Value and intention

54 hours changed how I see mentorship

Have you heard about the Startup Weekend? In short, it is an event where you only have 54 hours to validate and build an MVP for a business idea. I've participated in a couple of those in the past, but it has been 10 years since the last one. On a recent weekend I got the opportunity to not only be invested into the event again but to be a mentor for the rest of the participants!

To be honest with myself, it hit differently from what I expected... In a good sense. Somehow I didn't feel the well-known Impostor Syndrome in this case and I felt really excited to mentor the rest of the teams. At first I was kinda shy but as the event advanced and teams started looking for me and my advice I felt like I was actually doing a good for them (just like the past mentors helped me back then when I was a participant too).

Mentorship isn't just sharing knowledge. It's building connection, career growth, networking and mutual learning. The friction of explaining forces you to explain with clarity. And the person on the other side? They're paying attention in a way an audience never does.

I remember repeating to myself phrases like "When I have time, I'll mentor. I'll speak. I'll run a meetup." This Startup Weekend made it obvious: It is up to me to make that time. Building this intentionally by assisting to local meetups, one-on-one conversations and speaks. It all sums up to my goal and my career, yes, but I also genuinely enjoy doing it. I will make time to get more involved.

Claude vs ChatGPT (my quick take)

Last month I mentioned that I switched from ChatGPT to Claude. One month later I can say that I'm sticking with Claude.

After a couple of weeks, the real test came down to this: Claude consumes tokens exponentially faster than ChatGPT. The Plus plan that carried me through weeks of heavy coding on ChatGPT runs dry after 2-3 days with Claude. I'd have to pay extra. So I had to choose: spend more for a tool, or stick with something cheaper that works?

I chose to pay more because what Claude does justifies the cost. It's better for writing, planning, coding, and orchestration. The Cowork integration gives Claude access to my filesystem, and it has this Dispatch feature that lets me send commands remotely to Claude. One night I was out and a colleague asked for a code review. I sent it through Dispatch, Claude reviewed the code, noted the comments, confirmed everything with me, and sent the summary to my colleague on WhatsApp. That changed how I think about what a tool can do for my workflow.

ChatGPT excels at image generation and voice models that actually process audio. Claude Design handles design work better. The Claude app works on my Intel processor where ChatGPT doesn't. Each tool has different wins.

I consciously chose the more expensive tool because it aligns with my day-to-day activities and the capabilities and scope outweigh the token cost. That choice told me something about my priorities. I'm optimizing for what actually lets me work better. My skills combined with a good AI tool, that's what a call a winner combo!

Beyond the day-to-day, I'm also thinking about building a local LLM server, something I can control completely, keep everything in-house. This idea comes from a privacy concern. For now, Claude is my main tool. ChatGPT is backup for things I can't do with Claude, like image generation. But the local server idea is sitting in the back of my mind as something worth exploring.