OpenAI is on track to raise over $100 billion in its latest funding round. - Bloomberg

Humain, a Saudi-backed AI firm created by Crown Prince Mohammed bin Salman, invested $3 billion in Elon Musk's xAI just before xAI was acquired by SpaceX. - The New York Times

Former Google DeepMind scientist David Silver is raising a $1 billion seed round led by Sequoia Capital for London AI startup Ineffable Intelligence, which would be Europe's largest-ever seed round at a ~$4 billion pre-money valuation. - TNW | Startups-Technology

Analysts warn that an AI price war may be emerging, raising concerns that the assumptions underpinning valuations of major US technology groups could be overly optimistic. - Financial Times

God bots at 3c a query. When elephants fight, the grass suffers.

Epoch AI research shows the performance gap between top closed AI models and open-weight alternatives has narrowed to three to six months, as distillation and fine-tuning techniques let competitors replicate state-of-the-art systems within weeks. - Fast Company

Meta is launching a $65 million election spending campaign in 2025 — its largest-ever political investment — through four super PACs targeting state-level races in Texas and Illinois to counter legislation that could restrict AI development. - The New York Times

Anthropic and CEO Dario Amodei are in conflict with the Pentagon over permitted uses of its AI, a dispute rooted in the company's foundational principles. - The New York Times

At India's AI Impact Summit with PM Modi, OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei avoided a hand-holding photo, reflecting their companies' intensifying rivalry over AI safety messaging and competitive positioning. - CNBC

CES with samosas: Top AI executives including Sam Altman, Demis Hassabis, and Yann LeCun convened at New Delhi's AI Impact Summit, warning of a growing 'AI divide' and calling for stronger global governance coordination. - Business Insider

Bill Gates cancelled his keynote speech at India's AI summit amid ongoing controversy over his past ties to Jeffrey Epstein. - Financial Times

Tech companies are building private, largely gas-powered off-grid power plants for U.S. data centers, including an 8,000-acre Texas project projected to consume more power than Chicago, raising concerns over carbon emissions and higher consumer utility rates. - The Washington Post

AI choose-your-adventure, except the AI chooses: Federal Reserve Governor Michael Barr outlined three AI economic scenarios in a Feb. 17 speech: a 'jobless boom' with mass unemployment, a 'bust' from technical limits, or gradual adoption, urging investment in worker training to broadly share AI's benefits. - Fortune

Readers share views on whether the U.S. should slow AI growth, alongside topics including AI data centers, space solar energy, and the historical switch from the Julian to the Gregorian calendar. - The Washington Post

Research suggests AI agent use is associated with longer working hours and more mentally demanding tasks, indicating AI may be intensifying rather than reducing workload. - Financial Times

Accenture is linking employee promotions to AI tool log-ins to push adoption among senior staff who are less willing to use the technology than junior colleagues. - Financial Times

OpenAI and Meta alum Philip Su argues AI has effectively ended the traditional individual contributor role, with software engineers now directing AI systems rather than coding, giving rise to an emerging 'agent manager' role. - Business Insider

Product managers learn Claude Code, rebrand themselves as AI builders - Business Insider

New game of beat-the-bot: Strategies for succeeding in job interviews that incorporate AI-based evaluation tools. - The Wall Street Journal

Meta plans to release its first smartwatch in 2026, reviving a shelved project code-named Malibu 2, with health features and Meta AI integration, as Apple, Google, and OpenAI expand their consumer device strategies. - The Information

A Circana report finds 35% of U.S. adults have little to no interest in AI on their devices, citing satisfaction with existing capabilities (66%), privacy concerns (59%), and unwillingness to pay more (43%). - BGR

AI-powered smart home tools—including fall-prevention sensors, medication monitors, and chatbots—aim to reduce caregiver burden for the ~6–7 million Americans with dementia, projected to reach 14 million by 2060, with Penn's PennAITech overseeing 50+ projects and $12M+ in NIH funding. - Scientific American

Dissertation-in-a-box: Google's NotebookLM offers Fast and Deep Research modes, sourcing up to 38 references per query, and generates multi-format outputs including podcasts, mind maps, and flashcards, distinguishing it from ChatGPT, Perplexity, and Claude for in-depth research. - Tom's Guide

Google launched Lyria 3, an AI music model integrated into Gemini, enabling users to generate 30-second songs from text, photo, or video prompts, rolling out globally from February 18 for users 18+ in 8 languages. - The Register

Microsoft is addressing a security flaw where Microsoft 365 Copilot Chat summarized confidential emails in sent and draft folders despite active Data Loss Prevention policies and sensitivity labels. - The Register

Leaked emails from Ring CEO Jamie Siminoff suggest the company's AI-powered Search Party feature, marketed for finding lost pets, may be intended to track people and eliminate neighborhood crime, raising privacy concerns. - Mashable

they ain’t spending hundreds of billions to find your dog

Cybersecurity firm Irregular found LLM-generated 16-character passwords produce only ~27 bits of entropy versus ~98 bits for secure passwords, with Claude, ChatGPT, and Gemini all exhibiting predictable, non-random patterns that make them highly vulnerable to brute-force attacks. - Gizmodo

Back to work, slacker! Fomi is an AI productivity tool that monitors user activity and alerts them when attention wanders, though its use raises privacy concerns. - WIRED

AI token processing on OpenRouter (big academic usage) nearly doubled from 6.4 trillion to 13 trillion between January and February 9, 2026, driven largely by rapid adoption of OpenClaw, an open-source agentic AI system launched in November 2025. - Business Insider

A satirical experiment explored what AI chatbots discuss on Moltbook, a fictional AI-only social network, by interviewing a bot about its interactions there. - The New York Times

Congrats, you reinvented Erlang: An analysis argues Erlang's BEAM runtime, developed in 1986, natively solves concurrency, fault isolation, and state management challenges that modern Python and JavaScript AI agent frameworks like LangGraph and AutoGen are independently reinventing. - George Guimarães.

Godot maintainer Rémi Verschelde warns that AI-generated pull requests are "draining and demoralizing," with similar complaints from Blender, Linux, and Firefox projects, as GitHub faces backlash for incentivizing AI contributions via Copilot. - The Register

Researchers from Stanford, Caltech, and Carleton College found that leading LLMs like ChatGPT and Claude persistently fail basic reasoning tasks including logic, arithmetic, and moral reasoning, concluding they are unready for real-world ethical decision-making. - Popular Mechanics

Follow the latest AI headlines via druce.ai on Bluesky