Built Different: AI won't be this cheap forever

Reid Hoffman, the co-founder of LinkedIn and one of the earliest investors in OpenAI, said something in a recent interview that every business leader needs to hear.

First, he said, "even most people who say, 'Oh yeah, I'm using AI' are not using it seriously enough." Second, he inferred the quiet part out loud... you really only have 24 months before things get REAL.

When asked how much of AI's potential we've seen so far, his answer was blunt. "Maybe 5%. Maybe 2%."

He's not alone. BCG's latest research found that only 5% of companies are "future-built" for AI. 60% are laggards. Future-built companies are seeing 2x the revenue growth and 3.6x better shareholder returns. The gap is accelerating.

But here's what almost nobody is talking about. The pricing window.

The Subsidy Is Real. And It's Ending.

OpenAI and Anthropic (the company behind Claude) are both pre-IPO. Burning billions to acquire users. That's why you can access the most powerful technology ever built for $20 a month. Google, already public through Alphabet, is locked in the same arms race. Subsidizing access to compete. When OpenAI and Anthropic go public, the ripple hits the entire market, including Alphabet's stock and strategy.

Axios reported this month that "AI may never be as cheap as it is right now." When shareholder pressure replaces venture capital patience, the $20/month tool becomes $200. Then $500.

Hoffman's 24-month window isn't about surviving AGI (artificial general intelligence, the theoretical point where AI matches human reasoning). It's about building muscle memory while the tools are cheap.

Companies Are Firing People Based on Fear. Not Evidence.

A Harvard Business Review study from January found that 39% of organizations already cut headcount in anticipation of AI. Only 2% made large cuts based on actual AI performance. That's a staggering gap between fear and reality.

Klarna cut 40% of its workforce betting on AI. Then reversed course and started rehiring because quality collapsed.

The smarter play is what the data supports. BCG and Harvard Business School found that AI-equipped workers completed 12% more tasks, 25% faster, at 40% higher quality. Junior employees saw a 43% performance boost. AI doesn't replace your team. It makes your team dramatically better.

Here's the compounding effect. Each additional AI agent you deploy doesn't just add value. It multiplies it. Agentic AI (tools that take actions on your behalf, not just answer questions) accounted for 17% of total AI business value in 2025. By 2028, that number hits 29%. Each new agent creates a marginal gain on top of the last.

But Let's Be Honest About the Risks.

AI is not magic. And blind trust in it is dangerous.

These tools hallucinate. They generate confident answers that are factually wrong. Over-reliance creates what I'd call "AI brain rot." The temptation to let AI think for you instead of with you. That leads to sloppy work and decisions nobody on your team understands.

The right approach: become an expert on the tools. Find the strategic points where AI creates real leverage. Sharpen your judgment, don't replace it. Don't let it run your business for you.

What This Means.

The window is 18 to 24 months. Not because robots are coming. Because the tools are artificially cheap. The operators who build muscle memory during the discount window will have a permanent edge.

Hoffman put it simply: "There aren't individual contributing workers anymore. We all deploy with a set of AIs." The question isn't whether you'll use AI. It's whether you'll learn it well before the price goes up 10x.

Speaking of which (shameless plug !)… next week I'm teaching Advanced Claude for Business Techniques. March 25-26. Live two-day workshop. Real workflows. Real business applications. If you're ready to move from "I've tried ChatGPT" to deploying AI in your work, register here. Comment DISCOUNT for a special 25% off discount code to the session.

Sources: BCG: The Widening AI Value Gap | HBR: Companies Are Laying Off Workers Because of AI's Potential | Reid Hoffman Interview

My friend Carl Noriega, a Built Different subscriber, recently acquired a Tier III data center this Quarter. He invited me to tour it. It was my first time inside one!

I wasn't prepared for the scale. The cooling systems. The power draw. The security. Everything engineered to keep servers running 24/7/365. Every prompt you send to ChatGPT or Claude, the compute happens inside a building like Carl's.

Here's the tension for our world. These facilities consume massive power and water. Often in markets where housing already competes for the same infrastructure. If we don't plan around data center resource demands, affordable housing loses the infrastructure war before it starts.

But the opportunity is real. Data centers are a rapidly expanding asset class. For operators who understand both the technology and the real estate fundamentals, there's a lane opening at the intersection of AI growth and real asset investment.

Congrats, Carl! Thanks for the education.

My daughter Koko and I read about Ada Lovelace in a new children’s book of hers, and I thought Ada’s story was really cool to share. Ada invented the theory of the computer algorithm in 1843, nearly 100 years before the computer’s actual invention.

Ada was the daughter of the poet Lord Byron. A Victorian-era mathematician who translated an article about Charles Babbage's Analytical Engine and added notes three times the length of the original. Note G contained the first computer algorithm in history.

She never saw a computer. But without her work, there is no AI. No proptech. She wrote the future before the machine existed. That's built different.