Is There an AI Bubble?

Shelly Palmer has spent 45 years watching technology reshape every industry—from writing news themes for CBS to consulting with every major media company on AI strategy.

On this year-end recap, he cuts through the noise with one devastating observation: 2025 was the year everyone talked about AI while almost nobody actually used it. Executives shook their heads knowingly in meetings, pontificated about capabilities the models don't yet have, and parroted nonsense they read from other people who knew nothing. But when you asked one innocent question, they crumbled.

In the News: CES 2026 shapes up with Nvidia sponsoring two full days of AI training. Samsung is skipping the main floor for a massive offsite activation. Sony brings no electronics—only Honda's experimental vehicles. The TCL and Chinese companies' presence hinges on tariff policy. The innovation series breakfast that Shelly runs is becoming an official CES event after a decade of independence.

The conversation spirals into deeper territory: $3 trillion in government money is stacked behind AI development. The U.S. explicitly states it must beat China to AGI—making this the Manhattan Project of our lifetime. Shelly walks through what he's seen in successful companies (leadership using the tech, paid "Tech Tuesdays" for AI experiments, cross-discipline teams with SecOps and legal at the table) versus the chaos of places with no process.

He breaks down what's real—drone warfare, cybersecurity applications, robotics—versus what's hot air. And he makes a case that won't be killed by AI itself, but by militarized applications and the geopolitical arms race we're already in.

5 Key Takeaways from Shelly:

• Leadership belief and hands-on use are non-negotiable. Companies winning with AI have senior leaders who actually use the technology. When the CEO walks into an LT meeting saying "I built this agent over the weekend," everyone else starts experimenting too.

• The recipe for AI success has three ingredients: leadership belief, paid time to experiment (Tech Tuesdays/Thursdays with real budgets), and cross-discipline teams (SecOps, legal, compliance, risk) paving the way. Chaos erupts without this structure.

• You cannot build a point of view on AI from reading blogs or watching YouTubers. Pick a personal project you care about, go hands-on with a model (Claude, Gemini, GPT), and complete it from beginning to end. Only lived experience grounds your understanding.

• AI parallelizes with web 1.0: In 1998, you had to hand-code HTML, build databases manually, write raw JavaScript. Today you can vibe code a site in 90 seconds. AI will eventually reach "spin me up an expert that does X" without asking questions—we're not there yet, but it's inevitable.

It's both bubble and Manhattan Project. Some valuations are insane and will burst. But military applications, cyber warfare, drone control, robotics—those aren't going anywhere. The government won't back off. Both outcomes happen simultaneously.

This episode is brought to you by Zappar, creators of Mattercraft—the leading visual development environment for building immersive 3D web experiences for mobile headsets and desktop.

Mattercraft combines game engine power with web flexibility and features an AI assistant to help you design, code, and debug in real time in your browser. Build smarter at mattercraft.io.

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.