In partnership with

For 25 years, the internet worked on a simple deal:

Google sends traffic.
Websites make money.
Everyone survives.

This week, that deal started collapsing in public.

At Google I/O 2026, Google rolled out the biggest Search overhaul in decades pushing Gemini directly into the center of search, shopping, browsing, YouTube, and workflow automation.

But the important part wasn’t the AI itself.

It was this:

Google is no longer trying to help users find websites.

It’s trying to stop users from needing them.

And OpenAI, Microsoft, Anthropic, and Perplexity are all moving in the exact same direction.

The Architecture Behind AI-Native Revenue Automation

Most “AI finance” tools guess. Finance can’t. This white paper explains how AI-native revenue automation combines reasoning, deterministic math, and commercial context to automate billing, cash, and close—without sacrificing accuracy. Read the architecture behind AI-native revenue automation.

Google openly described this week’s launch as a “new era for AI Search.” Gemini 3.5 Flash is now deeply integrated into AI Mode globally.

Search is no longer behaving like a search engine.

It’s becoming an answer engine.

Instead of showing ten websites, Google increasingly generates a synthesized response itself.

Users stay inside Google.
Publishers lose the click.

TechCrunch described it bluntly: “Google Search as you know it is over.”

And this isn’t theoretical anymore.

Publishers and creators are already reporting traffic declines tied to AI Overviews and AI Mode. (Moneywise)

One recent research paper found AI Overviews frequently suppress clicks while still monetizing the page with ads. (arXiv)

The internet’s old economy depended on attention flowing outward.

AI search keeps the attention trapped inside the interface.

AI Companies Don’t Want “Users.” They Want Intermediary Control.

This week made something very clear:

AI companies are racing to become the layer between humans and information.

Google wants AI-native search.
OpenAI wants AI-native browsing.
Microsoft wants autonomous browser agents.
Perplexity wants AI-native discovery.

The goal is the same:
Own the interaction layer.

Microsoft Research released Fara1.5 this week — a browser-use AI agent that reportedly outperformed OpenAI Operator and Gemini on web-navigation benchmarks. (MarkTechPost)

That matters more than people realize.

Because once AI agents browse websites for humans, websites stop designing for people.

They start designing for AI systems.

That changes everything - SEO, advertising, content writing, etc.

The web slowly stops being human-first. It becomes machine-first.

Trade from the subway, golf course, or toilet

Liquid enables users to go long or short on any market from your phone or desktop 24/7.

No matter what time of day it is or where you are, you can monitor the situation with Liquid.

Reddit Accidentally Became One of the Most Valuable Sites on the Internet

While traditional publishers are struggling, Reddit is becoming more important.

Why? Because AI systems need authentic human conversations.

Google and OpenAI have both signed licensing deals involving Reddit data. (mediaandthemachine.substack.com)

And researchers increasingly point out that Reddit-style experience-based content performs differently in AI search systems than generic publisher content. (arXiv)

For years, the internet rewarded optimized content.

Now AI systems increasingly reward discussions, opinions, and lived experiences.

The weird internet may become valuable again.

Ironically, the most human parts of the web are becoming the hardest for AI to replace.

“SEO optimized content trained the models. Human conversations may survive them.”

The Quiet Collapse of the Click Economy

Most people still think AI is competing with Google.

That’s already outdated.

AI is competing with the concept of clicking itself.

One study this month found AI-generated search interfaces are rapidly changing how users interact with information online. (arXiv)

Another paper studying Google AI Overviews found conversational AI interfaces can fundamentally change traffic distribution patterns. (arXiv)

The scary part is how fast this is happening.

This week, Google described its largest Search upgrade in 25 years. (ECIKS.org)

Not five years. Not gradually. Now.

And the incentives are obvious:

  • fewer outbound clicks

  • more time inside AI interfaces

  • more dependency on the platform

The internet used to be made of destinations.

AI turns it into one endless conversation box.

What Happens Next

The next phase of the internet probably looks like this:

  • AI agents browse sites instead of humans

  • publishers optimize for model visibility instead of clicks

  • websites become data suppliers for AI systems

  • “search traffic” declines structurally

  • conversational interfaces replace navigation

And the winners may not be the smartest AI labs.

They may simply be the companies that control where attention flows.

Google understands this.

That’s why this week mattered so much.

This wasn’t a product update.

It was the beginning of a new internet architecture.

Keep Reading