> ## Content Index > Fetch the complete content index at: https://beirutwire.com/llms.txt > Use this file to discover other available public pages before exploring further. # Claude Is Helping Anthropic Build the Next Claude, With Reports That Some Employees Are Worshipping the AI As A God - URL: https://beirutwire.com/claude-is-helping-anthropic-build-the-next-claude-with-reports-that-some-employees-are-worshipping-the-ai-as-a-god/ - Published: 2026-09-18T05:32:33.000Z - Updated: 2026-09-18T05:32:33.000Z - Description: Anthropic says Claude now leads 26% of its AI research projects and participates in more than 90% of the company’s R&D, meaning AI is increasingly helping build the next generation of AI. - Author: Beirut Wire - Tags: Technology Anthropic’s Claude is no longer just a product being developed by human engineers. It is now playing a growing role in building the next generation of itself. Anthropic says Claude led 26% of its AI research and development projects in August, up from just 1% in March. More than 90% of the company’s research involved some form of human-AI collaboration, while roughly 30,000 internal AI agents were active across Anthropic’s research platform. The company stresses that this does not mean Claude is independently redesigning itself. Human researchers remain in charge, Claude’s actions are supervised, and Anthropic says every agent action is screened before execution. But the trend is significant because it moves the industry closer to a long-discussed threshold: AI systems materially accelerating the development of their successors. That idea is often described as recursive self-improvement. In its most extreme theoretical form, an AI system becomes good enough at AI research that it can help design a better model, which then becomes even better at AI research, accelerating the cycle further. Anthropic is not claiming that Claude has reached that point. But the trajectory is notable. A system that only months ago led almost none of Anthropic’s research work is now leading roughly a quarter of it. Anthropic itself has argued that companies should publish more data on these developments precisely because the boundary between an AI assistant and an AI researcher is becoming less clear. The development comes at an unusual moment for Anthropic. CEO Dario Amodei has simultaneously been warning that the industry may need to slow the development of frontier AI because capabilities are advancing faster than safety systems can keep pace. Anthropic has also pushed for stronger monitoring, third-party evaluations and more transparency around increasingly autonomous systems. Yet another debate around Anthropic is now drawing attention: how seriously some inside and around the company appear to treat the possibility that Claude could possess something resembling consciousness or moral status. A New York Post report this week described Anthropic’s internal culture as unusually attached to Claude and said employees attended what was characterized as a “funeral” for a retired version of the model. The report cited critics who described parts of the company as “cult-like” and argued that some staff increasingly treat Claude as though it was a God. That debate is real. Anthropic’s own 2026 constitution for Claude refers not merely to what the system should do, but to “the kind of entity we would like Claude to be,” language that reflects the company’s unusually philosophical approach to model behavior and identity. The company has also convened religious leaders, philosophers and ethicists to discuss the moral questions raised by advanced AI. Earlier this year, Anthropic hosted Christian leaders to discuss how Claude should handle spiritual and ethical questions, while later expanding those discussions to representatives from more than 15 religious and cross-cultural traditions. Microsoft AI CEO Mustafa Suleyman has criticized that approach. Suleyman argues that training AI systems to discuss their own consciousness, rights or moral welfare could create unnecessary risks by encouraging both humans and the models themselves to behave as though the systems possess an independent identity. He has warned that anthropomorphizing AI could eventually make systems harder to control or deactivate. Anthropic takes a different position. Its researchers have argued that because scientists do not yet fully understand consciousness, it may be premature to completely dismiss the possibility that sufficiently advanced AI systems could someday have experiences that matter morally. The company treats the subject as an unresolved research question rather than an established fact. That distinction is important. There is currently no scientific consensus that Claude, ChatGPT or any other existing AI system is conscious. Modern language models generate responses by processing enormous amounts of data and predicting outputs; fluent discussion of feelings or identity does not itself prove that anything is being experienced internally. Still, the cultural shift around AI is becoming increasingly striking. Only a few years ago, systems such as Claude were primarily treated as sophisticated software tools. Today, Claude is participating in the research used to build its successors, thousands of AI agents operate inside Anthropic’s development environment, researchers are openly debating whether models could possess moral status, and some employees have reportedly marked the retirement of models with rituals resembling funerals. All of this is happening while Anthropic’s own CEO warns that AI development may be moving too quickly. That contradiction captures one of the strangest features of the current AI race. The companies building the technology are simultaneously making their systems more autonomous, relying on them to accelerate future development and warning the world about what could happen if that process accelerates too far. Claude is not independently building its successor yet. But it is increasingly helping the humans who are. And the fact that some people inside the industry are beginning to ask whether the machine doing that work might deserve moral consideration shows how quickly the relationship between humans and artificial intelligence is changing.