Good morning. It's Tuesday, July 14, and we're covering ASML's clash with U.S. export claims, why enterprises may be giving away their AI advantage, and economists' growing warnings about AI's impact on jobs.

Plus: a poll on when AI providers become competitors, and a guest post discussing why AI is humanity's most transformative invention.

Nearly 200 economists and technology leaders warn that AI could reshape the economy faster than past technological revolutions, leaving policymakers unprepared. A new statement, We Must Act Now, urges governments and researchers to better understand AI's economic impact and labor disruption. Signatories include 15 Nobel laureates, OpenAI and Anthropic chief economists, Eric Schmidt, and Jack Clark, with concerns centered on rapid white-collar job displacement over the coming decade.

Most Americans favor giving the public a larger stake in AI companies rather than leaving AI gains with private firms. A Verasight survey finds 69% support requiring major AI companies to transfer 50% of their stock into a public sovereign wealth fund, echoing Senator Bernie Sanders' proposed legislation. The poll comes as Goldman Sachs estimates AI could temporarily displace 15 million U.S. workers over the next decade before creating new jobs.

Apple's long-term chip roadmap points to AI demands rather than faster consumer performance alone. Bloomberg reports future M6, M7, and M8 processors are being designed with increasingly powerful AI capabilities, with the rumored M7 Ultra supporting up to 1.5TB of RAM by 2029. The roadmap also includes new Apple Pencil models as Apple expands AI-focused hardware across its product lineup.

Every morning the Forward Future team argues about the stories, papers, and launches that matter. What's real, what's hype, and where AI is headed next. These are the public notes.

Monday's team traced one worry through the week's news: the companies that supply the models, chips, and compute want the app layer too. They saw it in a YC partner decamping for Anthropic, a data paradox flagged by Satya Nadella, and OpenAI's well-timed removal of Codex usage caps.

▸ What a YC managing partner's move to Anthropic signals
▸ Why showing your data to a model provider is riskier than showing it to a cloud
▸ Why OpenAI dropped Codex's usage window right as Anthropic users hit outages

Read Friday's Zeitgeist →

OpenAI has folded Codex into ChatGPT, shifting the app toward knowledge work while making traditional chat a secondary feature.

_{^{Guest contributor: Gopi Kallayil is Google's Chief Business Strategist for AI, advising leaders on AI, innovation, and business transformation. He is also a TEDx speaker, author, and former McKinsey consultant.}}

For over 12,000 years, human beings have been building tools to solve problems. It's what separates us from every other species on the planet. We build tools to improve the human condition, and we’ve been remarkably successful at it.

We discovered how to create fire, which let us cook food and stay warm. The wheel transformed transportation, while the printing press democratized knowledge. Steam engines powered the Industrial Revolution, electricity lit up the world, and then the internet connected us all. → Read the full article here.

ASML is pushing back against allegations from the Trump administration that one of its extreme-ultraviolet (EUV) lithography machines—the only systems capable of producing the world's most advanced chips—may have reached China. The Dutch company says it can account for all 340 EUV machines it has built, none are in China, and no EUV-specific components or equipment have ever been shipped there.

While U.S. officials have not publicly presented evidence, the dispute has intensified broader tensions over China's semiconductor ambitions, export controls, and the balance between national security and commercial interests. The debate also extends beyond EUV technology, with Washington seeking tighter restrictions on ASML's older deep-ultraviolet (DUV) systems, even as China advances its chipmaking capabilities using those tools and pursues its own domestic EUV technology. → Read the full article here. (Paywall)

A new concept dubbed the "Reverse Information Paradox" argues that AI has inverted economist Kenneth Arrow's classic information paradox: instead of sellers risking valuable knowledge, enterprise customers increasingly expose proprietary expertise every time they use AI models. The essay contends that prompts, workflows, and institutional context become valuable "learning exhaust" that can improve AI systems, raising concerns that organizations may unintentionally transfer competitive advantages to model providers.

It argues that enterprises should retain ownership of their AI-generated knowledge, build private learning environments, and maintain control over evaluations, model orchestration, and fine-tuned systems. The central claim is that, in the AI era, competitive advantage will come not just from protecting data but from protecting an organization's ability to continuously learn, adapt, and compound its own intelligence. → Read the full article here.

Meta Expands AI Data Center: Louisiana's Hyperion project grows to a $50B, 5GW AI hub, fueled by state tax incentives and infrastructure investment.

TSMC Hits Record Revenue: TSMC's Q2 revenue rose 36% to a record T$1.27T, beating estimates as AI chip demand continued to surge.

Anthropic Localizes Claude Pricing: Claude now offers rupee pricing in India, its second-largest market, though UPI payments remain unavailable.

Nolan Rejects AI Hype: Christopher Nolan says AI won't replace human creativity, calling that idea "nonsense" despite its growing role in filmmaking.

Hermes Triples to $1.5B Valuation: Nous Research, maker of the Hermes agent, is finalizing a $75M+ Robot Ventures-led round at a $1.5B valuation.

Intel Expands Ireland AI Hub: Intel will invest €5B to expand its Ireland chip plant, increasing AI processor output and foundry capacity to challenge TSMC.

Thanks for reading. See you next time!

— Matthew Berman, Nick Wentz & the Forward Future Team