Detailed Analysis
A Reddit post titled "Claude killed my other AI trainer" offers a firsthand account from a physical therapist and former personal trainer describing how Anthropic's Claude, paired with its "Cowork" functionality, has replaced ChatGPT in their daily workflow. The user describes a workflow in which Claude ingests personal health metrics and then autonomously writes and modifies daily workout routines and meal plans directly into an Excel worksheet. The standout feature, according to the poster, is not just the quality of Claude's recommendations but its ability to directly manipulate a working document—eliminating the manual copy-paste step that has long been a friction point in AI-assisted productivity tools. This direct file-editing capability reportedly made GPT-based alternatives "defunct" for this particular use case.
The significance of this anecdote lies less in the novelty of AI-generated fitness plans—numerous apps have offered algorithmic workout personalization for years—and more in the emphasis on agentic, file-native integration. "Cowork" appears to reference Anthropic's push toward Claude functioning as an embedded collaborator within existing productivity software rather than a standalone chatbot interface. This reflects a broader industry shift from conversational AI (where users ask questions and receive text responses) toward agentic AI (where the model takes action within a user's actual tools and documents). For professionals like personal trainers or physical therapists, this shift is meaningful because it closes the gap between advice generation and advice implementation, removing a key reason human intermediaries were historically needed.
The poster's closing remark—"the writing is on the wall, it's coming for my job"—captures a growing undercurrent of anxiety among professionals in coaching, health, and wellness fields who provide personalized, data-driven guidance as their core value proposition. Personal training and physical therapy have traditionally been considered relatively insulated from AI automation due to their hands-on, relational nature. However, the administrative and analytical components of these jobs—reading biometric data, designing progressive workout programs, adjusting meal plans based on goals—are precisely the kind of structured, data-in/data-out tasks that large language models with agentic capabilities are increasingly able to perform without human oversight.
This anecdote fits into a broader pattern of Anthropic positioning Claude as a productivity-first AI assistant, competing directly with OpenAI's ChatGPT not just on reasoning quality but on real-world task execution within existing software ecosystems like spreadsheets and documents. As AI models gain the ability to directly read, write, and modify files across an increasing range of applications, the practical barrier between "AI as advisor" and "AI as replacement" narrows further. While a single Reddit testimonial is anecdotal and unverified, it is illustrative of a sentiment spreading across knowledge-work and service professions: the value of human expertise is increasingly being challenged not by AI's ability to converse, but by its expanding ability to act.
Read original article →