Dhirendra

Now @PlanetScale is pushing the same idea that helped YouTube massively scale MySQL: Vitess + sharding + horizontal scaling. Instead of replacing MySQL, they scaled it. Pretty interesting to see where this goes next. Massive-scale database engineering is fun stuff. Waiting for the updates.

@kunal_twts Yes it is for sure

@cmd_dj221 Go is solid

@SrujanGowda_10 I would love to see it crash, that's how we learn and grow, we make things and then break them on purpose to learn

Read an article today about DB locks and decided to run an experiment. Writing thousands of pings a second directly to a database is a great way to watch it crash instantly. I'm building a demo this week with Kafka and Postgres just to test this out. I want to put Kafka in the middle as a buffer and see exactly how it handles that massive amount of writes without crashing the DB

@Arpit_2023 I guess I also need to do some cleaning

I have unfollowed a lot of dead accounts Let me know if I accidentally unfollowed any active ones

@arcbitbit Yes it's a very good alternative and honestly is used in many modern systems now as it can also help in microservices architecture and ML models also, so slowly its becoming first choice these days

@cmd_dj221 FastAPI is indeed a better alternative, no abstraction everything is on developer.

If I had to restart Python web development today, I would NOT start with Django. Not because Django is bad (it’s amazing and comes with batteries included). But because there’s a catch: Django abstracts a lot: - authentication - sessions - admin panel - many under-the-hood implementations Great for productivity. But sometimes it hides the internals beginners should actually learn. I faced this myself. Early in my career (back in 2018–19, before AI tools everywhere), I was building fast but not fully understanding what was happening behind the scenes. If I had to start over: I’d spend a 6-8 months with Flask or FastAPI first. Build: - CRUD APIs - authentication from scratch - queues, caching, middlewares - understand request flow & app lifecycle Then move to Django. You’ll not just use the batteries included, you’ll understand them too. Sure, you can absolutely go deep with Django itself. But exploring micro-frameworks first gives you both: speed + deep understanding. You can absolutely start with Django as well but this is the path I’d choose today for myself.

@Parthak32315 Welcome bro 🤗

@cmd_dj221 thank youu sirr jii

@Anishdotcom You are learning dsa concepts or solving problems? Or both ? How u basically proceeding ??

2/n - Revised Decision Trees - Practiced lot of DSA today - Worked on Project could have worked more - Steps 8k - Gym legs - Read : Forge you future

@Anishdotcom Ghante ka sahi istemal kar rhe ho aap, hope i can also utilise my time well

1/n - Studied Random Forest from Campus X - Practiced Dsa - Working on a project - Steps : 15k - Gym : Arms done - Skin Care routine done - Watched Anime

@Anishdotcom Welcome 🤗

@cmd_dj221 damn sirji thanks

@Yami_codes Congo 🥳🎉

Pass hoooo gayaaaaaa 🥳🥳🥳

@neuroquark That's really very good progress you are making that too fast, keep doing

finished writing a simple api in golang using the net/http package. it's a basic task creator with no auth and 3 endpoints also built a docker image for it and was pretty happy to see it come out at 20.6 mb. way better than my previous bun image that somehow ended up above 250 mb. still not sure if that was because of my skill issue

@parthdot404 Great analogy that's really the thing

back in the day we used to have slot machines, now they just ask llm and expect their desired output

@OjasSharma276 it's a decent amount to start with especially if one doesn't have to pay rent

Is 50k per month for a beginner decent? Only from a job, especially if someone lives at home.

@parthdot404 @mahanot_dikshit Too much workload sirji, baaki aap batao how things going

@cmd_dj221 @mahanot_dikshit Sirji long time no see?

Atp i’ve understood followers doesn’t mean anything on X. Logo ke 30-40k followers h and 20-50 likes ni aa rhe

@OfficialVyshu Messy days are there for all of us, don't worry it will pass and do share your work

Day-2 To be specific i don't want to mention the journey i was in, i just want to mention the productive day i was going on Today i continued on the remaining topics of the programming language Thinking to wrap this up quickly with in couple of days Yup hoping for the best :)

@Harry_The_Nerd I go in afternoon if no calls are there that too in office gym 😕

Me entering the gym just before the closing time after a whole fooking day of PRs and Meetings at work -

@neuroquark So whats the exact solution for this? Any steps people should take? Or any site where they can check which library is compromised??

fyi: using bun doesn't completely protect you from npm supply chain attacks it still fetches packages from the same npm registry. sure it blocks lifecycle scripts by default but it maintains an allowlist for the most popular packages. if one of those gets compromised, its malicious scripts will still run. also it can't protect you from attacks hidden directly inside the codebase itself

@neuroquark Yes correct but It's not just about the language, it's primarily about the architecture, concurrency model, and how you use OS level I/O facilities in your application.but yeah most languages now incorporate these concurrency and sync behaviours

@cmd_dj221 if i'm not wrong then i guess most languages can solve this problem

Modern servers handling millions of connections? It all traces back to a problem from the late 90s called C10K. It emerged in the late 90s and asked a simple but hard question: How do you design a server that can handle 10,000 simultaneous client connections? "C = Clients / Connections" "10K = 10000" Why was it hard? When this concept emerged in late 90s, Back then servers mostly used 1 thread/process per connection Which had limitations like: - high memory usage - context switching overhead - OS scheduling bottlenecks Servers struggled to handle even a few thousand connections. The C10K problem pushed innovations with whom we are familiar today like: - event-driven architectures - async programming - lightweight concurrency - efficient I/O & kernel support Few Systems that solved this C10K effectively are: - Nginx - Go(goroutines) - Netty(Java) - Rust async frameworks. In simple terms just think it like: Earlier each customer gets dedicated waiter and in new one smart waiter can handle many tables using a proper system. Key takeaway is the C10K problem is solved by Avoiding 1 thread per connection and using event driven or non blocking i/o instead. But with solutions now, the another problem evolved called C100K and even C10M.