In this episode, Boris — a strategy-turned-founder and serious AI Agent practitioner, who’s also training for Hyrox and trail races — finally unpacks what it really means to use an agent like a colleague. What happened in AI coding after Claude Code launched in summer 2025? Why does he say coding is solved? And how does he use agents to own his training data, track body battery, and write his own plans? 92 minutes of substance.
1 · After Claude Code: What Actually Happened to AI Coding
Claude Code shipped in summer 2025, and even Boris didn’t expect the answer he ended up landing on: coding is solved. But the agent story has just begun — WorkBuddy, Firecrawl, human-in-loop. Each one is a new question.
2 · WorkBuddy: A Chinese-Speaking Claude Code
The Chinese-localized Claude Code already running inside Chinese workflows — the product and the ecosystem that grew up around it.
3 · Hyrox Prep: Trails Meet the Gym Floor
How Boris moved from climbing to trail running to Hyrox. Patrick’s (Kipchoge’s coach) view on training load, how he chooses between trails and road during a Hyrox block, and how he has agents write his training plans.
4 · Agent as a Colleague
For developers, for trail runners, for everyone wiring agents into their daily work — the next real question isn’t whether to use them, but where, how deep, and how often.
About the guest
- Boris — strategy-turned-founder, AI Agent power user
- Also training for Hyrox / trail races — uses agents for his own training data
- This episode: the most thorough take yet on what it means to “agent-as-colleague”
Why this episode
After Claude Code launched in summer 2025, an answer took shape in AI coding that even Boris didn’t expect: coding is solved. The agent story is just starting — WorkBuddy, Firecrawl, human-in-loop — and meanwhile he’s wired agents into his own training data and body-battery tracking.
Key topics
- coding is solved — AI can already ship production-grade code
- WorkBuddy — Chinese-localized Claude Code, already in use
- Hyrox prep — trail running meets the gym
- agent memory / context / human-in-loop — the three-piece stack
Closing
“AI Agents are not the future — they are now.” But how to use them, where to use them, how deep to go — that is the next question, shared by developers, creators, and athletes alike.