Ed Zitron has been calling the AI bubble since March 2024. Newsletter with 80,000 subscribers. Bylines in The Atlantic and The Guardian. But Kelsey Piper at The Argument argues his case keeps getting weaker while his accusations keep getting wilder.
Zitron started with a reasonable position in 2024. Companies weren't really using AI. It wasn't adding value. Investors were betting on future potential that might never materialize. That argument made sense then.
Twenty-five months later, things look different. GPT-4-level intelligence costs roughly 1/1000th of what it did at release, driving a massive explosion in AI code generation. About 30% of Fortune 500 companies have enterprise deals with AI startups. More than half of Americans use AI chatbots weekly. OpenAI reports $2 billion in monthly revenue. By most metrics, progress from 2024 to 2026 outpaced the previous two-year period, as measured by agentic benchmarks.
So Zitron escalated. His recent work leans on allegations that OpenAI and Anthropic are running fraud operations like Enron or FTX. He's suggested OpenAI is lying about revenue, that AI companies are systematically misrepresenting their capabilities to keep investor money flowing. Piper calls this what it is: an admission that the economic case collapsed. You can't argue companies aren't using AI when 30% of the Fortune 500 has enterprise deals. You can't argue costs aren't falling when they've dropped three orders of magnitude. When the facts don't support your thesis, you can change your thesis or accuse everyone of fraud. Zitron chose the latter.