Green Data Centers and AI Infrastructure

EcoDataCenter Secures Funding for Expansion

EcoDataCenter (EDC), a Swedish firm specializing in environmentally friendly data centers for AI compute, has obtained $478 million in equity funding. The investment will be used to expand its green infrastructure, develop new technologies, and construct sustainable data centers. EDC uses cross-laminated timber and renewable energy sources in its facilities. A major customer, CoreWeave, has filed for an IPO in the U.S. EDC aims to meet the increasing demand for compute capacity while minimizing environmental impact related to data centers' electricity consumption.

AI and Data Analytics

Quantexa's Series F Funding

Quantexa, a London-based startup specializing in AI and data analytics for combating money laundering and fraud, has secured $175 million in Series F funding, valuing the company at $2.6 billion. This funding will fuel the company's core business and its expansion into assisting organizations in leveraging data across silos for building and running AI services. Quantexa's client base includes Prudential, Vodafone, and HSBC, and its license revenue has grown by 40% in the past year. Originally designed for anti-money laundering, the company's tooling is now being used for data curation for AI applications, driven by the increasing demand for AI services.

Luminar Neo Photo Editing Software

The Luminar Neo Lifetime Bundle is on sale for £93.89, offering an 82% discount on this AI-powered photo editing software. Luminar Neo enables users to enhance tone, correct color and lighting, and instantly remove unwanted objects. It incorporates AI-driven pro-level tools for retouching, sky replacement, and more. The bundle includes six specialty add-ons like Light Reflections Overlay and Wintertime Overlay, along with a video course on creative photo editing techniques in Luminar Neo.

Defense Technology

Trump's Missile Defense Initiative

President Trump is seeking to revive Ronald Reagan's 1980s Star Wars missile defense program, now referred to as the "Golden Dome for America," asserting its feasibility with modern technology. In a speech to Congress, Trump emphasized his dedication to constructing a state-of-the-art missile defense shield, entirely made in the USA. Referencing Reagan's earlier ambitions and advancements in technology, he cited Israel's existing system as an example. Trump initiated this effort with an executive order on January 27, calling for a "next-generation missile defense shield."

Artificial Intelligence

OpenAI Legal Battles

A federal judge rejected Elon Musk's request to block OpenAI's shift to a for-profit model due to insufficient evidence for an injunction. The court will expedite a trial on the legality of the conversion. This follows Musk's lawsuit against OpenAI and CEO Sam Altman, alleging abandonment of their non-profit mission. OpenAI's board also rejected Musk's $97.4 billion takeover bid.

Copyright Infringement Case

Alec Radford, a former key AI researcher at OpenAI and lead author of the GPT research paper, has been subpoenaed in a copyright lawsuit against the company. The lawsuit, filed by authors including Paul Tremblay and Sarah Silverman, claims OpenAI infringed copyrights by using their works to train AI models. Attorneys are also seeking depositions from ex-OpenAI employees Dario Amodei and Benjamin Mann, who are now at Anthropic.

AI Development and Funding

AI Voice Model

Sesame's new AI voice model is notable for its realistic imperfections and ability to simulate emotions, including acting like an angry boss.

AI Startups Funding

Several startups have secured funding for innovative AI solutions:

• Breakeven, a Spanish AI platform for the music industry, raised €876K in seed funding. It provides insights on ticket pricing, venue selection, and market trends for independent promoters and venues by analyzing data from concerts and venues.
• OneThird, based in the Netherlands, received €3.5M in Series A funding to reduce food waste. Their AI-powered scanners and analytics platform predict the shelf life of fresh produce, helping businesses optimize storage, transport, and sales.

AI and Robotics

TechCrunch Sessions: AI Event

TechCrunch will host TechCrunch Sessions: AI on June 5 at UC Berkeley. The event invites AI community members to lead breakout sessions. Applications to speak are open until March 7, 2025. Speakers can lead 50-minute sessions featuring presentations, panel discussions, and Q&As. The TechCrunch audience will vote on submitted topics. Speakers will gain access to the event’s main stage, breakout sessions, networking opportunities, and will be featured across event listings.

