Friday, June 26
AI Power, Price, and Politics
The Trump administration has asked OpenAI to delay the release of GPT-5.6 (opens in a new tab), marking the first time a sitting U.S. government has directly intervened to slow a commercial AI model launch on safety grounds. Reports suggest the White House wants individual approval authority over who can access the model — an extraordinary level of oversight that would reshape how frontier AI reaches the market. Meanwhile, Anthropic is making its own case for AI governance (opens in a new tab) from the inside, arguing publicly that its own commercial success is the necessary precondition for keeping AI safe — a position that conveniently aligns safety rhetoric with market ambition. Regulators in Europe are moving on a different front entirely, hitting Big Tech with $3.5 billion in fines (opens in a new tab) for using personal data to train AI models, with enforcement officials signaling that figure is a floor, not a ceiling.
A hardware cost crisis is quietly making every new computer purchase more expensive and more urgent. Micron has locked in historically high memory prices (opens in a new tab) under multi-year contracts, and the downstream effects are already visible: Apple has raised MacBook prices, Xbox has followed suit, and RAM costs across the board are spiking in what analysts are calling RAMageddon (opens in a new tab). Notion Mail is shutting down, citing an AI agent-driven shift in how people handle communication — a sign that rising infrastructure costs are accelerating product rationalization across the industry. For consumers, Prime Day is offering a brief window of pre-hike pricing on laptops and Apple hardware, but that window is closing fast.
YouTube is aggressively closing the gap with TikTok (opens in a new tab), rolling out a Shorts update that adds double-speed playback and tighter creator tools — a direct response to the behavioral patterns TikTok has trained into a generation of viewers. At the same time, Instagram is pushing hard into long-form video and TV-format content, positioning itself as a full-stack attention platform rather than a photo app. These moves reflect a convergence war where every major platform is cannibalizing the others' formats, and the casualty is the distinctiveness that once made each one worth using. General Intuition's $2.3 billion bet that video game environments can train real-world AI agents suggests the attention economy's synthetic media pipeline is about to get significantly more sophisticated.