Finding the right balance between my work as a founder and my personal life, including my fitness, has always been a top priority for me. I was pretty committed to staying fit, hitting the gym regularly since 2019 – but the pandemic came and I gained everything back and more.

Earlier this year I decided to get back into shape and focus on strength training. I set clear goals and consulted friends (who are fitness coaches) until I felt confident enough to work out on my own.

I used to have a personal fitness coach and subscribe to weekly meal plans, which were great (but pricey). This kept me wondering if I could achieve similar results on my own—and save some money. This led me to experiment with ChatGPT.

I soon realized there was a gap in the market for a new kind of fitness app, which inspired me to build Fit Senpai.

We pushed hard, launching the MVP after just 6 weeks of development, and I’m proud to say we’ve hit $260 in revenue. It’s not a headline-grabbing number, but it’s a milestone, and every milestone matters.

It was once just an idea and just a ChatGPT wrapper (which I know some of you hate). But hey, it’s now starting to generate revenue.

The next steps are to grow our user base and ramp up marketing efforts or even build our own LLM.

We’re on this journey, step by step, rep by rep.

  • Blackadder_@alien.topBEnglish
    1·
    1 year ago

    This is pretty cool. Can you eli5 what exactly did you do under the hood with GPT. Is there any proprietary data you have used?

    I’m mulling over a project to help first time founders and your explanations would help quite a bit how to think about my project

    • Impressive-Towel-705@alien.topOPBEnglish
      1·
      1 year ago

      So sorry for the late response. I’d say the comprehensive prompt used in the tool. I continue to improve and iterate based on user feedback. I’ve spoken to professional fitness trainers and dietitians to make sure it’s providing the best output possible.

    • Impressive-Towel-705@alien.topOPBEnglish
      1·
      1 year ago

      I hear you. I still have to update the data— adding more reviews. As you noticed the reviews were back in Aug. thanks for pointing this out!

  • IkHaalHogeCijfers@alien.topBEnglish
    1·
    1 year ago

    … or even build our own LLM.

    Don’t do this, you don’t need a LLM to generate a fitness plan. It’s like getting groceries with a monster truck. Sure it can work but its totally overkill, unsophisticated and expensive. Also, generating meal plans with with LLMs can be dangerous since its mathematically inaccurate. You don’t need AI to generate a meal plan or fitness plan unless you want them to be hyper personalized. In that case, you need recommender systems, not a LLM.

  • psilosyn@alien.topBEnglish
    1·
    1 year ago

    Anyone thinking of following this kind of thing for getting healthy, it is absolutely not the way to go.

    For those who decide to go forward with such an unsophisticated approach, enjoy breaking your body down and having nobody to help you fix it, or paying thousands later in life for someone to help fix all the damage you did following oversimplified routines.

    • pknerd@alien.topBEnglish
      1·
      1 year ago

      You still have a chance. 80% ppl can’t use gpt including your mom and Granny

  • Various_Low3880@alien.topBEnglish
    1·
    1 year ago

    small UI note, on screen sizes > 600px > 800px (like most tablets will have) your hamburger icon gets covered up by the sign up button, and at some sizes its not clickable because the sign up button covers it too much.