Tech billionaires are building underground shelters and buying remote properties. Mark Zuckerberg has reportedly created a 5,000-square-foot underground space at his Hawaii compound. He calls it a basement. Neighbors and reports suggest it looks more like a bunker. Reid Hoffman, LinkedIn co-founder, says about half of the super-wealthy have apocalypse insurance. New Zealand is a popular destination for these backup homes.
According to BBC News, the rise of artificial intelligence has added to these fears. Some computer scientists worry that machines could soon match human intelligence. This moment is called artificial general intelligence, or AGI. OpenAI co-founder Ilya Sutskever reportedly suggested building a bunker for top scientists before releasing AGI. OpenAI boss Sam Altman said in December 2024 that AGI will come sooner than most people think.
Predictions and Promises
Tech leaders predict AGI could arrive within five to ten years. Sir Demis Hassabis of DeepMind gave this timeline. Anthropic founder Dario Amodei wrote that powerful AI could emerge as early as 2026. Supporters say AGI will cure diseases and solve climate change. Elon Musk claims it could bring universal high income. Everyone would have their own personal robot assistant, he said.
But many scientists are skeptical. Dame Wendy Hall, a computer science professor at Southampton University, says the scientific community sees AI as amazing but nowhere near human intelligence. Fundamental breakthroughs are needed first, says Babak Hodjat of Cognizant. Current AI tools spot patterns in data but do not feel or think. They lack consciousness and cannot learn the way humans do.
Marketing or Real Threat
The Debate Over AGI
Neil Lawrence, a professor at Cambridge University, calls the AGI debate nonsense. He compares it to an artificial general vehicle. No single vehicle can fly, drive, and walk. The same applies to intelligence, he argues. Vince Lynch, CEO of IV.AI, says AGI is great marketing. If you build the smartest thing ever, people will give you money. He does not think it will arrive in two years.
The human brain has 86 billion neurons and 600 trillion synapses. Machines cannot match this biological complexity yet. Humans adapt instantly to new information. AI systems must be retrained. President Biden passed an executive order in 2023 requiring AI firms to share safety tests. President Trump later revoked parts of it. The UK set up the AI Safety Institute two years ago to study risks. But some experts say the focus on AGI distracts from real problems AI already creates today.
