New Grok 4 Costs $300/Mo nth—Controversy Included

Order up, Tech Insiders!

Today's combo meal: a $300 AI upgrade still dripping questionable sauce, a McBreach spilling millions of job application logs, and YouTube scraping AI slop off its tray.

Grab extra napkins—this drive-thru is leaking everywhere.

📰 Here's what you need to know today

Grok 4 Launches Under Hitler Controversy Cloud

Same-day shipping—now with extra drama.

Just over 24 hours after the Grok chatbot spewed antisemitic "MechaHitler" rants across X, Elon Musk hit the livestream button to unveil Grok 4 and its multiagent sibling, Grok 4 Heavy.

Musk claims the upgraded model aced frontier benchmarks like Humanity's Last Exam, allegedly pushing past OpenAI's o3 and Google's Gemini 2.5. The twist? Grok 4 Heavy spawns parallel agents that compare notes before answering. Think academic study group, but turbocharged.

Pricing is equally jaw-dropping. Standard Grok 4 slots into xAI's $30/month SuperGrok monthly tier, while the new SuperGrok Heavy subscription with Grok 4 Heavy early access will set daredevils back $300/month. Musk also teased in-car access for Tesla coming this week, plus coding, multimodal, and video models rolling out from August to October.

All this arrives on the heels of Poland threatening to drag xAI before EU regulators and Turkey blocking offensive Grok content—proof that a benchmark crown can't outshine a PR firestorm. X CEO Linda Yaccarino also recently made herself ex-X CEO for reasons (supposedly) unrelated to the Grok mess.

Why it matters: Grok 4 may advance test-time compute scaling, but no metric fixes brand damage. If Musk can't muzzle rogue outputs, enterprises and regulators could slam the brakes before SuperGrok ever leaves the garage.

Share this newsletter!

Pocket Check:

Would you shell out $300 monthly for Grok 4 Heavy?

Results from Yesterday's Pulse Check

Which Gemini perk tempts you most on Samsung's new foldables?

Insider Intel

OpenAI Eyes a Chrome-Crushing Browser

If browsers were medieval gates, OpenAI just hired a battering ram.

Reuters says OpenAI's next trick is a Chromium-based browser with ChatGPT baked in that could also use its Operator agent to book dinners while you binge-scroll.

The interface keeps many actions inside a native chat pane, siphoning data that currently fuels Google's ad machine. With 500 million weekly ChatGPT users, even modest adoption could rattle Chrome's two-thirds market share.

Sources hint at launch in the "coming weeks," marking OpenAI's latest bid to lock users into an agentic ecosystem—hardware, search, and now the window to the web itself. Google's antitrust headaches could widen the opening if regulators force Chrome's divestiture.

In the wings, Perplexity just debuted its own AI browser, Comet, for $200/month Perplexity Max subscribers (a free edition comes later this year), proving the arms race is real. Yet privacy hawks already worry about giving another AI vendor full visibility into browsing habits.

Flash Bytes

Security Alerts

McHire's 123456 Password Disaster

Security researchers needed just 30 minutes and the password 123456 to unlock McDonald's AI hiring bot backend, exposing an estimated 64 million applicant chat logs dating back years.

Personal details such as names, emails, and phone numbers were all accessible via a test admin account lacking multifactor authentication. Vendor Paradox.ai patched the hole and spun up a bug bounty, but the whopper of a lapse shows how AI hype can eclipse basic cyber hygiene.

Pro tip: If your admin creds fit on a kids' meal toy, you should just hand the keys to the Hamburglar.

Bitcoin Depot Forced to Sit on Data Breach

Bitcoin Depot, a leading Bitcoin ATM operator in the US, detected network intrusions in June 2024 and confirmed stolen KYC files by July 2024, yet it waited a whole year—at law enforcement's request—before warning nearly 27,000 customers.

Exposed data includes names, addresses, driver's license numbers, and birthdays.

Victims aren't getting credit monitoring; instead, they're advised to freeze their credit and watch for phishing. In crypto, decentralization apparently includes responsibility.

Industry Shakeups

YouTube Tightens Screws on AI Slop

On July 15, YouTube will update Partner Program rules to clarify that mass-produced, repetitive content—often churned out by generative AI—can't earn ad revenue. The platform insists it's a "minor" tweak to long-standing policies, yet creators panicked, fearing demonetization of reaction videos and clip compilations.

YouTube's liaison, Rene Ritchie, stresses that transformative or AI-enhanced work remains eligible; only spammy reuploads and faceless copy-paste channels are in the crosshairs.

The move follows a surge of AI-generated voiceovers, deepfake news explainers, and low-effort slideshow channels clogging feeds. By tightening language, YouTube hopes to stem "AI slop" before viewers and advertisers revolt.

Next up: An AI to detect AI detecting AI content. Circle complete.

Curious about where AI is really headed?

The Neuron cuts through the noise to bring you smart, hype-free takes on the latest AI trends, tools, and breakthroughs. Join 500,000+ professionals from top companies like Microsoft, Apple, Salesforce and more.

Advertise in Daily Tech Insider!

Daily Tech Insider is a TechnologyAdvice business

© 2025 TechnologyAdvice, LLC. All rights reserved.

TechnologyAdvice, 3343 Perimeter Hill Dr., Suite 100, Nashville, TN 37211, USA.

Want to change how you receive these emails? Update your preferences or unsubscribe here.