After switching to software development, the sedentary lifestyle caught up. Running became necessary, not optional. When AI took off, I used ChatGPT to build a half marathon training plan. It worked—1:48:17, beating the sub-2 hour goal. Then came the full marathon: 3:43:11, sixteen minutes under the sub-4 target.

The plans worked, but the process was broken. The conversation context grew too long, responses lagged, and maintaining the thread became a chore. So I built StrideRun to solve those problems. For the nerds: stateless AI architecture avoids context bloat entirely. Each plan generation is independent, fast, and reliable. This is the tool I wish existed when I started.