Thread

I'm currently doing something similar although it is not with ChatGPT. I'm having an AI teach me Python. I told it why I was interacting with it initially, i.e. that my goal is to learn Python. I told it some relevant information about my education background with respect to coding so it could use my previous experience whenever appropriate to map aspects of Python to other forms of coding with which I have some experience. I told it to generally adhere to WordPress's standards for brevity and succinctness. I told it to make sure to define any words that are commonly used in discussion of Python that are uncommon in use outside of Python discussions. Likewise I told it to make sure to define any words that are commonly used in the discussion of Python that are also common words outside of Python discussion and have different, specific meanings. I told it to make use of bullet points, numbering, and other information organization methods whenever appropriate. I told it to make use of visual aids whenever appropriate. I told it to give me 2 challenges after each lesson to demonstrate my understanding of the concepts presented in the lesson. So far that general set of instructions has yielded some great 101 level Python lessons. I'm picking it up. I'm pretty new at this, so please correct me if I'm wrong, but #vibecoding is where one writes prompts for AI to write some kind of code. If that's the case, then it seems to me that vibecoding comes down to two main skillsets: 1. Knowledge of coding 2. Ability to compose written language precisely and accurately. It seems to me that improving one's writing skills while studying coding would be the two most beneficial pursuits for someone interested in development of vibecoding skills. I've heard that Jordan Peterson has some kind of app that helps people learn to become better writers. I imagine that might be a helpful tool for pursuing item #2 from above.

Thank you for your input. I actually started with a very similar approach using ChatGPT for a python study plan. I think it is a great learning tool and helped me a lot, especially because of exercises and the 1:1 teaching where you can deep dive what ever interests you the most right now. I am sure I would have learned much more with the current version of ChatGPT (not even special learning AI) and free time instead of school and even university... But that's another topic. I had a project (finance tracker and planer) I started coding after some lessons. I later shifted to let ChatGPT write most of the code and make sure I understand all of it and look at the used Syntax etc. There are a lot of logical problems it can't solve in its own, especially when the project gets more complex, so you have to solve these on your own, but for the coding it worked surprisingly well, is fun and keeps me more motivated because I see more rapid results. I used python and streamlit for a web based program I can access through a VPN and also made a small flutter app to save cash transactions. I agree with your two vibecoding improvement tips and think those are best learned by doing projects and experimenting. But a third point would be tools. With the rapid evolution of AI I guess there are better AIs for vibecodeing or certain prompts and tricks that streamline the experience. I simply have no overview and am searching for a better entrance in that rabbit hole.

Thanks for the response. I will keep your experience in mind as I make my own progress in these areas. I know a few people who have been using AI e.g. nostr:nprofile1q9z8wue69uhk67trx4hk6c35wachwmnxwy6n2vmddd5hgmpnwp58wdnhdahnwmfhwam8jdntwe3kz6tcv3hxsc3hv4jxwctzdeukgtn0de5k7m30qqsfurf693da7utpkmey3ez67jq284ddjgucf4cv5sshndf9uf724wqurnqlh . I think there are some Bitcoiners out there who have pivoted to focus on AI. I feel like nostr:nprofile1qyfhwumn8ghj7ctvvahjuat50phjummwv5q32amnwvaz7tm9v3jkutnwdaehgu3wd3skueqqyzu7we2xhgry2mknq8v7227yn7jguu9xhu3g90n6rtnjj3mpyq3ackdvvhl has done a lot of content on AI that could possibly help you out with an overview but I don't know. That's just a hipshot suggestion. If I find any overviews or good resources on AI I'll try to remember to share them here for your benefit and the benefit of others. I'll ask a couple people at work if they have any thoughts on AI overview content. My guess is that we're going to have to seek out podcasts and immerse ourselves...40 hours per week 😉 #40HPW

A key question to probe in your journey is by looking to see what's left after the dust has settled. Where you look is critical. If you have to look outside your own mind, then you haven't actually built any foundation. I will give you an example: TIL Learned that folding hose is 10x easier than coiling it. 1. create a oval circle, 2. drag it from the middle until the two ends mostly meet creating a sharp V, 3. walk to the middle and repeat step 2 and 3. End of example This is a divide and conquer strategy. I looked after the dust has settled and found a cool program that is still operating years later. The foundation is rock hard. I can see that the efficiency delta is: Coiling: BigO(n) Folding: BigO(n * log(n)) I didn't have to move my eyes away from the screen to arrive to the solution. I can do it outside on a nice sunny day. Software development is extremely inhumane as a career.