giving some of that power to our users.” Monaco said that she doesn’t really consider the competition because Paradigm doesn’t think of itself as an AI-powered spreadsheet. She said she thinks of it as a new AI-powered workflow that happens to be in the familiar form of a spreadsheet but won’t necessarily stay that way forever.
Empirical Security combines the power of global threat intelligence with localized, organization-specific insights delivering highly accurate threat prioritization based on the specific context of the organization
Cybersecurity startup Empirical Security raised $12 million in new funding to develop and deploy custom artificial intelligence cybersecurity models tailored to each organization’s unique infrastructure and threat landscape. Empirical’s platform offers dual-model architecture that combines the power of global threat intelligence with localized, organization-specific insights. The models are trained on about 2 million daily exploitation events sourced from internet-scale datasets, while local models are fine-tuned using customer-provided and curated internal data. That, delivers highly accurate threat prioritization based on the specific context of the organization. The approach is designed to allow cybersecurity teams to make faster, evidence-based decisions, backed by predictive models that highlight the most critical vulnerabilities. The idea is that instead of relying on generic risk scores that may not reflect the actual danger to a specific business, the company’s local models provide actionable intelligence customized to a given company’s operational environment. Empirical also emphasizes explainability and decision support as key elements of its platform. The platform gives security leaders the ability to justify their strategies with data by integrating risk-based analysis with measurable prediction outputs. The transparency is especially valuable in boardroom discussions, compliance reporting and budgeting, where clear articulation of cybersecurity priorities is essential.
Open-standard RISC-V instruction set architecture’s ability to tailor the architecture to the application driving its adoption across a diverse range of compute platforms from datacenter service processors, AI accelerators, chip makers and custom embedded cores
The open-standard RISC-V instruction set architecture has evolved from an academic project into a legitimate disruptor, increasingly embedded across a diverse range of compute platforms. Amid a wave of industry consolidation, geopolitical shifts, and an AI-driven workload explosion, RISC-V is accelerating toward mainstream relevance with some of the largest chip players backing it. A strong ISA needs a strong ecosystem, and RISC-V is seeing meaningful support across all levels of the stack. Android is being ported to RISC-V, with full RVA23 compliance under development. Ubuntu and Red Hat Enterprise Linux are already available on SiFive hardware, giving software developers familiar toolchains. Tenstorrent, a new AI chip start-up co-founded by legendary chip architect Jim Keller, and also a SiFive customer, is designing RISC-V-based CPUs targeting AI and high-performance edge workloads. Meanwhile, Ahead Computing, another RISC-V start-up with ties to Intel talent, is quietly making strides in server-class designs. Finally, all of the “Magnificent Seven” tech giants reportedly use RISC-V in some capacity, with five of them working directly with SiFive (and likely others). These include datacenter service processors, AI accelerators, and custom embedded cores—proof that RISC-V’s flexibility fits many application use cases. What makes RISC-V different isn’t just the open license, it’s the ability to tailor the architecture to the application. Unlike x86 or Arm, which impose fixed instruction sets and licensing constraints, RISC-V lets chipmakers design exactly what they need, nothing more, nothing less. That’s a major win in the AI era, where model diversity and workload complexity demand hardware tuned for domain specific workloads and energy efficiency. In the US, DoD-aligned programs are increasingly turning to open ISAs to avoid reliance on foreign-controlled IP. Even GlobalFoundries’ acquisition of MIPS plays into the RISC-V narrative, expanding IP offerings while avoiding competitive overlap.