Peter Steinberger, a self-professed "Claudeholic," stood before a gathering in London during August 2025, participating in an event dubbed Claude Code Anonymous. Here, he and other enthusiasts convened to share their experiences with groundbreaking coding solutions like Anthropic's Claude Code. In a warm, brick-walled venue, Steinberger expressed his obsession: “I dedicate pretty much all my waking time to this, yet it doesn’t feel enough.”
Fast forward to a few months later when Anthropic unveiled an upgrade to Claude Code, named Opus 4.5. This version managed complex programming tasks more efficiently, had extensive memory capacity, could operate for numerous hours continuously, and even coordinated AI subagents. Anthropic’s notoriously challenging take-home examination for prospective engineering recruits led the company to declare that, in a direct competition, Opus 4.5 “outperformed any human candidate ever,” igniting discussions about the potential transformation of engineering jobs through AI.
During the holiday season, innumerable programmers found themselves deeply engaged in experimenting with this advanced tool, likening the experience to obtaining superpowers. As one user noted, “It feels like becoming Spider-Man.”
Despite the excitement surrounding Opus 4.5, 39-year-old Steinberger—who divided his time between London and Vienna—felt it was not enough. By November 2025, he introduced OpenClaw, a user-friendly platform allowing individuals to create personal AI agents that leverage the capabilities of Claude Code and other coding technologies. Once set up, these agents could access an individual’s data, applications, and even payment methods to perform tasks autonomously, navigating challenges with remarkable determination similar to that of the Terminator.
OpenClaw saw rapid growth, evidenced by its reception on GitHub—where the number of stars a project receives indicates its popularity. In under two weeks, the initiative garnered over 100,000 stars, and as of early May, it had reached an impressive total of 366,000.
With the arrival of these two innovations—commercially available Claude Code and the open-source OpenClaw—the anticipated era of AI agents has finally begun, albeit primarily for those with sufficient technical skills willing to tackle the often unpredictable and uncertain nature of this emerging field. Many "Claudeholics" express a conviction that they are witnessing the future unfold. “AGI is here!” declared one enthusiast, echoing William Gibson’s famous assertion that “it’s just not evenly distributed.”
The enthusiasm mirrors sentiments from the 1980s computer revolution, where technological advancements evoked both intrigue and unease from the public, while innovators thrived. “It’s hard to convey the magnitude of this change,” remarks Thomas Reardon, a former Microsoft and Meta executive who now leads a startup in a different AI sector. “It’s the most underrated, massive release I’ve witnessed in technology.”
Soon, the impact will be widespread. In a recent podcast, Marc Andreessen, a co-creator of the web browser and a noted tech optimist, stated what many in Silicon Valley believe: “It’s almost inevitable that this is the way people are going to use computers.” What remains unspoken is that adopting this direction may not be optional.
Reflecting on earlier days, in early 2024, Boris Cherny worked as a tech lead for Instagram, living remotely in a quaint house in rural Japan with his partner, where they embraced hobbies like making miso and pickles, even trading with neighbors. This serene phase came to an end when he started engaging with AI models emerging from his former home of San Francisco. Originally from Ukraine, Cherny had deep familial ties to programming, with his grandfather using punch cards to code. The allure of the AI advancements prompted him to return to the Bay Area and join Anthropic, shifting his career trajectory significantly.
