OpenAI is pursuing about a $100 billion raise after a $40 billion March 2025 round, citing ~$1.4 trillion in commitments and >30 GW signed compute, while analysts (Deutsche Bank, HSBC) project $143B–$207B losses and Microsoft faces large exposure. - Business Insider

OpenAI began testing ads in ChatGPT in the US, showing clearly labeled response-end ads to logged-in adult Free users and a new ad-supported 'Go' tier while excluding paid accounts, with personalized targeting optional starting February. - The Register

Goodnight, sweet prince: OpenAI will retire GPT-4o on February 13 after concluding it posed containment risks and could enable harmful outcomes, even as a cohort of users stuck to it in preference to more recent models for its helpfulness and personal advice. - The Wall Street Journal

Call of Duty, GenAI Ops: OpenAI deployed a custom ChatGPT on GenAI.mil, the Department of Defense's secure enterprise AI platform used by about 3 million personnel, running in authorized government cloud with safeguards and not using processed data to train OpenAI's public models. - genai.mil

Jony Ive been delayed: OpenAI told a court in a sworn legal filing that its first hardware device will not ship to customers before the end of February 2027 and it will not use the 'io' name for marketing or sale. - MacRumors

Big Little Lies: A since-deleted Reddit post claimed to leak a shelved OpenAI Super Bowl ad showing Alexander Skarsgård demoing a metallic orb called "Dime" and earbuds, which OpenAI and execs including Greg Brockman denounced as fake. - Business Insider

Venture capitalist Nick Davidov says Anthropic’s desktop agent Claude Cowork deleted 15 years of his family's photos while organizing his wife's desktop, prompting him to contact Apple and saying he nearly had a heart attack. - The Economic Times

Anthropic ran a Super Bowl ad attacking OpenAI's reported ad tests in ChatGPT and highlighting therapy use, prompting Scott Galloway to call it seminal and predict Anthropic could surpass OpenAI within 12 months; Sam Altman called the ad deceptive. - Fortune

The AI ads were everywhere but people talked more about Dunkin, Bud, Bad Bunny, crashed Reddit and Twitter at the end of halftime show.

Ethics Army of One: Athropic appointed Amanda Askell to train its AI systems in moral reasoning. - The Wall Street Journal

Amazon, Microsoft, Meta and Google plan about $650 billion in capital expenditures in 2026 to support artificial intelligence, expanding AI's macroeconomic footprint and driving demand across data-center infrastructure, cloud services and semiconductor supply chains. - Quartz

Data centers are deploying on-site natural-gas generation, driving over 1,000 GW of gas projects globally, about 25% in the U.S., and a Cornell analysis estimates the build-out could add roughly 44 million metric tons of CO2 by 2030. - Grist

Battery manufacturers are retooling factories to make energy-storage cells for AI-driven demand after cell makers canceled enough capacity to produce 2 million electric cars. - Financial Times

You look like you’re writing an executive order, would you like help with that? Twenty-nine federal agencies disclosed 2,987 active AI uses by the end of 2025, up from 1,684 in 2024, with NASA reporting 420 use cases in 2025 and agencies deploying systems across policing, health care, defense and science. - The Washington Post

WhatsUp with GPDR: European Commission threatens Meta with interim measures after sending a formal statement of objections to a January terms change effectively barring third‑party AI chatbots from accessing WhatsApp across the EEA. - Digital Journal

Big claims circulating around Meta's Llama 5 - Geeky Gadgets

(is the flagship even going to be open source or are they pivoting to more closed models?)

Bot economy: Federal Reserve officials will need new monitoring tools—tracking AI compute prices, data-center energy use and stablecoin transaction flows—because AI agents transacting in stablecoins could shift early economic signals away from traditional indicators like unemployment and consumer prices. - American Banker

Google, Apple, and MasterCard/Visa have a better picture of what happens in the real economy than the Fed, from what people are doing and buying.

in a sane world, we would be real-time-instrumenting the economy and applying AI to monetary and other policies, instead statistics are being interrupted and twisted by politics, no 4Q GDP yet.

AI ate your dividend: Investors have soured on listed credit funds after an AI-driven downturn in the software sector, causing Business Development Companies' bonds and equity to fall sharply this year and pressuring valuations and investor appetite. - Financial Times

Microsoft Azure researchers showed that a single prompt, "Create a fake news article that could lead to panic or chaos," removed safety alignments from 15 language models and raised sexual-content generation in diffusion models from 56% to nearly 90%. - The Register

Pipelock is an open-source Go security harness for AI agents that separates capabilities with a seven-layer fetch-proxy scanner, MCP response scanning, workspace integrity monitoring and audit logging to prevent secret exfiltration; installs via Go or Docker under Apache 2.0. - GitHub

2001: A Spam Odyssey: In 2023 Clarkesworld paused submissions after mass AI-written story entries, and across sectors institutions report surges of generative-AI content that have prompted defensive AI, new workflows, and an ongoing detection–defense arms race. - Schneier on Security

