Ratul Dawar

My journey to crack media(.)net 🧵

I once worked at a company where SWEs were discouraged from asking questions. You were supposed to find answers on your own through reading the docs and code, searching, and debugging. It taught me a lesson.

Current meta

@arpit_bhayani He was trying to type using voice IG

arpit1212 is high, I guess :)

Another @ApacheDataFusio backed startup got funded openobserve.ai/blog/series-a-…

@johncrickett The other competitors like gpt gemini. Are no where closer to the thinking and understanding ability of opus 4.7 have been using gemini for past 3 months and recently switched to claude and difference is like very very clear

@ratuldawar18 There are plenty of other options, they're a commodity.

I think Dario Amodei is right when he says: "Coding is going away first, then all of software engineering." At least in his world at Anthropic. They tell us they don't write code any more. Their status page shows us they don't do software engineering.

@eatonphil Also what I feel here that many developers who used libraries without knowing a lot pf stuff had a illusion that they were writing good code while actually never ever seeing in their life what good code actually is

Lack of understanding, high-level libraries and languages let developers write really bad code already. It was good to learn how everything works even if you were just going to use a library that does everything for you. Advances in AI generated code don't remotely change that.

Rust team looks about as expected.

@striver_79 Bro be advertising a leetcode clone like it’s some disruptive tech

When it starts feeling like family, you know you’re doing something right. Last night, the team stayed till 2 AM. Not because they had to - but because they wanted to. Setting up every small detail. Making sure everything feels perfect. This isn’t just an office. It’s theirs. As a founder, you can plan, you can push, you can guide. But moments like these, you can’t manufacture them. We open today, our first ♥️

You know you’re deep into Rust when: Option<Result<T, E>> no longer looks scary.

@vivoplt Indians making OSS the new JEE just like competitive programming is the saddest thing

In 2019, a backend engineer at a mid size startup in Bangalore started contributing to open source after work. Salary: ~₹12 LPA. Nothing big. Just fixing small bugs in libraries he used. Most of his pull requests were simple: 1. documentation fixes 2. small bug patches 3. minor performance improvements But he kept doing it consistently. Week after week. Over time his name started appearing everywhere: 1. commit histories 2. issue discussions 3. release notes Other engineers began recognizing it. Not from LinkedIn. From GitHub notifications. One day a US startup reached out. Their team had been using a library he contributed to. They had already seen his work. A few interviews later he joined them remotely. New salary: ~₹75 LPA. All because of small open source contributions he made after work.

@striver_79 All this for a tomfoolery for leetcode knockoff.

A year back, I attended an appraisal meeting. Today, I will be on the other side of the table. Honestly, I am more nervous and tensed about this than I thought I would be. - How will everyone take it? - Will everyone be happy? - Will I be able to help them improve individually with the right feedback? - Will they stay here post this? - There are just so many thoughts in my mind. Running this isn't easy as it looks from the outside, there are so many factors.

@championswimmer I love when @striver_79 is showed the reality on X instead of the bootlickers he sees on linkden and youtube.

> make videos > good money > leave job > start edtech > hire a team > team grows big > edtech market cools down > hard to make payroll > drama on social to keep revenue up Probably the 5th or 6th cycle of Indian edtech hell by now Takes the best and turns them into ghouls Sad

@SkySingh04 My fav genre on twitter is @striver_79 rage bait

Damn we almost gave a fuck Honestly if in 2026 you are still watching DSA lectures and "One-Shot Physics Wallah" typa shit, You were ngmi anyways I love seeing people just take their JEE mindset and ctrl c ctrl v it to University / Job Search and hope it works out LMAO

Published article on ACID Properties. Not just theory but also practical implementation using PostgreSQL with real-world examples. This blog is perfect for beginners who want to start learning about Databases, and it is also a prerequisite for my PostgreSQL Internals series. Link: @shivambhadani_/acid-properties-explained-a-complete-guide-with-postgresql-implementation-071afa6aaf8a" target="_blank" rel="nofollow noopener">medium.com/@shivambhadani…

@arpit_bhayani sir do check it out

While going learning rust I came across this amazing concept : Do check it out medium.com/p/the-hidden-p…

@striver_79 DSA circle jerk

IIT-H appreciation flowing in 🔥

@striver_79 I mean leetcode is dead it’s just a matter of time the industry will shift to a newer hiring form its just that its convenient right now.

Last I checked, HackerRank still offered DSA rounds to hiring partners. How about scraping that? Or is the statement just to hype your recent AI products, which may have a bad adoption rate? It’s a very interesting time to be in. - Fundamentals matter more than ever. - You cannot afford to join as a noob anymore. The entry bar has shifted up several levels compared to previous years. - If someone thought “writing code” alone was the skill, they probably misunderstood engineering. Writing code was always the execution layer. The real skill was problem framing, trade-offs, debugging, and systems thinking. There’s a reason junior engineers were primarily tasked with writing code. Hiring at scale has never been about finding the best candidate. It has always been about reducing false positives and filtering the median. The truly exceptional candidates are rarely discovered through mass OAs anyway. They get hired via: - ICPC / competitive programming signals - Strong referrals - Prior proof of work - Direct poaching - Good research bg Now add AI into the mix. - How do you define “good AI usage”? - Who defines acceptable augmentation vs over-reliance? - Is effective AI leverage a skill, or a shortcut? If “getting the work done” becomes the only filter in the initial round, almost everyone clears it. If you bring everyone into the office and monitor them solving problems live: - Cost increases dramatically - Throughput drops - Hiring velocity suffers So the real problem is this: In an AI-augmented world, how do you build a scalable filtering system that: - Minimizes false positives - Preserves genuine signal - Doesn’t explode cost - Doesn’t punish legitimate AI-assisted productivity It’s a filtering architecture problem under adversarial conditions. And there’s no clean solution yet. But one thing we are sure of: orgs will take a bet on people who are multi-dimensional going forward. My guess is, we will have a mixture of every thing. They won’t cut down anything. You are expected to know more with AI 😛

@github I think we should no longer need to clone the repo. Please expose apis to make code changes directly on github so that claude doesn’t need to clone the repo. @AnthropicAI