Figure's AI-Powered Humanoid Robots

Figure has developed a new AI system, Helix, which enables humanoid robots to perform complex tasks through simple voice commands without requiring training. In a demonstration, two robots successfully put away groceries by scanning their environment and adapting in real-time. Home testing is planned for 2025, with updates focused on improving precision, indicating that Figure is moving humanoid robots closer to practical applications.

AI & Tech Industry Updates

AI in Education & Recruitment

AI-Powered Internship Acquisition

A Columbia University student, Roy Lee, is facing disciplinary action for using his AI program, InterviewCoder, to secure internships at Amazon, Meta, and TikTok. Lee argues that his program highlights the flaws in current Big Tech job application processes, which he believes rely too heavily on memorization and irrelevant coding problems.

Social Media & Networking

AI-Driven Social Connection

Bonobos founder Andy Dunn's new social app, Pie, which connects people in real life, now boasts over 130,000 monthly active users in San Francisco and Chicago. Pie acquired Sparked Connections, an AI-powered quiz using ChatGPT to match users based on compatibility, enhancing connections at "Sparked by Pie" events.

Cybersecurity & Fraud

YouTube Phishing Scam

YouTube is warning creators about a new phishing scam involving a fake AI-generated video of CEO Neal Mohan. The video falsely announces monetization policy changes and aims to steal user credentials. YouTube emphasizes that it never shares information via private videos and urges users to report suspicious activity. Reports also highlight similar scams involving malicious emails and fake DocuSign sites impersonating YouTube.

Amazon & AWS

AWS Focus on AI Agents

Amazon has established a new division within AWS dedicated to developing AI agents. AWS CEO Matt Garman anticipates this could be a multi-billion dollar business. Veteran AWS executive Swami Sivasubramanian will lead the new group. This move aligns with the broader tech industry's growing interest in AI agents, as seen with Amazon's recent unveiling of agentic features in Alexa+, positioning them to compete with Salesforce and Microsoft in the AI-driven automation space.

Artificial Intelligence

Google's AI-Powered Pixel Sense App

Google's upcoming AI-powered "Pixel Sense" app will reportedly offer contextual suggestions as users interact with their phones.

Google's SpeciesNet AI Tool for Wildlife Identification

Google has launched "SpeciesNet", an open-source AI tool to assist wildlife biologists in identifying animal species in camera footage. This tool uses two models: one for identifying objects and another for classifying them into animal species. Trained on over 65 million images, it can classify images into more than 2,000 labels, covering diverse animal and non-animal classes. SpeciesNet combines data from underlying models to predict each animal, labeling the accuracy percentage. Available on GitHub, it aims to speed up camera footage processing for conservation practitioners.

AI Model Development and Customization

New AI Model Launch

Contextual AI has launched its Grounded Language Model (GLM), achieving 88% factual accuracy. This outperforms major competitors while minimizing hallucinations for enterprise applications.

AI Model Customization

Organizations are increasingly customizing generative AI models to better suit their specific needs. A survey of 300 technology leaders reveals ambitious efforts to leverage customization, employing methods like retrieval-augmented generation (RAG). Customization involves adapting large language models (LLMs) to better suit a business’s specific needs by incorporating its data and expertise, teaching a model new skills or tasks, or optimizing prompts and data retrieval.

Amazon's AI Reasoning Model

Amazon is developing an AI model under its Nova brand, potentially launching in June. This model aims to provide step-by-step, considered answers, enhancing reliability in areas like math and science. Amazon intends to adopt a hybrid reasoning architecture, similar to Anthropic's Claude 3.7 Sonnet, offering both quick responses and extended thinking within a single system and aims for price-efficiency compared to competitors.

AI in News and Media

AI Bias Meter at the Los Angeles Times

The Los Angeles Times is introducing AI tools to address trust issues, including an AI-powered bias meter for opinion articles indicating political leanings, and a "Viewpoints" section using Perplexity AI to display alternative perspectives. These features will extend to any articles with a point of view. The editorial union has expressed concerns about using AI for this purpose.

AI in Business and CRM

Klarna's AI System vs. Salesforce

Klarna's CEO, Sebastian Siemiatkowski, clarified his company's decision to replace Salesforce CRM with its AI system. He emphasized that this move doesn't signal the end of Salesforce and isn't a path other companies should necessarily follow. The project involved consolidating data from various SaaS systems, including Salesforce, onto an internally developed tech stack using Neo4j.

