In an unusual social media post, Anthropic co-founder and CEO Dario Amodei acknowledged that the AI industry is suffering from a fundamental crisis of trust. Amodei noted that ordinary people remain deeply skeptical of tech companies and governments, suspecting they are "cooking up some new way to screw them over."

Power Concentration and Scaling Laws

Amodei addressed the debate over AI regulation, arguing that the concentration of power in the industry is not primarily driven by policy but by "scaling laws." This refers to how a model's performance improves as the resources used to build it increase. Even open-weight models, he noted, tend to shift power toward those with the most computing capacity and hardware.

To counter these risks, Amodei advocated for "rules of the road" that could institutionally constrain frontier AI companies. He expressed support for a FINRA-like entity and the Trump administration's stance on AI testing, aiming to balance safety with innovation.

A Shift in Rhetoric Ahead of IPOs

The CEO's stance on employment has notably softened. Having previously predicted that AI could eliminate 50% of white-collar jobs, Amodei now describes the technology as a "multiplier of output." This shift in messaging mirrors similar pivots by OpenAI CEO Sam Altman as both companies move toward potential IPOs.

Delivering Results Over Marketing

Amodei argued that marketing campaigns cannot fix the public's negative view of AI. "The thing that will work is actually curing cancer," he wrote, admitting that the most accurate criticism of companies like Anthropic is their failure to deliver on grand promises to benefit the world. He revealed that Anthropic is currently seeking to improve trust by focusing on breakthroughs in biology and medicine.

Vowing to avoid empty promises, Amodei concluded that honesty regarding AI risks is the only path to building long-term public credibility.