Descriptions:
Matthew Berman delivers an impassioned editorial framing the US government’s decision to require OpenAI to stagger the release of GPT-5.6 as a watershed negative moment for the AI industry. Rather than a general public launch, GPT-5.6 Soul preview will first go to a select group of vetted partner companies — an arrangement Berman describes as de facto AI regulation arriving through the back door.
Berman places direct blame on Anthropic and CEO Dario Amodei, arguing that Anthropic’s sustained fear-based marketing campaign — the restricted rollout of its Fable model, public accusations that Alibaba was illicitly distilling Claude’s capabilities, and repeated warnings about AI danger — effectively coached the US government into intervening in frontier model distribution. He traces a direct line from Anthropic’s messaging strategy to the regulatory framework now emerging around model releases. The video also notes that Fable itself was rolled back after just days of availability due to government pressure, a situation Berman calls entirely self-inflicted.
The video rounds up reactions from industry figures including Box CEO Aaron Levy, who stated plainly that “we now have de facto AI regulation.” Berman warns that this arrangement disproportionately harms startups and independent developers who depend on frontier model access to compete, while doing nothing to slow training runs at the major labs themselves. He closes with a call to support open-source AI projects — including from non-US labs — as the most meaningful counterweight to an increasingly centralized and gatekept AI ecosystem.
📺 Source: Matthew Berman · Published June 26, 2026
🏷️ Format: Opinion Editorial
