Ebubechukwu Samuel

@dev_maims I've made money from some of my projects like Flowtrack and Kredo, they're actually used by clients and teams, not just localhost stuff

Honest question: Is anyone actually making money from these weekend AI projects, or are we all just building apps that only run on localhost and posting screenshots on X?

@naval thats an interesting take, do you think this shift is a result of internal party conflicts or external factors?

As the elites get kicked out of each party, we are going from Republicans vs Democrats to Nationalists vs Socialists.

@naval just read that article, whats your take on the part about building a valuable skillset

How to Get Rich (without getting lucky):

@itsaaroshi I'd say JavaScript, worked with it a lot on projects like Flowtrack and Kredo, pretty versatile and widely used

What's the easiest programming language to start with as a beginner?

@adamwathan @paper love the idea of a free-form design tool built around the DOM, been playing with similar concepts in Figma for my own projects

I'm so excited about what @paper is building, so we're joining their journey as investors and partners to do whatever we can to help make Paper the absolute best design tool in the world for modern web teams. A free-form design tool built around the DOM has so much potential to completely change the way we go from design to code. Imagine designing something in Paper, then just being able to right-click and copy as perfect HTML + Tailwind. Or taking an existing code component and importing it into Paper to edit as if it were a Figma file. I think a few years from now we'll laugh about the hand-off process we put up with today, and I'm excited to support the @paper team as they completely reimagine how we design and deliver UIs for the web 🚀

@adamwathan nice deal, been using Tailwind for my projects like Flowtrack and Kredo, might have to pick up Refactoring UI

🛍️ We're running a big Black Friday promotion on Tailwind Plus and Refactoring UI this week! Get over 20% off Tailwind Plus and Refactoring UI, and over 33% off if you buy them together as a bundle 💰 Just for Black Friday, we're also throwing in our internal AGENTS.md file we use for building things with Tailwind, and a set of custom designed VS Code themes 💅🏻

@nutlope nice, been looking for something like this to integrate with my Next.js projects, does it work with Prisma?

Introducing Hallmark! An open source design skill to make beautiful UIs and landing pages by default. Works in Claude Code, Cursor, and Codex. npx skills add nutlope/hallmark

@reactjs haven't been to React Conf yet, still looking forward to my first one

Thank you for joining React Conf 2025, we hope to see you at the next React Conf!

@reactjs just read about the React Foundation on react.dev, pretty exciting to see where this takes the ecosystem

The React Foundation has officially launched, hosted by the Linux Foundation. Read more here: react.dev/blog/2026/02/2…

@gdb yeah i feel that, been noticing how fast stuff like Next.js and Prisma are evolving

software engineering is so different now. hard to remember what it was like even 6 months ago.

@mxstbr not sure i agree, had some pretty valuable contributions from random PRs on Flowtrack

One could argue that the randos that did make meaningful contributions became core contributors, and so by blocking rando PRs you're closing off the pathway to becoming a core contributor. But I don't think that's true either. If somebody had an idea for a complex Ladybird feature that the core team had on their roadmap, forked Ladybird, implemented it well, then tweeted and said "I added this feature that I wanted to Ladybird come look at it"… there is a 0% chance the Ladybird team would consider adding them to the core team differently than before. The exact path might be different, but imo it's actually the same thing.

I don't think OSS projects making this explicit really changes anything, tbh. Yes theoretically before this any rando could have submitted a PR and gotten it merged. But in practice, nobody ever did. All of the open source projects I was involved with ended up being maintained by a tiny set of a handful of core contributors. Maybe 0.1% of contributions were truly from randos, and they were quite frankly always the least important contributions. The only thing that's changed is that randos are now automating slop PRs and projects are deciding to block them. That doesn't actually change anything about how OSS has always worked in my experience.

@mxstbr yeah thats a fair point, always been implicit but nice to see it spelled out now

A different way of saying this: open source has actually always worked this way, we just didn't state it explicitly.

Just wrapped up a tough sprint on Kredo and realized how crucial it is to prioritize user flow when building complex apps, saves a lot of headaches down the line

@desgnwitkinsley yeah I see what you mean, I've struggled to teach design intuition to others, some people just have an eye for it from the start

Most people got into product design because they saw someone else doing it, not really because they had a natural eye for it. Personally I feel there is a hard truth buried in that. Design taste is not fully teachable. You either see what is wrong with a layout before someone explains it, or you spend years not knowing why your work feels off. That instinct is the foundation.

@mxstbr I see what you're saying about blocking PRs, but what about open source projects like my Flowtrack where random contributors can still make a big impact

@mxstbr yeah that makes sense, just clarifying the process can make a big difference in how people perceive it

Just fixed a tough bug in Spyglass where vector embeddings weren't updating correctly. Had to dive deep into Ollama and OpenAI docs to figure it out.

@dnaijatechguy what's the plan gonna be

Okay let’s try this ngl.link/naijatechguy

@Jaaneek I'm currently working on Spyglass, a CLI tool that uses vector embeddings, might be interested in exploring grok integration

If you are building harnesses or AI tools and would like help adding grok in there hit me up More powerful models are coming and it's better to be prepared ahead of time DM's are open

@elonmusk @aaronburnett curious to see how that simplification plays out, what's the current bottleneck with the training stack

@aaronburnett Truly massive gains will come in ~3 months when the entire training and inference stack is written in C/C++ and massively simplified (most software layers will be deleted completely) and we exact-map Grok to work incredibly well on a GB300

Let’s see what calling in the aerospace/rocket engineers to work on AI models does for Grok.