Web dev - How would you go at it today?

[Neiluj]

Hi all,

A short off topic question, humbly pooling community wisdom:

• Let's say you have an app idea (nothing too complicated but also not a static website), and while you can mostly read an interpret code, web dev is not your current daily occupation.
  You start with a "blank canva" (albeit the idea is pretty well defined - and you know it fits a need, albeit somewhat niche) and have no attach to one or another tech stack.

• Naturally, you want to start simple before adding more relevant features (so the potential to further develop and add on the existing is important)

• Ideally, you would want to have a first version in the next 8-10 months and while your time cannot be 100% dedicated to this, you are ready to devote a fair amount of resources to this project (whether it is time or money). You are willing to fail but do not hope or expect to. Succeeding means your app will be used, nothing more, nothing less. But it also means it would open the door to other projects.

• Ultimately, you imagine a PWA or web app with mobile apps, with auth integration (OIDC, LDAP etc..) and easy to update, maintain and deploy (did I hear Cloudron app?).
  You are not interested in being at the forefront of the tech trends (AI, AR/VR etc.. ) at this stage, yet you wouldn't close the door on possibilities.

**

Question: Which way would you go to make this happen, in today's world?

Would it be to go at it alone, learn from books, training, people etc.. and make this happen?
Would it be to go the AI way and see how far this leads, while learning?
Or find a someone (contractor dev or "co founder" or else) with whom you convince to share your idea and hope to be able to rely on a long cooperation?
Something else?

What's is your best advice/experience with this?

**

There is such a compelling community of Cloudronians that I though it would be interesting to ask.
Thanks in advance for your thoughts!

[robi]

LLMs are all the rage to get a feel for the implementation and iteration of whatever your app does. Be it from a napkin sketch or from a Canva design.

If it shows interest, get customers to pay for you making what they want, then if needed, hire a dev/team to keep going and polish it + support.

[scooke]

I have this exact train of thought several times a week! The ultimate question being: where do I start?

[Neiluj]

@robi Thanks for this.

• LLM: I really need to up my skills on this. Do you know of good resources for developing via LLMs? Any recommendations about a particularity relevant LLM to use in this context?

• Customer interest / hiring: without putting the cart before the wheels, I am also wondering how to determine a ballpark figure of the cost involved here. Say I shop around and get offers back, is there/do you know of a good way to evaluate if this is fairly market priced or simply out of proportions?

Naturally, there are far more questions around the corner - starting with choosing a relevant dev stack depending on the choosen direction.

Any sharing of related resource or experience, is very much appreciated.

[Neiluj]

@scooke indeed!
in approaching these situation, I am a fan of breaking things into "ingestible" bits / smaller actionable parts.
But it doesn't mean I have cracked this nutshell...

[jdaviescoates]

@Neiluj said in Web dev - How would you go at it today?:

starting with choosing a relevant dev stack depending on the choosen direction.

Unless there is already a budget to pay people, in my limited experience the dev stack is normally just chosen by which ever developer you manage to get on board for free in the early days.

Products/ projects often go through complete re-writes as things develop, so I'd try not to be too perfectionist about it in the early days when you just want a Minimum Viable Product to be able to test your assumptions (says me, a terrible perfectionist who doesn't listen to his own advice and overthinks everything )

[scooke]

@Neiluj said in Web dev - How would you go at it today?:

breaking things into "ingestible" bits / smaller actionable parts

You know, I tried this exactly ... with ChatGPT! I've managed to get one app idea into a visual, working, state. I gave the command, and because I said the app has to be cross-platform, ChatGPT opted to give a bunch of code in Flutter. I then decided to install all the required software (flutter, android studio, etc.) to see if the code actually worked, and it did. From that initial command, I then iterated on it and asked, "with the code you just gave me, add this functionality" a few times and wow. I'm currently stuck on getting it to provide me with code to plug into a db backend, but it's progress.