AI Reading for Monday October 13

In Big Tech's knotty game of AI Twister, what happens when a giant slips? As firms forge interdependent AI partnerships—OpenAI's $300 billion compute deal with Oracle, Meta's $10 billion Google Cloud contract, and Nvidia's $100 billion investment in OpenAI—vendor-financing poses risks. - Business Insider

Explaining tech deals - Reddit

(This is how money is supposed to work, right?)

Microsoft, OpenAI staff and its nonprofit arm hold nearly 90% of startup equity; a recent Nvidia deal will dilute shareholders and complicate investor payouts. - Financial Times

OpenAI has filed antitrust complaints with European Union regulators against several U.S. tech companies, seeking regulatory intervention to curb chatbot competition by appealing to European authorities instead of U.S. regulators. - The Washington Post

Leaked beta Android code-name Calpico suggests OpenAI is adding user-to-user DMs, including group chats, assistant renaming, and blocking; experts warn the feature raises privacy risks because OpenAI currently uses chat data for training and lacks E2EE. - BGR

Individuals are increasingly using OpenAI's ChatGPT and Google's Gemini to manage debt, improve saving strategies and obtain investment advice, reflecting a shift toward AI-driven personal finance support as an alternative to traditional advisors. - The New York Times

Salesforce's Agentforce 360 is a bet that every enterprise will want to have a fleet of AI agents working with humans. Agents, including voice and Slack agents, will reason, take actions and adapt to a company’s culture, voice and business logic. - SiliconANGLE

Slack is revamping Slackbot into a personalized AI assistant that creates plans, schedules meetings, finds files and searches workspace content, integrates with Outlook and Google Calendar, has been deployed to 70,000 Salesforce employees and will roll out broadly by year-end. - The Verge

We launched an AI agent and it flopped. Here’s what we learned—and why we're trying again - Fast Company

Moved from chatbot to agent because:

1. The AI isn’t the product, the outcome is.

2. The product must map to real workflows 

3. Don’t make me think. The best AI is invisible.

AI automation engineers embed within teams to build AI-powered workflows using tools like Zapier, ChatGPT, Airtable and Cursor, prototyping and scaling solutions (e.g., automating recruitment messages, interview summaries) to improve efficiency and free staff for strategic work. - Fast Company

The agentic solutions and vertical AI apps space is getting pretty crowded. - IT Pro

A study analyzing over 14 million website visits finds AI browsing constituted under 1% of daily online activity, mostly ChatGPT used for academic or professional tasks, and higher use correlated with narcissism, Machiavellianism and psychopathy. - psypost.org

AI Is Juicing the Economy. Is It Making American Workers More Productive? - The Wall Street Journal

AI may widen pay gaps by amplifying superstar productivity, experts warn - The Wall Street Journal

No U.S. manufacturing boom as it lags artificial intelligence growth, a divergence experts say could reshape workers' prospects and wealth distribution, and requires investment in AI retraining and infrastructure to sustain economic growth. - The Washington Post

Klarna’s CEO says AI cut their workforce from 7,400 to 3,000 — without layoffs. He warns tech leaders are “sugarcoating” AI’s impact: mass job loss is coming. 2/3 of support now run by bots. Fewer humans, higher pay, similar cost. “Let’s be honest about what’s next.” - Fortune

Just-in-time app generation, what if your app could let you describe a feature and just add it? - Hacker Noon

Google's CodeMender is a Gemini-powered agent that autonomously detects and fixes software vulnerabilities using static and dynamic analysis, fuzzing and SMT solvers, and has patched 72 flaws in large open-source projects. - SlashGear

London-based autonomous driving startup Wayve is in early talks with Microsoft and SoftBank to raise $1–2 billion, seeking roughly an $8 billion valuation to accelerate its AI-driven vehicle technology development. - Financial Times

The Dutch government has taken control of Chinese-owned chipmaker Nexperia under the Goods Availability Act to safeguard chip supplies for European industries, a move that escalates tensions with China over access to advanced semiconductor technology. - Financial Times

Nearly half of top AI talent is Chinese, yet rising U.S. visa restrictions curb their flow to Silicon Valley—where Meta, Google and Nvidia rely on them—while China’s new K visa seeks to attract foreign scientists. - Rest of World

Anduril Industries launched EagleEye, a suite of AI-enabled helmets, visors and glasses using Anduril’s Lattice software and Meta’s augmented-reality waveguide tech, winning a $159 million U.S. Army contract. (so we meet again, Mr. Zuck and Mr. Luckey) - Business Insider

Like with chemical and biological weapons, all states should agree on ethical, legal best practices and international collaboration to manage risks and ensure responsible defense-sector AI use. - Financial Times

Meta-appointed adviser Robby Starbuck, appointed in August under a lawsuit settlement, keeps spreading disinformation about shootings, vaccines, transgender people and political violence, prompting criticism from anti-extremism and LGBTQ+ groups. - the Guardian

Jacob Silverman's book Gilded Rage says some Silicon Valley elites have shifted to openly embracing Donald Trump, deepened ties to the U.S. defense sector and AI investors, and pursue 'exit' sovereignty strategies like charter cities. (and apocalypse bunkers) - Politico

Assembly language is still a thing: your right to peacefully assemble may not be infringed by AI - WIRED

Police departments increasingly use artificial intelligence to analyze data from crime scenes, digital surveillance and other sources, enabling faster processing while raising civil-rights and privacy concerns and prompting debate over ethics and oversight. - Slate Magazine

Schneier: AI reshapes U.S. politics by automating campaign tasks like fundraising, targeting, and data analysis, aiding organizers' deliberations and recruitment; lack of regulation and oversight raise misinformation and surveillance risks. (repeat of American Prospect essay) - Schneier on Security

How generative AI can reduce climate skepticism and increase support for climate action, presenting opportunities for policy, education and communication while posing risks including misinformation and ethical concerns. - Nature

Hollywood demanded copyright cuardrails from Sora 2 - users complained that's no fun - Yahoo

Monica Hesse tested an AI tool to animate her deceased grandmother and found the experience distressing, highlighting technical shortcomings and ethical, emotional and grief-processing concerns about AI-driven resurrection of the dead. - The Washington Post

Aheadform's Origin M1 reignited the uncanny valley debate after studies, including from the University of Castilla-La Mancha, found moderate anthropomorphism boosts trust while excessive realism reduces comfort; the global service-robot market is projected to top $293 billion by 2032. - Decrypt

John Searle, Philosopher Who Wrestled With A.I., Dies at 93 - The New York Times

Follow the latest AI headlines via SkynetAndChill.com on Bluesky