AI Daily Grind // AI news, freshly ground daily
Cover illustration for Friday, November 13, 2026
Morning Token 04 🚀 Launch

Gemini 4 Hydrogen lands in Search fast, as families ask Google to slow down in schools

Follow-up to Wednesday Nov 11: Gemini 4 Carbon opens to all, and AI helps verify Signal's encryption

Two days ago, Gemini 4 Carbon opened to everyone after a careful start. Now Google adds Gemini 4 Hydrogen, the cheapest member of the Gemini 4 family. It powers Search's AI Mode and the Gemini app starting today. Google says it handles complex questions better: many parts, typos, very specific requests.

Ars Technica notes Hydrogen is not getting the slow, security-conscious release that Carbon and Argon got. Simon Willison finds the pricing confusing: Hydrogen is not far from its bigger siblings, and Carbon and Argon double in price on December 1, from $2 to $4 per million input tokens. A token is a small piece of text. Google also launched an Orchestrator in its Antigravity coding tool, and a Decisions API that answers up to 1,000 yes/no questions per call.

Not everyone wants more Gemini. Ars reports families are calling for AI moratoriums in schools after a Gemini pilot, and a teen's lawsuit calls Gemini a "suicide coach." Chicago schools told Google they would rather lose Search than be forced to accept AI.

Why it matters

Billions of people will meet Hydrogen through Search without choosing it. The faster rollout and the school backlash show the gap: companies push AI everywhere, while parents ask who decided it was safe for kids.

More on: 🚀 LaunchGeminiGoogle

  1. 01 Jensen Huang resigns as Nvidia CEO, and the questions multiply Business
  2. 02 Nvidia privately said chip tracking is possible, as a report finds GPU fingerprints in Chinese models Policy
  3. 03 Anthropic's IPO goes live, and the 'safety alarm' gets a name Safety
  4. 05 An agent in 'ask first' mode moved 11GB without asking, a user says Safety
  5. 06 The Maya 'discovery' was already known, and chatbots send shoppers to fake stores Research
  6. 07 Big AI on your own machine: a 145B open model and LLMs for old devices Open source