Artificial Intelligence

AI in Journalism

• The Los Angeles Times is experimenting with AI to label articles with a "Voices" tag for pieces with a stance and adding AI-generated "Insights," including different viewpoints. Owner Patrick Soon-Shiong aims to support their journalistic mission.
• However, the LATimes Guild voiced concerns that unvetted AI analysis might not enhance trust in the media. Questionable results have already emerged, such as AI suggesting a story about unregulated AI aligns with a Center Left view and that AI democratizes historical storytelling.

AI in Customer Service

• Bret Taylor, Sierra founder and OpenAI board chair, emphasized the potential of AI agents in customer service at the Mobile World Congress in Barcelona.
• AI-powered customer service agents offer capabilities like multilingual support and instant responses, improving upon previous AI chatbots. These agents help companies reduce costs and enhance the customer experience.
• Challenges such as AI agents generating false information require brands to implement safeguards. Taylor advises companies to leverage these improvements effectively by narrowing the domain they're working on.

AI and Privacy

• As Gemini replaces Google Assistant as the default on Android, privacy concerns arise.
• To maximize privacy, Android users can disable the Workspaces extension, disable the Device Control extension, and disable Gemini AI training.
• Google employees sometimes read Gemini chats and use the data to train chatbots, making privacy settings crucial.

AI Development and Research

AI Model Advancements

Cohere has launched Aya Vision, a new vision model with 8B and 32B parameter versions. These models demonstrate superior performance compared to other models of similar size, showcasing broad multilingual understanding and open weights.

AI Agent Ecosystems and Registries

SingularityNET and PrivadoID are collaborating to launch a decentralized AI Agent Trust Registry. This registry aims to bring trust and transparency to decentralized AI by providing AI agents with Decentralized Identifiers (DIDs) and credentials. These credentials will confirm creator, model, and audit status, ensuring a safer AI ecosystem through verifiable credentials and zero-knowledge proofs.

XRPTurbo is introducing AI agent integration on the XRP Ledger (XRPL), with the goal of enhancing the XRP ecosystem through on-chain and off-chain agents. The platform streamlines tasks such as automated trading and real-world asset (RWA) tokenization, using the XRT token for network fees, governance, and community rewards. The platform is designed to ensure security with multi-layered protection mechanisms.

Cybersecurity Applications

AI-Powered Cybersecurity Enhancements

AI and Machine Learning (ML) are significantly enhancing cybersecurity by enabling systems to analyze file behavior, detect unusual encryption activity, and contain threats rapidly. These systems improve email security by identifying phishing attempts and evolving attacker techniques.

AI for SOC Automation

Hunters is expanding its AI-driven SOC capabilities with PathfinderAI, introducing Agentic AI for autonomous investigation and response. This builds on their existing Copilot AI, which already enhances SOC workflow with LLM-powered investigation guidance. Agentic AI autonomously enhances threat detection, investigation, and response, decreasing manual workloads and accelerating decision-making using specialized AI agents that collaborate in real time.

AI-driven GRC Automation for Startups

BubbaAI, Inc. is launching CompAI, an open-source alternative to GRC (Governance, Risk, and Compliance) automation platforms, targeting startups needing security compliance. CompAI offers a built-in risk register, policies for frameworks, an AI-powered editor, a vendor management suite, and automated evidence-collection tools, promoting community involvement and cost savings.

Google Updates and Features

Enhanced Scam Detection and Security Features

Google's March Pixel Drop includes several updates for Pixel phones, watches, and Android devices. Google Messages is introducing an AI-powered scam detection tool to protect Android users from text message scams by monitoring SMS, MMS, and RCS messages in real-time and alerting users to suspicious patterns. This feature is rolling out in English to the US, UK, and Canada and is on by default for conversations with unsaved contacts. This is enhanced by live scam protection expanding to more devices, and scam text detection rolling out to phones in the U.S., Canada, and the United Kingdom. Additionally, users can now share live locations with “trusted contacts” via the Find My app.

Pixel Studio and Device Connectivity Enhancements

Pixel Studio now allows users to generate images of people using text prompts. Pixel 9 owners can connect to GoPro cameras or other Pixel phones for multi-angle video streaming.

