In 2020: Experts thought AGI was 50 years away. In 2026: The median estimate has plummeted to roughly 5–8 years away. The Reality Check: While an AI might be "smarter" than you at math or coding by 2027, it may still struggle to understand why you're sad or how to fix a leaky faucet for several more decades. "Smart" in a computer sense is very different from "wise" in a human sense.
2024–2025 (The Reasoning Wave): AI reached "Ph.D. level" in specific fields like biology, chemistry, and physics. Systems like OpenAI’s o1 and Claude 3.5 began using "Chain of Thought" to solve complex logic problems that previously stumped AI. 2026–2028 (The Agentic Wave): This is where we are now. AI is moving from answering questions to executing tasks. Experts (like Demis Hassabis of Google DeepMind and Sam Altman of OpenAI) suggest that "Weak AGI"—AI that can do most office jobs at a computer—is likely to emerge by 2028. 2030–2040 (The General Intelligence Wave): This is the median prediction for Full AGI (Artificial General Intelligence). This means an AI that can learn any task a human can, including physical ones, and can invent new things rather than just remixing old data.
2030–2040 (The General Intelligence Wave): This is the median prediction for Full AGI (Artificial General Intelligence). This means an AI that can learn any task a human can, including physical ones, and can invent new things rather than just remixing old data.