The Third AI Winter Is Coming. And You'll Pay for It...
Artificial intelligence has already experienced two AI winters—periods when grand promises collided with reality, funding dried up, and entire research areas went into hiding for years. The first time was in the 1970s, the second in the late 1980s. And both times, the problem wasn't that artificial intelligence proved impossible. The problem was that the economics stopped converging.
Today, the third round begins. But now the stakes are incomparably higher.
Modern AI requires enormous computing power, data centers, graphics cards, electricity, and water. Training advanced models already costs hundreds of millions of dollars, and major companies are investing billions and trillions in infrastructure. Artificial intelligence is gradually transforming from a computer program into a massive physical industry.
But investors aren't the only ones paying for this race.
In this video, we'll explore how two pages by John McCarthy in 1956 launched the history of artificial intelligence, why neural networks emerged, how the first and second AI winters occurred, why perceptrons proved a dead end, how video cards from the gaming industry suddenly revolutionized machine learning, and how ChatGPT turned AI into a global economic race.
And then, the main thing.
What happens if the economics of artificial intelligence don't converge again? Who will bear the costs of data centers and electricity? Why could AI hit the labor market even before true artificial intelligence emerges? And why could the disappearance of junior specialists be far more serious than another stock market crash?
This isn't a story about the rise of the machines.
This is a story about money, electricity, computing, the labor market, and the price of a seventy-year-old promise to create artificial intelligence.
Investors and companies survived the first two AI winters.
This time, it might not be them who bear the bill.