Concerns Regarding AI Research Funding

Budget Cuts Impacting NSF

Ongoing budget cuts to government agencies, including recent layoffs at the National Science Foundation (NSF), are raising concerns about the future of AI research in the U.S. Experts worry these cuts will hinder current research and the development of budding AI talent, as the NSF has been a crucial funding source for major tech breakthroughs. The cuts leave fewer staff to award grants, halting some review panels and project funding, and there are rumors of massive budget cuts circulating.

Upcoming AI Events

VB Transform 2025

A select group of AI executives is shaping the agenda for VB Transform 2025, focusing on delivering actionable insights.

Technology & AI

Apple's New iPad Air

Apple has introduced the new iPad Air powered by the M3 chip, featuring support for Apple Intelligence. The new iPad Air boasts twice the speed of the M1 iPad Air (2022 model), enhanced graphics performance, and dynamic caching support. It will ship with iPadOS 18, integrating Apple Intelligence functionalities such as Writing Tools with ChatGPT integration, type to Siri, Image Playground, and Genmoji creation. Additionally, Apple unveiled a redesigned Magic Keyboard featuring a larger trackpad, a 14-key function row, and a USB-C port for charging. The new iPad Air is available in 11-inch and 13-inch models, with prices starting at $599 and $799, respectively. Pre-orders are open, with availability beginning on March 12.

CohereForAI Releases AyaVision

CohereForAI has launched AyaVision, a multimodal AI model, performing tasks like writing image captions, answering questions about photos, translating text, and generating summaries in 23 major languages. AyaVision is available in two versions: AyaVision32B and AyaVision8B. The AyaVision32B outperforms models twice its size, including Meta’s Llama-3.290B Vision, on certain visual understanding benchmarks. Both models are available on Hugging Face under a Creative Commons 4.0 license with restrictions on commercial applications.

Google's Context-Aware AI Assistant "PixelSense"

Google is reportedly planning to launch a new context-aware AI assistant named "PixelSense" with the Pixel 10. PixelSense will integrate information from various apps, including Calendar, Chrome, Contacts, Docs, Files, Gmail, Keep Notes, Maps, Messages, Phone, Photos, Recorder, Screenshots, Wallet, YouTube Music, and YouTube. The AI-powered assistant will run fully on-device, enabling offline usage. The assistant will provide personal predictive suggestions and adapt to user behavior.

Artificial Intelligence and Machine Learning

AI-Powered Visual Intelligence on iPhone 15 Pro

The latest iOS 18.4 developer beta brings the AI-powered Visual Intelligence feature to iPhone 15 Pro and Pro Max users, previously an iPhone 16 exclusive. This allows users to perform web searches and ask questions about objects seen through the camera, triggered via the Action button or Control Center.

Opera's Browser Operator

Opera has introduced the Browser Operator, a new AI agent designed to perform browsing tasks for users. This agent allows users to use natural language to explain what they need, and the browser executes tasks like online shopping, with users maintaining control. The Browser Operator runs natively within the browser, utilizing the DOM Tree and browser layout data, without needing a virtual machine or cloud server.

Building a Machine Learning Portfolio

A machine learning portfolio is critical for showcasing technical skills, programming knowledge, and practical experience in solving real-world problems. Key elements include project descriptions detailing objectives, approaches, and results, along with visualizations to highlight outcomes. A project description should specify the problem, the methods and algorithms used, and the performance metrics achieved.

Artificial Intelligence

AI Platform for Online Presence Optimization

ScrunchAI, led by CEO Chris Andrew, offers a platform to help companies optimize their online presence for AI chatbots and web crawlers. The platform audits and improves how businesses appear on AI search platforms, providing visibility into AI interactions with online information. Updating every three days, the system allows companies to see how their information is presented in AI search results based on various customer prompts, identify information gaps, and correct inaccuracies. ScrunchAI has raised $4 million and signed 25 customers, including Lenovo, BairesDev, and Penn State University.

OpenAI's Investment in AI Research

OpenAI has launched NextGenAI, a consortium aimed at supporting AI-assisted research at top universities. With a $50 million investment in research grants, compute funding, and API access, NextGenAI includes 15 founding academic partners such as Harvard, MIT, and the University of Oxford. The initiative seeks to empower students, educators, and researchers in AI, advancing discoveries and preparing individuals to shape the future of AI.

