I remembered getting fascinated seeing the git commits on other developer profiles. But 2020 surprised me with mine. Didn’t even realised when i reached this number.
Check out my profile here-> https://t.co/HpBmDshpHx
From a CTO at a startup:
"We interview devs by giving them a task to build an app on the spot, from scratch (2x BE endpoints, some frontend.) They can use AI, ofc - and we dig into why they did this or that.
What is surprising: 14/15 devs from Meta failed this screening."
Next.js right now feels like a weird paradox. You can’t fetch data in Client Components, but you also can’t mutate Server Components once they’ve mounted. So we’re left in this awkward middle zone where you create these tiny Server Components that just fetch data, and then pass it down to Client Components that basically act as static UIs pretending to be dynamic. It’s like writing a “fake SPA” inside an SSR framework it works, but it feels like coding gymnastics for something that should be simple.
And yeah, technically, you can get around it by making Client Components call Server Actions, even though Server Actions were never designed for general data fetching. Fvck it, no other easy way. It works, but it’s hacky. You know it’s wrong, but it’s also the only pattern that doesn’t make you lose your mind mid-build.
The bigger issue is how the Next.js docs completely skip over optimistic updates. They talk mutations, but they never mention the real world case when you want to instantly reflect a change on the UI before the server confirms it. Every real app needs that, but it feels like this case was never even thought about. Components rendered by the React Server can’t be modified after mounting by design. Anything that could possibly change needs to live inside a Client Component but then Client Components can’t do data fetching. Even during SSR. So we end up with these ultra-split architectures that look elegant in theory and feel broken in practice
The end result? Server Components that only exist to fetch data, and Client Components that carry around static HTML versions of your app. It’s the new form of SSR spaghetti just with fancier buzzwords.
Honestly, I’m done forcing it. Whatever my next personal project is, imma use TanStack Start. At least there, data fetching, mutations, and optimistic updates make sense. The flow feels natural again. You don’t have to fight the framework to build a simple CRUD interaction. Lamo 😋
btw I love tanstack
RAG is a important topic in AI that you need know about.
If you don’t have time for a course to learn about RAG, this thread will teach you everything about it in 2 minutes:
I have worked with pretty much every major AI model/provider. They all have strengths, but man do they have some weird issues too.
I've been keeping a list. Decided to share because why not? Here's everything I hate about Gemini, Claude, ChatGPT and Mistral.
@signulll Hey, I'm so excited to share this with you! 🤩 It's an AI platform called https://t.co/MsqbIpLR0C, and I think you're going to love it. 🥳 Go check it out and let me know your thoughts! 😄 #AI#tech#newplatform
🚀 Introducing Dashgen: Your centralized AI Assistant Hub. Seamlessly integrate multiple AI models using your own API keys, all in one intuitive platform. Experience the future of AI interaction here: https://t.co/hlOnFIodrK
#AI#MachineLearning#AIInnovation
@atulrp@AppLaunchKit It's looking really cool, @atulrp! This stack will definitely save a lot of time and boilerplate. Excited to see what can be built with it! 🚀
Devanagari is an extremely elegant script. But this was never explained to us in school.
A thread on the awesomeness that is devanagari.
Let's start with the things that my teachers did *not* teach me in school: