Last year, Dario Amodei published a widely-read essay predicting AI could automate the majority of knowledge work within years. Sam Altman made similarly sweeping statements about job displacement. Now both are softening those predictions, according to Fortune, and the timing is not subtle: both OpenAI and Anthropic are navigating IPO territory.
What's actually changing in their messaging
This isn't a full reversal. Neither Altman nor Amodei is arguing that AI won't change how work happens - they're walking back the timeline certainty and the catastrophist framing.
The shift lands somewhere around "AI changes what jobs look like" rather than "AI makes human labor structurally obsolete." That framing happens to be both more defensible as a factual claim and considerably more useful for enterprise sales, government relations, and IPO road shows. Telling institutional investors that your product will eliminate the jobs of their knowledge workers is a complicated pitch. Telling them AI makes workers 40% more productive is a cleaner one.
The credibility problem
Both executives made the original predictions in public, with apparent conviction. Walking them back invites a simple question: were they wrong then, or are they strategic now?
Probably some of both. Job displacement predictions have consistently underestimated how long it takes for technology to move through actual workplaces rather than demos and benchmarks. Restructuring jobs, retraining workers, and building organizational trust in new tools takes years - not quarters. The imminent-apocalypse framing was always more speculative than the confidence behind it suggested.
But the IPO timeline adds a plausible incentive for moderation. Both companies need enterprise customers who see AI as a productivity multiplier, not a liability. They need regulators who view them as manageable rather than existential. And they need engineering talent that isn't spooked by their own leadership's predictions about what the technology will do to the broader job market.
For the people actually using these tools daily, the macro predictions have always been somewhat beside the point. Developers, marketers, and content creators aren't evaluating whether AI will eliminate their profession by 2028 - they're deciding whether to use Claude or ChatGPT to draft a brief this afternoon. That ground-level adoption is what's driving the market, independent of how the apocalyptic rhetoric from the top gets recalibrated.