alright, i’ll admit — this week was a bit of a cop-out.
ultimately, i want this blog to be the most accurate/honest account of my process in building these various applications. thus, i can say that this past weeks project was (by design) intended to be a rather simple, and straightforward endeavor. the prior weeks project (healthviz) consumed quite a bit of my energy, so i decided that during thanksgiving week, i would build something that didn’t require quite as many lines of code, and/or hardcore thought.
although this is only the end of week 3/beginning of week 4, i feel like i’ve learned a ton. the applications that i’ve deployed thus far have been rather simple, yet the sheer process of coming up with a concept, cranking out the necessary code, finding useful APIs, debugging, deploying, getting feedback, etc. has been incredibly eye-opening for me. in many ways, i feel as though this project has fundamentally changed my life — at least in the practical ways that i approach how i manage my time, and what i think about allocating energy towards. there is always a problem for me to solve, or a new idea to test out — it is quite difficult to build an app in a week, nonetheless an app that actually works, and is useful in some way. some weeks i will build an app that is completely hashed out, and requires minimal ongoing maintenance/addition, whereas some weeks i will merely lay the foundation of a project that i will continually build on top of.
a reiteration of my thesis behind the 52 apps, 52 weeks project:
we (humans) spend way too much thinking about problem-solution, when in reality we should be putting our heads down, and moving the world forward through tangible, iterative progress.
a reiteration of my thesis behind the 52 apps, 52 weeks project:
the process is just as important, if not more important than the ideas/outcomes. find a problem, and start working towards solving that problem, even if the end is unclear. throughout the problem-solving process we can iteratively reshape our approach, and maybe even identify/solve new problems along the way — but none of this can be accomplished if one doesn’t initiate the journey.
a chess analogy — if you simply play good moves, and make incremental progress, tactics will begin to flow.
the project — “the portal”
with the portal, you enter in some text (a prompt) and the machine learning in the background will output a unique image based on your prompt.
from a technical standpoint — all i did here was use the OpenAI API/DALLE-2 model, connect this to a basic streamlit app, allowed the user to input text, deployed it for free using render.com, and now i essentially have my own customizable interface to DALLE, that other people can use as well.
‘the portal’ as it stands now, will serve as a foundation that i will continue to add onto over time. certainly, i would love to replace DALLE with Stable Diffusion, and perhaps build a more customized model. i also think building out an image->image capability would be interesting to try as well.
nonetheless, the portal is now available online — the-portal.app (if you are browsing on mobile, it will only work with the latest iOS update)
if you have any feedback or suggestions, feel free to email me: seppy_robot@proton.me
the code will soon be available in the stuff-squared repo on my github:
