founding eng. @niledatabase. 21. i love postgres performance, analytics & fast bikes.
ahmedabad
Joined July 2024
- Tweets3.7K
- Following564
- Followers811
- Likes11.1K
Pinned Tweet
hello there! i got a job.
i'll be joining @niledatabase as a founding engineer, which means I GET TO WORK ON DATABASES FOR A LIVING.
i've been and i'm really excited for this. feels like a rare opportunity to work on infra like this full-time with a team that's deeply strong at what they do. i could not think of a better way to learn and to start my professional career. thanks so much @gwenshap and @sriramsubram for this.
we're so so so back.
i am happy for sunil as he moves ahead in life but in all honesty this is equally shocking to read as was watching ironman die in avengers endgame.
regardless, goated run by the goat.
one thing i've noticed this week is that a lot of people seem to read a new CVE as "someone recently introduced this vulnerability."
that's often not what happened at all.
a CVE being disclosed today can mean a bug that has quietly existed in the codebase for months or years was only just found, reported, and understood.
with how insane this week has been across PHP, Node, Railway, etc., that distinction is very important.
and the slightly unsettling part is that "we just discovered it" != "nobody knew about it before us."
some vulnerabilities can sit there for a very long time. it's entirely plausible that someone sufficiently capable found one earlier, used it, and simply never told anyone.
i've had a hard time finding people irl who are ambitious about systems, or even tech in general, and actually want to build things since i come from a not-so-tech-savvy city.
i've also found that this is a common problem with many of us not blessed to be living in sf.
so i'd like to compensate for that by meeting more of you here on x. if you're building interesting things, especially around systems/databases/infrastructure, i would love to have a chat.
as i try and build systems from scratch, i keep realizing the same thing again and again.
it's how hard it is to build conviction about which problems are worth solving when you're still early in your career. this is further accelerated by the advent of agentic coding.
llms lack conviction and so far i think it can only be built by having human mentors and real failures.
a lot of the tedious work that used to force you to understand a system can now be done much faster. that's useful, but I wonder if it also means junior engineers get fewer chances to develop a feel for how the system behaves, where it breaks, and what is worth investigating.
i don’t have a clear answer yet. i'm trying to spend more time working directly with Postgres and learning from people who have operated it for a long time here at @niledatabase.
also this.
nitter.cf/carllerche/status/2112…
"Good compression is an art" love it.
in all honesty, its depressing how many things i enjoyed doing have become prompts.
it kinda sways me away from working on things. it's been a while since i touched a side project because my brain has defaulted to thinking "it's just so easy" and that makes it so i never get to it.
i crave difficulty and fail to understand the new way of finding it. does anybody have any help/advice regarding this?
for a second i thought he was talking about RAM and arenas.
have i been left too far behind in the llm users race?
coding by hand is therapeutic. i’ll find myself writing a postgres proxy in rust at 2am on weekends and it genuinely feels healing.
absolutely recommend it, even the boring, tedious parts. i think those might actually be my favorite bits now.
winit retweeted
the point of being a hater is having taste. For everything you hate on you must have something you just as enthusiastically recommend. Otherwise you’re a cancer upon society
Isn't the end level of "build your own software" a massive AI labs/inference-provider profit since that'd drive the token demand to the moon and beyond.
i.e, everyone stabs everyone else and then in the end we're all dependent on our ai overlords for token subsidies. everyone prompts everything.
Some acquisitions feel so inevitable that the surprising part is finding out they hadn’t happened already.
The team behind @Deno_Land, including @rough__sea (he knows a bit about Node.js), is joining @Cloudflare.
They’ll help make Workers better for Node.js developers and advance workerd as a core primitive for what comes next. cfl.re/4zlzlSa
