OpenAI: Industrial policy for the Intelligence Age - Ideas to keep people first. - OpenAI

OpenAI released a policy blueprint urging a public wealth fund, taxes on automated labor and a shift toward corporate and capital-gains taxation, incentives for employer trials of four-day workweeks with no pay loss, and accelerated electricity-grid expansion. - Business Insider

Altman and CFO Friar said to be at odds on IPO, with her reporting to Simo who is now on leave, and excluded from some meetings. - The Information

sounds like Altman is intent on banking it ASAP and thinks things are going just great, but if you make a deck and a roadshow and it raises a lot of questions and red flags and IPO investors demand a big haircut or it falls post-IPO, it’s very damaging to the company’s long-term prospects.

Documents: OpenAI and Anthropic have the same challenge: funding giant computing costs, as they project profitability to investors with and without training costs, and report inference costs exceeding half of revenue. - The Wall Street Journal

Ronen Farrow on the OpenAI beat: New interviews and closely guarded documents shed light on the persistent doubts surrounding Sam Altman, with some calling him a liar and a sociopath. - The New Yorker

IRGC spokesman Brig. Gen. Ebrahim Zolfaghari warned in a video of "complete and utter annihilation" for OpenAI's reportedly $30 billion, 1GW Stargate AI data center in Abu Dhabi, citing satellite imagery and threats to other U.S. tech firms. - Tom's Hardware

Silicon Data's indexes show GPU prices rising across generations, with Neo Cloud H100 up ~20% (2.20→2.64) and B200 up ~22% (4.40→5.35) over three months, while hyperscalers command roughly 3x rental premiums. - Business Insider

Former crypto-mining firms pivot to AI data centers, growing market cap from $2.1 billion to $48.5 billion - Business Insider

OpenClaw surged to over 150,000 GitHub stars and runs locally with deep system access, while Google’s Antigravity and Anthropic’s Claude Cowork automate coding and domain workflows, prompting warnings about increased data‑leak, security and governance risks. - VentureBeat

Kuse CEO Xiankun Wu built OpenClaw-based autonomous AI employees that run 24/7, handle an estimated 60–70% of work and boost team output 3–10×, prompting a human-only Slack channel to curb AI-driven task overload. - Business Insider

Tech founder quits startup to homestead in British Columbia, building OpenClaw AI assistant to monitor 22-acre property - Business Insider

AutoAgent, an open-source library by Kevin Gu (thirdlayer.inc), autonomously iterated an agent harness via a meta-agent loop and, in a 24-hour run, reached #1 on SpreadsheetBench (96.5%) and #1 GPT-5 on TerminalBench (55.1%). - thirdlayer.inc

Power user/shadow IT employees are independently piloting AI tools for task tracking, workflow automation and in-app assistance, producing efficiency gains that have prompted executives to consider formalizing training, elevating internal champions and developing governance to scale AI adoption. - Fast Company

How to set up your agents: The author configured iTerm2 profiles and a one-click arrangement to launch color-coded tabs and parallel Claude Code sessions that auto-load project context, run startup commands (cd, clear, claude), and provide status summaries for two app projects. - ZDNET

Writing a set of sqlite dev tools in Rust with an agentic assistant - Lalit Maganti

A practical guide says make AI the default—master prompts, build daily, use AI across development/design/content, and combine tools—so AI-native individuals and small teams can build full-stack applications in days and accelerate product and marketing. - C# Corner

An author describes a three-AI workflow using Claude for strategy and critique, Gemini for research and structured summaries, and ChatGPT for execution, claiming it reduces bottlenecks and rework while improving outputs. - Tom's Guide

Skynet for retail: AI-enhanced ceiling cameras now analyze foot traffic, shelf stock and checkout activity, used by 42% of retailers (Honeywell), with the global video-analytics market forecast to grow from $12.71B in 2024 to $37.84B by 2030. - BGR

Alibaba's Accio, launched in 2024 and trained on 26 years of proprietary transaction data, reached over 10 million monthly active users by March 2026 and lets small sellers compress weeks of product research and supplier sourcing into a single chat. - MIT Technology Review

