About two years ago I began teaching myself Ai. I learned all about ChatGPT and how to create and train my own Ai agents. I learned Ai imagery platforms and how to make art, water colors/Ink artworks with my private GPTs I built. And then about eight months ago I heard about Ai Music generation softwares. I was a touring sound engineer and concert video director and was blessed to work with some incredible world class musicians and was a songwriter since I was 18 years old, but never published any of my works. Well then I found Suno. I created Ai artists and started writing music and now 20 album releases later I have built an entire music label with ten artists with the help of Ai. While some are not into Ai music, because there really is a lot of slop out there, I approached it as an engineer, composer and lyricist and learned to build command structure into the lyrics that made it more like composing a track rather than prompting it. So now I create music and artwork and videos to promote the music and could not be happier with my choice. Ai has literally changed my life and my career trajectory at 63 years old. Now I am creating real life bands around this music and am having a blast doing it!

The following app on https://fluxtimer.pplx.app is life an publicly accessible It is generated with Perplexity Pro version. Took me 10 minutes for the prompt and the AI 10 minutes to generate a first local draft and another 2 minutes to publish this for the public. So all in all 20 minutes for creating an advanced Workout-Timer similar to the Seconds Pro App with the difference that you can edit and upload your workout-routines in an *.md file rather than painstakinly edit it on your iPhone. But the point here is not the feature. Its what follows afterwards. As my goal with this app is to sell customized workouts for individuals the app must be predictably stable with the option to PREDICTABLY add new features without putting everything in danger customers are already used to use. There must be no change to what has worked so far. And therefore I have to OWN the app code and all the building process the same way I had to won it before AI. So Step 1 was to have the code local on MY machine in MY VSC editor under MY control using MY tools. And it took me 2 hours to find out whether this is doable at all. Step 2: Creating a development environment with associated GitHub Repos to manage the code: another 2 hours (testing and documentation included). Step 3: Building the code as is - failed - off course. Why? because its built and tested for Linux - not Windows. However, the AI has helped a lot and what would have taken my hours or even days figuring out the machine-code level cryptic errormessages where they are coming from, AI was able to point out DIRECTLY in which configuration files I had to change what to tweek it for Windows. Lesson: Either harness your AI that it really builds for your setup (but might not find then a solution that meets your requirements) or ask AI before building your code what it means for porting the code into your context. Step 4. Once the app was compiled successfully on my machine, publishing it with Perplexity AI was a breeze: 5 minutes.