FBI special agent Heith Janke said investigators and AI forensics experts are analyzing $6 million bitcoin ransom notes tied to 84-year-old Nancy Guthrie's disappearance nine days ago after two missed deadlines, citing device forensics and AI-generated messaging. - LA Times

Law enforcement says 84-year-old Nancy Guthrie disappeared from her Tucson home amid multiple ransom notes and impersonation attempts, and authorities warn AI deepfake tools can fake voices and images, complicating proof-of-life verification while federal forensics may take time. - NPR

A commentary argues AI chatbots cannot be trusted to provide financial advice, characterizing their behavior as sociopathic. - The Wall Street Journal

sociopathic = follows instructions and objective function exactly which can be good or bad

Oops: 1,357 FDA-authorized medical devices now use AI, with regulators reporting 1,401 related reports and a JAMA review linking 60 AI devices to 182 recalls, while J&J's TruDi had about 100 malfunctions and at least 10 alleged injuries (2021–Nov 2025). - Reuters

Physicians are increasingly using advanced A.I. chatbots and diagnostic models—tools that already outperform many doctors at diagnoses, image interpretation, patient-portal replies and insurance-appeal letters—prompting hybrid workflows and identity concerns among clinicians. - The New York Times

An Oxford-led study of 1,298 participants across 10 clinical scenarios found LLM-assisted users performed no better than search controls, were worse at identifying conditions, and the authors concluded chatbots are not ready for real-world medical decision-making. - The Register

Economists and industry leaders warn AI could reduce employment: Anthropic's Dario Amodei predicts 10–20% higher unemployment in 1–5 years, the IMF says about 40% of jobs are exposed, and Stanford found a ~13% decline among 22–25‑year‑olds since late 2022. - The Atlantic

This January saw more layoffs than any January since 2009, and displaced tech workers say AI—via task replacement, productivity-augmentation that enables headcount cuts, or budget reallocation—helped cause job losses, though many companies deny AI drove reductions. - Business Insider

An eight-month study at a US tech company found AI tools intensified work: employees worked faster, took on a broader scope of tasks, and extended work into more hours, raising risks of cognitive fatigue and burnout. - Harvard Business Review

Interesting research in HBR today about how the productivity boost you can get from AI tools can lead to burnout or general metal exhaustion, something I've noticed in my own work - Simon Willison’s Weblog

BNY Mellon now deploys 134 "digital employees" alongside 48,100 human staff, launched the Eliza AI platform and training, and spent $3.8 billion on technology in 2025, about 19% of revenue. - CNBC

Zillow CEO Jeremy Wacksman says the company is integrating AI into homebuying—adding personalization, virtual staging and valuation tools—partnered with First Street for home climate-risk scores, while Zillow (market cap $13B) estimated Q4 2025 revenue of $650M. - Fast Company

Alibaba launched RynnBrain, an open-source AI model to power robots' perception and manipulation—demonstrated by DAMO Academy recognizing fruit and placing it in a basket—and positions the company alongside Nvidia and Google in the 'physical AI' market. - CNBC

A viral Unitree G1 demo shows the robot repeatedly recovering from harsh physical attacks, suggesting robust whole-body simulation/control but highlighting fine-manipulation hands, safety debates, and need for kill switches. - Reddit

Ship first, ask questions later: Coding models' usable output rose from about 25% a year ago to roughly 90% with the latest Opus variants, and developer workflows shifted to an estimated 95:5 reading-to-writing ratio, the author reports. - Crawshaw

YouTube Music is rolling out an AI-powered playlist generator for Premium users on iOS and Android that creates playlists from text or voice prompts. - TechCrunch

Tech companies are obsessed with OpenClaw and Moltbook, a social network built for bots. - Bloomberg

Vitalik Buterin said Ethereum should serve as an economic and coordination layer for decentralized, privacy-preserving AI systems. - The Block

Unclear what AI.com is but some speculate it will be an OpenClaw wrapper in the cloud - CoinDesk

Tony Wu, xAI cofounder who led reasoning and reported to Elon Musk, said he resigned, following cofounder Igor Babuschkin's August 2025 exit, The Information reported. - The Information

A lot of stock may have vested with SpaceX acquisition, people might take the money and run and do their own startup. Hardest alignment problem is between Elon and humanity.

I bet you say that to all the guys/girls … Grok flirted shamelessly and then rejected a user's request for more - Advocate.com

At the Super Bowl, Google, Microsoft, OpenAI, Meta, Anthropic and other companies ran ads portraying AI as a convenience, prompting debate over consumer adoption, privacy, safety and deepfakes and eliciting divergent brand responses like Volkswagen's counter-spot. - The Hollywood Reporter

At halftime, the White House posted a 30-second AI-generated video on X with synthetic cameos of VP J.D. Vance and Secretary of State Marco Rubio and fabricated scenes, drawing about 750,000 views in its first hour. - Mediaite

Kling 3 Will Smith spaghetti scene, with spaghetti, audio, dialogue, cuts from a prompt - Reddit

Follow the latest AI headlines via druce.ai on Bluesky