-
WSJ: Taiwan Steps Up Probe Into AI Hardware Smuggling (Jun. 30, 2026)
Taiwan prosecutors probed alleged smuggling of Super Micro AI servers with Nvidia chips to China, summoning six people on charges of forgery and breach of trust. -
WSJ: How Companies Are Managing AI Token Spend (Jun. 30, 2026)
Companies face rapidly rising AI costs as broad employee use, always-on agents, and larger models drive token consumption. CIOs use cloud-era tools, including FinOps, dashboards, token caps, showback/chargeback, and smaller or open-source models, to control spending. -
Barrons: This AI Power IPO May Already Be Pricing In the Good News (Jun. 29, 2026)
Innio’s AI-power IPO already prices in the good news, leaving little upside. Wall Street sees strong demand for its gas-powered data center business, and analysts say the optimism is largely baked in. -
WSJ: 3 Economists Answer the Question: Is AI Going to Kill Jobs? (Jun. 28, 2026)
Economists disagree on AI’s impact: some predict major job, wage, and labor‑share losses, others say AI will complement work and create new roles. -
Jack Clark: Import AI 462: Superpersuasion; self-sustaining AI; paths to ASI (Jun. 22, 2026)
Across four experiments with 6,923 people, state-of-the-art AI systems consistently out-persuaded expert humans, raising more donations and shifting opinions, mainly by producing more information faster. -
WSJ: Inside the Room Where America’s Brightest Game Out How to Avoid an AI Apocalypse (Jun. 13, 2026)
Forty experts ran a “Paper Prosperity” scenario showing AI could boost GDP yet increase underemployment, unrest, and erode the social contract. -
WSJ: Chatbots Are Replacing Therapists With Little Scientific Evidence Behind Them (Jun. 29, 2026)
People are turning to AI chatbots for mental-health support, using therapy-focused bots like Tee and Ebb for coaching and crisis detection. -
Robusta: You really shouldn't copy-paste errors into Claude Code — Robusta Blog (Jun. 28, 2026)
Don’t copy-paste errors from Claude Code; give the agent safe, isolated access—staging databases, headless browsers, API keys—so it can test end-to-end, reproduce failures, and fix them autonomously. -
NY Times: A.I. ‘Employees’ Might Disrupt Work in Unexpected Ways (Jun. 29, 2026)
Companies are treating A.I. agents as employees, but managers check A.I. work less carefully, creating risks. A.I. systems also favor A.I.-written work, push overly rational choices, and can create unseen, self-reinforcing biases. -
Ars Technica: South Korea to spend $1T on more memory chip production and humanoid robots (Jun. 29, 2026)
South Korea is backing physical AI, funding a world-model, Hyundai pledged $5.8B for robot factories, and plans humanoid deployment by 2028. -
WSJ: OpenAI Limits Access to New Models, Citing Government Security Concerns (Jun. 26, 2026)
OpenAI will limit initial access to GPT-5.6 to a small group approved by the Trump administration, offering models for reasoning, everyday tasks, and heavy workloads. -
The Transmitter: Neuroscience News and Perspectives: AI can’t solve the brain without data that fit together (Jun. 28, 2026)
Brain foundation models arose where years of data integration, standards, and infrastructure existed, not merely from bigger models. -
The Texas Tribune: San Marcos ban data centers, testing its local control (Jun. 30, 2026)
San Marcos became the first Texas city to ban data centers within city limits, using home-rule zoning to block developments over water, energy, and noise concerns. -
Noah Smith: Will AI make companies outsource more, or less? (Jun. 29, 2026)
Before the pandemic, many worried that US business activity was falling, with fewer startups and small businesses being started.