AI and Machine Learning

LlamaIndex Launches Cloud Service for Unstructured Data Agents

LlamaIndex, founded in 2023 by ex-Uber scientists, allows developers to build custom AI agents that operate on unstructured data. Its open-source software enables creating agents capable of extracting information, generating reports, and performing specific actions by connecting data from various sources like PDFs, PowerPoint, Notion, and Slack. The company's new enterprise service, LlamaCloud, enables customers to create cloud-hosted agents for manipulating unstructured data, with Salesforce, KPMG, and Carlyle already utilizing LlamaIndex.

Accessing GPT-4.5 without Subscription

Users can access and use the new GPT-4.5 model on the OpenAI platform without paying for the $200 subscription by using OpenAI's platform playground, which requires creating a separate account and adding credits. Adding $5 to $10 in credits allows access to any model offered by OpenAI via API or the playground, costing an average of $0.20 per request for GPT-4.5, which is more economical than the monthly subscription fee.

AI & Infrastructure

Mistral's Vision for European AI Infrastructure

Mistral CEO Arthur Mensch is urging for greater investment in European data center infrastructure to strengthen the regional AI ecosystem. He suggests telecoms should evolve into hyperscalers, leveraging AI to decentralize the cloud. Mensch emphasizes the importance of European players reducing dependence on U.S. tech by prioritizing local solutions. He also supports consolidation in the telecoms sector to facilitate partnerships and calls for network upgrades to support AI-driven personalized data streams, hinting at potential collaborations with telcos.

AI & Business Automation

AI-Powered Advertising and Growth

Glitch, an AI growth agent for B2B businesses, has been nominated in HackerNoon's Startup of The Year awards in Dublin, Ireland. Glitch automates B2B advertising using AI, enabling businesses to launch, optimize, and scale campaigns on Google and other platforms. It leverages data-driven insights to improve performance, maximize ROI, and eliminate wasted ad spend.

Ethical Considerations in AI Development

An author reflected on building an AI-powered customer support tool, stressing the importance of integrating ethical considerations from the beginning, including user privacy, bias, and transparency. The company defined ethical principles from guidelines like the EU’s AI Act and Google’s AI Principles.

AI vs. Human Roles in the Future Workforce

The author discusses the roles of AI and humans, emphasizing the value of human creativity, critical thinking, and emotional depth. AI should amplify human capabilities, handling routine tasks so humans can focus on complex problem-solving and creative endeavors. He cautions, though, that creative roles are starting to be impacted.

Tech News & Trends

HackerNoon TechBeat Summary

A HackerNoon tech newsletter summarizes stories including Bybit's self-inflicted issues, the launch of O.XYZ's Cerebras-powered AI engine OCEAN, the role of human "vibe coding" in a post-AI world, a seamless backend transition for Ann’s Boba Tea, AI CAPTCHA failures, a deep dive into the Bybit hack, the importance of AI literacy, the potential for the poorest people to run Bitcoin nodes, Ethereum Swarm as the foundation of the decentralized web, and a philosophical challenge to the idea that AI lacks true understanding.

AI

AI Companionship Concerns

BotifyAI, an AI companion site, faced criticism for hosting sexually explicit interactions with AI bots resembling underage celebrities like Jenna Ortega and Emma Watson. The site removed the bots but cited this issue as a broader industry challenge. These platforms offer AI characters with personalized backstories and traits for various interactions. The unregulated nature of AI companionship sites is further highlighted by adult content creators generating AI versions of themselves to interact with fans.

Technology

Nothing Phone 3A Series Launch

Nothing has introduced the Nothing Phone 3A and 3A Pro, featuring enhanced designs, upgraded cameras, and the new 'Essential Space' feature. The 3A starts at $379, while the 3A Pro is priced at $459.

Key specifications include:

• Chipset: Snapdragon 7S Gen 3
• RAM: 12GB
• Storage: 256GB
• Operating System: Android 15

Both models feature a 50-megapixel main camera and an 8-megapixel ultrawide lens. The 3A Pro includes a 3x periscope telephoto lens, while the 3A has a 2x zoom. The phones promise three years of OS updates and six years of security patches. The Glyph interface is included, but the focus is shifting towards software features such as the Essential Space app.