Retailers are trying to solve the ridiculous problem of inconsistent sizing (do these run large or small?) to reduce fit-related returns running up to 15.8% of sales. - CNBC

Google added Gemini AI to Google Maps as an "Ask Maps" feature that plans daylong itineraries, finding playgrounds, kid-friendly restaurants, routes and lesser-known spots to aid local discovery and reduce decision fatigue. - The Verge

Microsoft Copilot parsed a 29-page Ohio homeowners policy to extract coverages, Google's Gemini generated a risk checklist flagging plumbing, electrical and termite concerns, and Anthropic's Claude identified a vole, illustrating AI can aid home maintenance but requires professional verification. - CNET

LM Studio 0.4.0 runs Google Gemma 4 26B model locally at 51 tokens per second on Apple Silicon - George Liu

Old Macdonald had an AI-AI-O: U.S. farm employment fell to 2.184 million in February 2026 (down 22,000 versus five years earlier) amid 38% of farmers aged 65+, and AI-enabled robotics are being adopted to address labor shortages and shift farmers toward strategic manager roles. - TechRadar

JPMorgan CEO Jamie Dimon warned in his annual letter that geopolitical conflicts (Ukraine, Middle East) and rising China tensions risk market volatility, criticized proposed Basel III/GSIB surcharges as adding up to 50% more capital, and flagged $1.8 trillion private-credit stress. - Quartz

AI is already causing rapid changes in how auditors review company accounts, prompting questions about whether regulators are prepared. - Financial Times

MIT FutureTech's study of 17,000+ expert evaluations found frontier AIs reached acceptable quality on ~50–75% of text tasks, hit 50% full-workday task success by Q3 2024 and project 80–95% by 2029, while under 19% of U.S. firms have deployed AI. - Yahoo

Between 2019 and 2023 generative models displaced entry-level translators, illustrators and copywriters, cutting translator incomes by about one-third while data-labelers were paid as little as $1.32–$3.74/hour and reported lasting psychological harm. - Hacker Noon

Your Work Trained the Model. The Model Replaced You. Philip K. Dick Wrote This Story in 1968.

A University of Pennsylvania study of 800 Taiwanese students found an LLM-based chatbot delivering personalized problem sequences outscored a fixed-sequence group on the final exam—equivalent to 6–9 months' schooling—and added about three minutes per problem; draft posted March 2026. - KQED

Fast Company reports that artificial intelligence is making cheating easier. - Fast Company

Is Medvi an AI-powered fraud machine? - The Potential Surface

Texas Gov. Greg Abbott reposted an image claiming to show the U.S. airman rescued after an F-15 shootdown over Iran, which a community note and observers flagged as AI-generated and the Pentagon had not released official photos. - Mediaite

Viral internet memes driven by algorithmic feeds have spread offline into slang, politics and institutional messaging, including DHS posts and a $100 million ICE recruitment effort. Pew found TikTok's top quartile produced about 98% of videos - The New York Times

Tab-complete-bomb: An exercise found an AI system 100x faster and 97% tactically viable versus 48% for humans, while drones, cyberoperations and battlefield AI reshaped conflicts and AI models face integration and governance challenges. Woe betide fools starting wars because they think AI gives them an edge. - Foreign Policy In Focus

Suno's AI platform Suno Studio (Premier, US$24/month) can be seeded with simply preprocessed copyrighted tracks to generate near-imitative covers of songs like Beyoncé's "Freedom", enabling export to streaming services and raising monetization and legal risks. - The Verge

In August 2025 Grammarly launched an "Expert Review" AI feature offering writing suggestions "inspired by" named figures, then removed the feature from review after user and expert backlash. - The Verge

Unbound AI launches uncensored adult video generator, raising deepfake and consent concerns - Unbound AI

Researchers trained cultured rat cortical neurons confined in 128 microfluidic wells (≈14.6 neurons/well) using a 26,400‑electrode array and real‑time FORCE learning to autonomously generate periodic waveforms. - Tom's Hardware

You're a n<redacted>, Harry - Reddit

Follow the latest AI headlines via druce.ai on Bluesky