The Internet Was Built to Train AI: Why You Deserve Credit

Source: Blockworks Original Title: Friday charts: The many, many architects of AI Original Link: https://blockworks.co/news/architects-of-ai

The Original Vision

“By 2026, Google’s main product will not be search but AI.” — Kevin Kelly, 2016

When Kevin Kelly met Larry Page at a party in 2002, he asked why anyone would invest in a company offering its product for free. Page’s response was prophetic: “Oh, we’re really making an AI.”

That was 20 years before ChatGPT.

For Kelly, it was an “aha” moment. Rather than using AI to make search better, Google was using search to make its AI better. Every query, every click, every link created was training the system.

But from today’s perspective, we can make an even bigger claim: The true purpose of the entire internet was to train AI.

As economist Alex Tabarrok recently told Tyler Cowen, history will remember the internet primarily as the “agar culture for the growth of AI”—the nutrient-rich petri dish where it all began.

“When we look back,” Tabarrok said, “we’ll think: ‘What was the internet?’ Putting everything online was for the AI. It wasn’t for us.”

Except Larry Page knew. And so did his co-founder Sergey Brin, the machine-learning expert who contributed code to Google’s Gemini.

The Real Architects

Google’s founders explained their vision clearly back in 2000: “The ultimate search engine would understand everything on the web. It would understand exactly what you wanted and give you the right thing—that’s obviously artificial intelligence.”

But who put all that information on the web? You did.

“When you type ‘Easter Bunny’ into the image search bar and click on the most Easter Bunny-looking image, you are teaching the AI what an Easter Bunny looks like,” Kelly wrote. “Each of the three billion queries Google conducts daily tutors the deep learning AI.”

We’re now up to 16 billion Google queries per day.

So you should be on the cover of Time Magazine this week—for the many years of important searching you’ve done. Thank you for your service.

The Infrastructure Reality Check

Data Center Power Crisis: Forecasts for data center power consumption have risen 36% since April alone. AI-related stocks were down partly because companies can’t get enough power to run all the data centers they need.

The China Advantage: China’s big advantage in the AGI race is clear: soon it will be generating 3x as much power as the US.

The Capex Reversal: As a percentage of sales, Microsoft (traditionally asset-light) is now spending more than double on capex compared to what Exxon (traditionally asset-heavy) does.

Market Signals

Retail Investing Trends: 61% of investors under 35 use YouTube for investment advice—a significant shift in how market information flows.

Broad Market Strength: The equal-weight S&P 500 hit an all-time high, suggesting gains aren’t concentrated solely in mega-cap AI stocks.

Food Price Reality: Contrary to social media narratives, food at home has gotten cheaper as a percentage of disposable income. The total food budget is unchanged only because we’re increasingly relying on convenience foods.

The Demographic Wild Card

Here’s a sobering stat: Every 100 South Koreans today will have only 6 great-grandchildren in total—not per person. Thankfully, AI should be handling most jobs by then.

Full Circle

Time Magazine named “The Computer” person of the year in 1982. In 2006, it was you—recognizing individual content creators populating the web with videos, blogs, photos, and comments.

This year, you should rightfully have joined nine US presidents as a repeat winner. Because without the content we all created, the “architects of AI” would have had nothing to work with.

Your training session is complete. Please proceed outdoors.