stuff @alma_inc | ex intern @cisco/ex research @ai4bharat https://xitter.cf/t.co/uyqch55y1f (30k users) | https://xitter.cf/t.co/UfNnMe4nMe (4.5k users) | https://xitter.cf/t.co/vcv1vDMWRC (1.5k users)
India
Joined April 2024
- Tweets27K
- Following8.8K
- Followers51.2K
- Likes8.1K
Pinned Tweet
lattedit is now a desktop app
Projects are real folders on your disk. Compiles offline with your own TeX. ~4 MB. Free.
Linux today, macOS & Windows soon:
lattedit.xyz/install
launching lattedit.xyz
A calm little LaTeX editor for your browser, Free, no account, works on your phone. LaTeX, freshly compiled.
check them out :) xitter.cf/cneuralnetwork/status/…
first tweet from Apple M5 MacBook Pro (Apple M5 Chip/16 GB/1 TB SSD/MacOS/Liquid Retina XDR), 35.56 cm (14 inch), Space Black
Used Treg right now to find some leads for AI startup founders.
It actually found so many results in such a short time and only cost me a few cents.
Gonna be a power user now :)
🤝 Paid partnership
Introducing OpenSource Recruiter for Claude
- Paste any job post → 50 ranked candidates
- Personal email + mobile, $0.20/candidate
- No $170/mo seat, no 30-InMail limit
- 0% markup, fully open source
Try it at treg.to/recruit
Repo link below 👇
I can't just express today how emotional I am with this MacBook purchase as well, right? I'll just tell you: two years ago I had asked for ₹7 from my girlfriend to have one cup of tea. That was my financial scenario two years back.
I just woke up and did shit every day. I still do shit every day and it just means so much that today I can go into a shop and buy this. I just cannot express in words how much this means to me as well because it's so... you feel proud of yourself.
Like from having ₹10 in your bank account to working hard every single day, showing up every day, and doing everything possible in your things. If, through multiple rejections from multiple companies, you still show up every day and do get this done, that just feels so good at this point: I'm getting some kind of fruits, some kind of rewards from all the hard work I have done. I just want to say that it is possible to work hard enough.
I faced this earlier as well and people are going to say this spec is bad. That is going to happen tomorrow. You could have gotten it for like 10,000 less and everything but that's never the point with this purchase.
Some people will have good experiences, some people will have bad experiences, but you'll never get an absolute answer to choose whether to buy. This is kind of like a statement to myself that this is an achievement for me.
That today I could go into a store and buy something that took me 2 years of hesitation: that is the happiness that comes to us. No amount of RAM, no amount of storage will decide that.
what actually happens when two devices have the same IP address on a network?
you might expect the router to reject one of them, but that's not necessarily what happens.
the interesting part is how ARP handles this, and why connectivity can randomly break.
you can inspect this on Linux:
ip neigh show
or watch ARP packets:
sudo tcpdump -eni eth0 arp
if you repeatedly see the same sender IP address advertised by different source MAC addresses, that's a strong indication of a duplicate-IP conflict.
you can also use arping -D to perform IPv4 duplicate-address detection.
the important distinction:
ARP does not maintain a globally agreed ownership table for IP addresses.
each host maintains its own neighbor cache.
that means different machines can simultaneously associate the same IP with different MAC addresses.
and that's why duplicate IPs don't always result in a clean, immediate failure.
sometimes the network works, sometimes it doesn't, and sometimes different hosts reach completely different devices.
i built a tiny DNS resolver in C because i wanted to understand what actually happens when we type google.com into a browser.
the interesting part is that DNS is a surprisingly simple protocol at the packet level.
here's how it works:
the most interesting part for me was parsing the response.
DNS responses can compress repeated domain names using pointers.
if the first two bits of a label byte are 11, the remaining 14 bits across two bytes encode an offset into the DNS message.
so instead of repeating google.com in the answer, the server can just point to its earlier occurrence.
a pretty neat space optimization that's been part of DNS since RFC 1035.
and that's pretty much the core of a basic DNS lookup.
encode a domain name, construct a query, send it over UDP, parse the answer, and extract the IP address.
of course, a production resolver needs much more: retries, TCP fallback, proper response validation, caching, and security considerations.
but implementing the packet format yourself is a really nice way to understand DNS beyond just calling a library.
i'll probably build more networking protocols from scratch in C.