Is AI Trending Up or Down in 2026? | AI Reality Check

In this episode of Deep Questions with Cal Newport, the host and AI commentator Ed Zitron conduct a critical review of three major AI developments in 2026, aiming to separate hype from reality. They examine the OpenClaw phenomenon, which sparked widespread media frenzy over AI agents forming a social network and managing personal tasks, only to reveal it was a simple Python library with no new AI technology—just LLMs doing what they’ve always done. The discussion highlights how credulous media coverage and investor hype inflated the significance of OpenClaw, leading to a brief but intense frenzy before fading into obscurity. Next, they dissect the Anthropic-Department of War controversy, where Anthropic claimed ethical boundaries around military use, yet was actively embedded in the war in Iran, revealing a contradiction between public image and reality. This was compounded by a shocking revenue affidavit showing only $5 billion in lifetime revenue despite massive funding, raising serious questions about financial transparency. Finally, they tackle the illusory data center boom, revealing that over 50% of announced data centers in the U.S. are not under construction, and that NVIDIA may have pre-sold far more GPUs than can be installed, creating a fragile system built on speculation and debt. The episode concludes with a stark warning: the AI boom is not grounded in real infrastructure or sustainable economics, and a reckoning—potentially an AI winter—is likely looming. The episode delivers a powerful reality check, arguing that 2026 has been a bad year for AI in terms of substance, despite the relentless hype. The core takeaway is that the AI narrative is being driven by media sensationalism, corporate PR, and financial speculation rather than technological breakthroughs. The real story is one of overvaluation, under-delivery, and systemic fragility. The hosts urge listeners to resist the 'dread laundering' that conflates minor issues like AI-generated content with existential threats, and to demand accountability from both media and tech companies. Ultimately, the episode calls for a more grounded, skeptical approach to AI—one that values verifiable progress over breathless headlines.