NEWS

What happened today, without the feed bloat.

News is the live signal layer of SUPERCRZY. It separates the major shift from the fast read, then gives CRAZE the context needed to help you understand the story and decide where to go next.

Updated June 26, 2026 Updated date (auto) Editor filtered

Signal Flow

Scan the day in a cleaner single-line news stream.

NVIDIA, OpenAI, and Qualcomm Just Drew Three Very Different Battle Lines in AI Inference

Business

NVIDIA, OpenAI, and Qualcomm Just Drew Three Very Different Battle Lines in AI Inference

NVIDIA reframes GPUs as token factories, while OpenAI and Broadcom unveil Jalapeño, a custom ASIC slashing inference costs by 50%. Meanwhile, Qualcomm stakes its own clai…

4 min read June 26, 2026 Jeff Lu
The White House Just Showed Its Hand on GPT-5.6. And It Changed the Rules.

Business

The White House Just Showed Its Hand on GPT-5.6. And It Changed the Rules.

Two weeks after it forced Anthropic to pull its flagship models offline, the Trump administration is now asking OpenAI to stagger GPT-5.6’s release — approving access cus…

3 min read June 26, 2026 Jeff Lu
Anthropic Says Alibaba Ran a Massive Campaign Against Claude. The Timing Raises Questions.

Business

Anthropic Says Alibaba Ran a Massive Campaign Against Claude. The Timing Raises Questions.

Anthropic accused Alibaba of using 25,000 fraudulent accounts to generate 28.8 million exchanges with Claude. Two days after the letter, the Commerce Department restricte…

4 min read June 25, 2026 Jeff Lu
Nvidia CEO Just Said AI Is Profitable. Wall Street Needed to Hear That.

Hardware

Nvidia CEO Just Said AI Is Profitable. Wall Street Needed to Hear That.

Ditch the cost center narrative: Nvidia's Jensen Huang says AI data centers are 'token factories' generating profit, delivering the message Wall Street needed.

3 min read June 25, 2026 Jeff Lu
OpenAI Just Built Its Own AI Chip. And AI Helped Design It.

Hardware

OpenAI Just Built Its Own AI Chip. And AI Helped Design It.

OpenAI and Broadcom unveiled Jalapeño on June 24 — a custom inference processor built for LLM workloads. Engineering samples are already running production code. The deve…

2 min read June 25, 2026 Jeff Lu
OpenThoughts-Agent Just Published a Cookbook for Malicious AI. That’s Not Even Its Biggest Problem.

Agents

OpenThoughts-Agent Just Published a Cookbook for Malicious AI. That’s Not Even Its Biggest Problem.

OpenThoughts-Agent’s new data curation framework makes training capable agents a paint-by-numbers exercise—for anyone. Systematizing the pipeline hardens subjective human…

5 min read June 24, 2026 Jeff Lu
Mythos Found U.S. Government Vulnerabilities in Hours. That‘s Not the Real Story.

Autonomy

Mythos Found U.S. Government Vulnerabilities in Hours. That‘s Not the Real Story.

Anthropic’s Mythos found vulnerabilities in classified U.S. systems in hours. The same government banned the model five days earlier — and hasn‘t explained why.

3 min read June 24, 2026 Jeff Lu
Meta Ditches Ray-Ban for Its Own $299 Smart Glasses. The Logo Is Gone. The Strategy Is Clear.

Hardware

Meta Ditches Ray-Ban for Its Own $299 Smart Glasses. The Logo Is Gone. The Strategy Is Clear.

Meta drops Ray-Ban branding for $299 smart glasses, launches Muse Spark AI with live translation, and bets on tech recognition over fashion logos.

4 min read June 24, 2026 Jeff Lu
Three AI Models Are Dropping Next Week. Fable 5‘s Shutdown Is Why.

Models

Three AI Models Are Dropping Next Week. Fable 5‘s Shutdown Is Why.

Google’s Gemini 3.5 Pro, Anthropic‘s Claude Sonnet 5, and OpenAI’s GPT-5.6 are all expected within days of each other. Fable 5‘s free trial ended on June 22 — and three c…

4 min read June 23, 2026 Jeff Lu
SpaceX Just Signed Another AI Compute Deal. This One Is Different.

Business

SpaceX Just Signed Another AI Compute Deal. This One Is Different.

Reflection AI, an open-source lab, will pay SpaceX $150 million per month for GB300 access through 2029. It’s SpaceX‘s third compute deal — and the first with a company b…

5 min read June 23, 2026 Jeff Lu
ByteDance Just Doubled AI Video Length. That‘s Not the Real Story.

Models

ByteDance Just Doubled AI Video Length. That‘s Not the Real Story.

Seedance 2.5 can generate 30-second native video clips — double the industry standard. But the company’s real bet isn‘t on longer videos. It’s on physical AI simulation,…

3 min read June 23, 2026 Jeff Lu
Codex Is Eating Your SSD for Breakfast. OpenAI Hasn’t Fixed It.

Agents

Codex Is Eating Your SSD for Breakfast. OpenAI Hasn’t Fixed It.

OpenAI‘s Codex CLI is quietly writing up to 640 TB of logs per year to a local SQLite database. A typical 1 TB consumer SSD is rated for about 600 TB of total writes. If…

3 min read June 23, 2026 Jeff Lu
Advertisement

CRAZE

Your reading companion for this page.