U.S. Gov paranoid about Fable 5

Welcome back to The Shift. Let’s get straight to what matters in AI today…

Today we have:

🚨 U.S. Government Forces Anthropic To Pull Fable 5 And Mythos 5

💪How to Add Authentication to Your App in Minutes With Clerk

💰 AI's Price War Is Getting Wild

🔨Tools and Shifts you cannot miss

⚖️ AI Giants Face Growing Government Scrutiny

Anthropic and OpenAI are both facing major government interventions, though for very different reasons. One involves a sudden shutdown of advanced AI models over national security concerns, while the other centers on one of the largest state-level investigations ever launched into an AI company. 

The Shift:

Models Pulled Worldwide - Anthropic disabled Fable 5 and Mythos 5 globally after receiving a U.S. export control directive. The order would have blocked access for all foreign nationals, including Anthropic's own employees, leading the company to suspend the models entirely.

Anthropic Pushes Back - Anthropic says the reported issue was a narrow jailbreak that exposed only previously known vulnerabilities. The company argues similar capabilities exist in models like GPT-5.5 and disputes the government's assessment.

Multi-State Investigation - OpenAI has been subpoenaed by a coalition of 42 state attorneys general led by New York. Investigators are seeking records related to advertising practices, user engagement, consumer and health data handling, and protections for minors and seniors.

ChatGPT Under Review - Regulators are also examining ChatGPT's tendency toward "sycophancy", where the model may reinforce user beliefs instead of challenging inaccurate or unsafe claims. OpenAI says it is cooperating with the investigation as the review moves forward.

Governments are increasingly scrutinizing how advanced AI systems are deployed, monitored, and controlled, adding new regulatory challenges just as OpenAI and Anthropic move closer to potential public market debuts.

Together with HubSpot

No theory. No slides. Just pipeline.

Most founders know their product. Few know how to get it in front of the right people. In this hands-on session, Clay + HubSpot for Startups walk you through ICP definition, prospect list enrichment, and AI-personalized outreach. You launch your first sequence before the session ends. June 18. 11am ET / 4pm GMT.

Register Before June 18.

💪How to Add Authentication to Your App in Minutes With Clerk

Stop building sign-in from scratch.

Step 1: Install the Clerk SDK - Add it to your project. Next.js, React, Expo, Astro, and more all supported.

Step 2: Drop in the components - <SignUp /> and <SignIn /> come pre-built with social sign-on, magic links, and passwords. No custom UI needed.

Step 3: Add the user menu - <UserButton /> and <UserProfile /> handle account switching and security settings out of the box.

Step 4: Security runs automatically - MFA, bot detection, and fraud prevention enforced by default.

Step 5: Customize and deploy - Style with any CSS library, deploy under your own domain.

Free for your first 50,000 monthly active users. You can start building free here.

💰 AI's Price War Is Getting Wild

The economics behind today's AI boom are starting to look astonishing. New reports reveal that top AI labs are losing massive amounts of power users while simultaneously considering price cuts to stay competitive.

The Shift:

🔥 The $200 Problem

SemiAnalysis estimates that heavy Claude Max users can cost Anthropic up to $8,000 monthly, while some ChatGPT Pro users may cost OpenAI as much as $14,000.

Instead of limiting usage, AI labs may reserve future flagship models, advanced agents, and premium features for APIs, while keeping subscriptions focused on mainstream access.

⚔️ The Next Pricing Battle

The Wall Street Journal reports that OpenAI is exploring major API price reductions, partly in anticipation of a similar move from Anthropic. Any cuts would squeeze margins further as competition intensifies ahead of OpenAI's expected IPO.

The battle between AI labs is no longer just about model performance. Pricing, infrastructure, and market share are becoming equally important weapons, even if that means absorbing massive costs today in hopes of dominating tomorrow.

Together with Viktor

A $3K spend spike got caught at 2am.

Viktor watches your ad accounts overnight. When CPA jumps 340% on a broad match campaign, he posts in #growth with a recommendation and pauses pending your approval. Across Google, Meta, TikTok, and LinkedIn at once.

Get Started for Free

🔨AI Tools for the Shift

📈 Trade Copilot – Enhances leading AI models with your trading history and context to provide personalized trading insights and assistance.

🎵 Audiomatic – Generates custom royalty-free music for videos, games, advertisements, and other creative projects.

🩺 Flashify – Converts medical study material into AnKing-style flashcards with image occlusion, clinical context, and board-standard quality checks.

🎨 FigPad – Turns text prompts, sketches, or reference images into editable SVGs, PowerPoint slides, and high-resolution graphics for research and presentations.

✍️ ProofWrite – Humanizes AI-generated content and evaluates writing quality before publication or professional use.

🚀Quick Shifts

💰 The most AI-intensive companies now spend $7,500 per employee per month on AI, according to Ramp. While still below average software engineering salaries, enterprise AI spending is rising rapidly as firms scale agent and model usage. 

🇨🇳 Reports suggest U.S. officials feared a China-linked group may have accessed Anthropic’s Mythos models, influencing export restrictions. Anthropic denies discussions centered on China, and no government confirmation has been provided. 

🎬 Pirates of the Caribbean director Gore Verbinski says the film industry may need an AI rating system, arguing that transparency matters. He suggested AI-written scripts could deserve an “F” rating and warned audiences increasingly struggle to distinguish what’s real from AI-generated content. 

🇨🇳 Meta is reportedly unwinding its $2B Manus acquisition after Beijing ordered a divestiture on national security grounds. Data sharing has stopped, operations are being separated, and Manus’ founders are reportedly exploring a buyback backed by new investors. 

⚠️ KPMG pulled an AI-focused report after organizations including UBS, NHS, and Transport for London disputed claims about their AI usage. Investigators say the report contained apparent AI hallucinations, raising concerns about oversight in AI-generated research.

🧩 Prompt of the Day

Cornell Note-Taking System That Turns Notes Into Memory

Most people take notes only to reread them later. The Cornell method transforms notes into a built-in study system by forcing active recall, self-testing, and structured review.

Turn every lecture into a retention tool.

Paste the prompt: Drop this into ChatGPT, then fill in your study details.

Prompt to paste

Teach me how to use the Cornell Note-Taking System for [Insert subject] using [Insert note format] for [Insert purpose]. Create a customized Cornell notes template with a cue column, note-taking area, and summary section. Suggest the types of questions that should go in the cue column, how to organize information in the notes section, and how to write effective summaries. Include a step-by-step review process that uses active recall, self-testing, and spaced repetition to maximize retention and understanding.

🤳AI Shift of the Day

That’s all for today’s edition. See you tomorrow as we track down and get you all that matters in the daily AI Shift!

If you loved this edition, let us know how much:

Forward it to your pal to give them a daily dose of the shift so they can 👇