He Predicted the AI Bubble in 2023 | Doug Clinton and Gene Munster on Why We're Still in 1996

We're still in 1996, not 1999

Clinton maintains his 2023 prediction that AI will create a bubble larger than dot-com, but believes the market remains in 1995-1996 rather than 1998, suggesting several years of runway remain before any peak.

Electricity converted to intelligence

AI fundamentally transforms electricity into intelligence, creating seemingly infinite demand for power and compute infrastructure with data center energy capacity representing the primary bottleneck.

Claude Code ignited enterprise demand

Anthropic's model releases made coding accessible to non-programmers, driving the company's revenue run rate from $9 billion to $45 billion in just four months.

Corporate AI budgets exhausted in months

CTOs at Uber and ServiceNow reported burning through entire annual inference budgets within the first four months as employees discovered massive utility in AI coding tools.

Models reaching professional competency

AI capabilities progressed from high school graduate level a year ago to college graduates with two years experience now, with PhD-level competency expected by year-end.

Acute knowledge worker unemployment coming

Munster predicts the next five years will see more severe white-collar job displacement than during the mobile or internet transitions as enterprises aggressively automate.

The 80/20 rule applies to AI adoption

High-performing employees who embrace AI will become supercharged and invaluable, while skeptical or slow-adopting knowledge workers face irrelevance and job loss.

New detective roles will emerge

As AI commoditizes existing information, new jobs will involve finding novel real-world data that models cannot access and feeding it back into enterprise systems.

GPT 5.5 currently leads rankings

Testing across 700 stocks places OpenAI's GPT 5.5 at the top, followed by Anthropic's Opus 4.7, with Google's Gemini 3.1 and xAI's Grok 3 tied for third place.

Different models excel at different tasks

Codex dominates production-level coding while Claude serves better as a thought partner for ideation, with enterprises increasingly using multiple models for different workflows.

Google lags in the coding revolution

Despite integrating Gemini effectively into search, Google has been slowest among major players to embrace the agentic coding revolution, with Gemini CLI not yet competitive.