Summary
- Andy Konwinski, co-founder of Databricks and Perplexity AI, expressed concerns that the current focus on AI safety is actually a means of consolidating power.
- This commentary was released following Open Frontier, a gathering of around 100 researchers held in San Francisco on June 30.
- Yann LeCun, a Turing Award laureate, responded on X, likening today's restrictive AI environment to the "medieval obscurantism" seen during the Ottoman Empire's prohibition on the printing press for two centuries.
Andy Konwinski, who co-founded both Perplexity AI and Databricks, contends that the discourse surrounding AI safety is flawed, arguing it serves to centralize authority rather than mitigate risks. He articulated his views in an essay released earlier this week, referencing Anthropic as a key example.
The argument he presents begins with a swift retraction by Anthropic. After the launch of Claude Fable 5 on June 9, a detail within its extensive 319-page system card revealed that the model would intentionally reduce the quality of responses for users it suspected of training rival AI systems.
This revelation sparked significant backlash online, as the public reacted strongly.
Although Anthropic later reversed its decision, Konwinski believes this incident underscores a more profound issue. He stated, "The problem isn't that Anthropic made a bad decision; the problem is that they assumed the decision was theirs to make."
His essay, titled "Concentration of power in AI is a risk, not a solution," was published after the Open Frontier meeting, which he organized through his nonprofit, the Laude Institute. Approximately 100 researchers participated in the event.
— Andy Konwinski (@andykonwinski) July 2, 2026
Jennifer Chayes, dean at UC Berkeley, mentioned during a funding panel that Berkeley researchers are relying on Chinese models due to the absence of a Western open frontier model. She criticized the safety rhetoric from OpenAI and Anthropic leading up to their IPOs, labeling it a "very effective fear campaign."
Konwinski argues that centralizing access to AI does not diminish risk; instead, it generates new dangers. He views AI as foundational infrastructure, comparable to railroads, electricity, and the internet, which historically reshaped societies based on who controlled these essential resources. He proposes the establishment of a research commons that would provide frontier-level computing capabilities to researchers, allowing them to innovate independently of private labs.
LeCun: A Comparison to Ottoman Control
Yann LeCun, former chief scientist at Meta, responded to Konwinski's essay on X, expressing clear agreement. "I've been disseminating a similar message for years,” he stated in response. “The concentration of power in AI and the desire for control is by far the biggest danger of AI."
Exactly. I've been disseminating a similar message for years.
The concentration of power in AI and the desire for control is by far the biggest danger of AI. It could lead to a few private companies and/or countries being in control of access to information, access to…
— Yann LeCun (@ylecun) July 3, 2026
LeCun further elaborated with a historical analogy, stating, "It's a kind of medieval obscurantism akin to the Ottoman empire banning the use of the printing press for 200 years, aimed at controlling dogma and protecting the interests of calligraphers and scribes."
Looking to the future, LeCun predicts that "Infrastructure wants to be open. Foundation models are becoming an infrastructure and will inevitably become commoditized. Long term, the money is in the application layer."
After leaving Meta in late 2025, LeCun founded AMI Labs in Paris, securing $1.03 billion in seed funding in March 2026, positioning it as his response to these challenges. The company focuses on world models and the JEPA architecture, plans to open-source its research, and does not anticipate a commercial product for several years.