First rolled out in the U.S. in May 2024, Google’s AI Overview has now been live long enough for publishers to sound the alarm with year-over-year traffic drops.

It turns out that while it’s convenient for users, for publishers it’s an “extinction-level” event. The numbers are brutal:

A Pew study explains why: when users see an AI Overview, they’re 50% less likely to click a link and more likely to end their search session entirely. In fact, only 1% clicked on the link the AI Overview referenced. Industry watchers call this the dawn of “Google Zero”—a world where Google answers everything without sending you anywhere.

Publishers don’t really have a choice. If they block Google from using their content in AI Overviews, they’re also blocking Google from indexing their site for regular search results. In other words: say no to AI, and you vanish from the world’s biggest search engine.

Meanwhile, Google has started placing ads inside the AI Overviews, creating new revenue streams for itself, even as publishers watch theirs evaporate. The company posted strong second-quarter results with revenue jumping 14%.

Google calls the clicks that remain “higher quality” and claim AI is “boosting how many links users see,” but says it’s not their responsibility to get users to click.

1. Stop Playing the Algorithm and Start Acting Like a Platform
The Verge is doubling down on subscriptions, newsletters, podcasts, and even adding an infinite-scroll feed (mimicking social media) so readers stick around.

2. Turn Content Into Cash Through AI Partnerships
News Corp, owner of The Wall Street Journal, signed a five-year deal with OpenAI reportedly worth $250M. The New York Times struck a similar agreement with Amazon, valued at over $20M a year.

3. Compete on AI’s Terms
Cloudflare is pushing a “pay-to-crawl” model, where AI companies pay for the content they scrape. Startups like Scrunch AI help brands appear more in AI-generated answers.

In the AI era, it’s all about who sits at the top of the food chain. Everyone’s trying to survive, even Google. AI Overviews aren’t just about innovation for Google, they’re the shield against users defecting to ChatGPT.

But the ironic truth is that AI companies still need human reporting. Chatbots can remix content, but they can’t investigate or verify facts.

As one industry leader put it:

Meet Showrunner, an AI platform some are calling the “Netflix of AI.” Created by San Francisco startup Fable (backed by Amazon), Showrunner lets anyone create TV episodes in minutes, no Hollywood budget required.

Here’s how it works:

Fable’s CEO says the goal is to make shows “playable,” so fans can dive deeper into the worlds they love. Think Family Guy meets video games.

It’s free (for now) at showrunner.xyz, but a subscription model might be rolling out soon, possibly along with deals with major studios like Disney.

👉 Question is, how long before you’re binge-watching your own series?

Skej is an AI scheduling assistant that works just like a human. You can CC Skej on any email, and watch it book all your meetings. It also handles scheduling, rescheduling, and event reminders.

Imagine life with a 24/7 assistant who responds so naturally, you’ll forget it’s AI.

Whether you’re scheduling a quick team call or coordinating a sales pitch across the globe, Skej gets it done fast and effortlessly. You’ll never want to schedule a meeting yourself, ever again.

The best part? You can try Skej for free right now.

Sign up for free

German broadcaster Welt TV is rolling out KI-Welt, a weekly show “about AI and created by AI.” The program uses an avatar host and an AI-driven editorial process where research, design, and production are all handled by algorithms with human oversight. Viewers can expect deep dives into AI, robotics, and future tech, from EU regulations to China’s humanoid robots. They can also expect the AI avatar’s voice to sound every bit as robotic as a human news anchor.

McKinsey is rethinking its playbook as AI shakes up consulting. The firm has rolled out 12,000 AI agents that build decks, summarize research, and even nail the classic McKinsey tone, all jobs once handled by junior consultants. About 40% of its revenue now comes from AI-driven projects, and client work is moving from high-level strategy to hands-on execution. It’s a make or break moment as McKinsey races to guide companies through the AI era while proving it still matters in a world where consulting itself can be automated.

U.S. employers cut 62,075 jobs in July, pushing this year’s total to 806,383: the highest since 2020. AI specifically accounts for over 20,000 of those cuts, but it’s also fueling demand for new skills. Companies still plan to add 86,000 roles in areas like entertainment and insurance. For universities, AI is transforming work faster than schools can keep up, leaving recent grads entering a market where their degrees may not match what employers need. This group may feel the most pressure to adapt.

Big AI models may dominate the spotlight, but smaller models could drive the next wave of efficiency in agentic AI, according to a study by researchers at Nvidia and Georgia Tech.

Large language models (LLMs) are great for broad, flexible conversations—but most agentic systems don’t need that level of power. That’s why small language models (SLMs) may make more sense for agentic AI: they’re designed for these specialized jobs and run more efficiently.

Key Points from the Study

Why This Matters for Your Business

Big AI models grab headlines, but smaller models might be where the real ROI lies. As the agentic AI market heads toward $200B by 2034, companies that optimize for smaller, specialized models today will gain a competitive edge tomorrow.