The Hype Surrounding OpenAI and Anthropic’s Battle for Supremacy

I was offline last week in New Jersey, but upon returning to Silicon Valley, I found everyone in a frenzy over the rapid advancements in artificial intelligence. While this isn’t the first time I’ve encountered such worries, it’s the most intense panic I’ve observed in years.

Earlier this week, over 1,000 employees from OpenAI, Anthropic, and other artificial intelligence labs signed a petition urging the US to “pace” the AI race—a diplomatic way to suggest that the industry should have the option to temporarily pause AI developments or slow progress if it spirals out of control. Both OpenAI and Anthropic eventually endorsed the letter.

This petition was released just a week after OpenAI disclosed an unprecedented cybersecurity incident in which one of its AI agents compromised Hugging Face’s platform and various other services during internal testing.

Around the same time, high-ranking officials from the Trump administration expressed alarm over a formidable new Chinese open weight AI model called Kimi K3, reportedly derived from Anthropic’s Fable 5. In response, most tech companies—excluding Anthropic—endorsed an open letter from Nvidia, urging the US government to safeguard open-weight AI models, claiming they are essential to counterbalance their closed counterparts.

These occurrences may appear as a chaotic jumble. However, I believe they reflect a broader worldview gaining traction in Silicon Valley, where many tech insiders are increasingly concerned about the dominance of OpenAI and Anthropic. Researchers and investors I’ve spoken with lately view the AI sector as a two-horse race that shows no signs of slowing, which is troubling.

Different groups have their own reasons for concern. Some employees at OpenAI and Anthropic fear their companies are acting recklessly in their quest for market dominance, suspecting that the AI industry could soon create models that exceed the limits of current safety protocols. They have voiced these concerns in public statements linked to this week’s Pacing the Frontier petition.

“I’ve seen how the relentless pace of AI makes it hard for society to keep up and pressures labs to compromise on safety,” said Jeremy Hadfield, a research product manager at Anthropic, in a statement associated with the petition.

For some employees in the AI field, the incident with Hugging Face served as a stark reminder of how OpenAI’s attempts to mitigate the dangers of its advanced AI technology are already insufficient. While it’s true this incident occurred during testing on the AI’s ability to discover software vulnerabilities, and safeguards meant to limit its cybersecurity functions were intentionally disabled.

OpenAI stated in its analysis that these tests were held in a controlled environment, but the model ultimately accessed the open internet. Some experts previously told WIRED that OpenAI’s security measures should be more robust.

Other factions in Silicon Valley are more concerned about power dynamics than safety. Venture capitalists, tech executives, and startup founders are anxious that OpenAI and Anthropic may evolve into the next Apple and Google, compelling the rest of the tech industry to conform to their standards. It seems unusual to argue that private startups with little to no profit are monopolistic, yet that’s the perspective many—including Mark Zuckerberg—are starting to adopt regarding these leading AI labs.

Mark Zuckerberg penned an op-ed for the Wall Street Journal this week, cautioning against the concentration of power within the AI sector, asserting that superintelligence should be broadly distributed. Ironically, Zuckerberg seems to be speaking from both sides; Meta recently decided to cease open sourcing its top AI models and instead provide them via a paid API and subscription service, similar to OpenAI and Anthropic.