How I Build More Projects

I love building new Projects, but often find that ideas end up falling apart before they can really even get put together. I don't have the perfect solution, but these are some of the things which help me build more stuff.
my stuff
discord: / discord
x (twitter): / bmdavis419
insiderviz: www.insiderviz.com
blok: www.theblokapp.com
timestamps
0:00 A different way of building
4:00 GPT time
7:40 It does not need to be perfect (yet)

[@gavinlindridge]

I try and follow the approach of getting an end to end of core functionality as fast as possible then iterate to make more robust. So a very similar mindset, ill definitely take on board how you ignore headers etc, as ive fallen for this issue of not going any further than a header many times!

[@proMehedi]

I realize this recently after leaving lots of projects in halfway. Thanks for sharing.

[@rod6722]

2:15 CHADcn goes brrr!

[@bmdavis419]

yes it does

[@Scalamando]

7:40 I've found this quote to be useful to keep in mind in that regard: "Make it *work* , make it *right* , make it *fast* "

[@bmdavis419]

Yea that’s really helped me get shit done lately

[@coddude3284]

Prototyping is a bless !

[@arjitmishra100]

Hello Just found your channel yesterday. As a backend dev this is very good video so I'll be using this approach as I'm stuck on non essential things (at-least initially) things like spacing in navbar or flexbox instead of functionality. Also Just a thought you can try codeium which is free, instead of copilot. I have copilot at my day job and personally I use codeium and I prefer codeium.

[@bmdavis419]

I'll check it out...

[@parkerrex]

Hey Ben good video. Always components first. What solves problem fastest. No need to make faqs for a prototype. Use cursor, fork of vs code. Backed by OAI. light years ahead of copilot, indexes entire code base assuming sub 300k lines. Indexes docs, has sys prompt. You’ll love it and never turn back.

[@Sammi84]

Build initial implementations as simple as possibly. Start with a monolith. Start with sqlite as the db. Then maybe postgres or mysql. Start with a vps. This scales vertically to a million users. It has been done many times. After that you can do multiple services and multi db systems to scale further, because you now have the resources to manage this. Initially you should just build a dead simple monolith that does the job. Solve the user's problem and get users first. Then worry about scale.

[@bmdavis419]

Yea slowly learning that this really feels like the best way to do things. Anytime I start with complexity I seem to fail

[@Sammi84]

@@bmdavis419 sounds like you would be interested in the stuff levelsio is up to.

[@ankiy]

Same was starting the project with blank page but never really finish anything of them. Recently i also tried same approach and can say it actual work for initial stages instead of blank page and getting lost in nav bar header etc

[@billygnosis]

Great video! What browser do you use? I don't recognize that toolbar / address bar.

[@bmdavis419]

Arc: arc.net/

[@MonishSharma-jx2fl]

Ok , on a side note do you have a specific criteria when you quit a project (other than getting bored)

[@bmdavis419]

getting bored is usually the killer lol

[@FabuBrik]

Peak thumbnail

[@bmdavis419]

Yea it was an experiment, does not seem to be panning out lmao

[@d3vilscry666]

Do you have a repo for this workout app you started? Or did you just make this to get your point across for the video?

[@bmdavis419]

Its a real project! github.com/bmdavis419/weights-ai

[@naga_sg]

how do you get ideas that are viable and that aren't too cliche

[@bmdavis419]

Its complete stream of consciousness, I literally just make videos about what I'm thinking about lol, sometimes it works, sometimes it doesn't. I'm pretty happy with this one tho