Maanas Tyagi

@emollick I dont think its gon be a big deal tbh

(Sorry, after seeing so many of these, could not resist): 🚨 BREAKING: Google just dropped a NEW paper that completely deletes RNNs from existence. No recurrence. No convolutions. Nothing. Just one mechanism. And it’s destroying every translation benchmark on the planet. The title alone is a flex: “Attention Is All You Need” Vaswani. Shazeer. Parmar. Uszkoreit. Jones. Gomez. Kaiser. Polosukhin. 8 researchers. 1 architecture. The entire field of NLP will never be the same. Here’s why this is INSANE → LSTMs took DAYS to train. This thing trains in 12 hours on 8 GPUs. 🤯 → 28.4 BLEU on English-to-German. That’s not an improvement. That’s a MASSACRE. They beat the previous SOTA by over 2 points. → English-to-French? 41.8 BLEU. At a FRACTION of the training cost of every model that came before it. → They called it the “Transformer.” The name alone tells you they knew. But here’s the part nobody is talking about 👇 They threw out sequential processing ENTIRELY. Every other model on Earth processes words one at a time. This thing looks at the ENTIRE sentence simultaneously and figures out what matters. It’s called “self-attention” and it’s basically the model asking itself: “which words should I care about right now?” Every. Single. Token. In parallel. Do you understand what this means? Training that used to take WEEKS now takes HOURS. Models that couldn’t scale past a few layers? This thing stacks 6 encoders and 6 decoders like it’s nothing. And the multi-head attention? 8 attention heads running at once, each learning DIFFERENT relationships in the data. I’m not being dramatic when I say this paper just rewrote the rulebook. RNNs are cooked. 💀 LSTMs are cooked. 💀 The future is attention. And attention is ALL you need. Follow for more 🔔

@pow00661 @Cereal_fan I mean to say cost of inference is high. As the tokenizers are largely trained on eng vocab. Check out sutra tokenizer and it's papers

@maanas_tyagi @Cereal_fan What do you mean "Inference is to high"??

Chatgpt is fluent in Indian languages What is the use of Sarvam AI again ?

@Cereal_fan Wdym tokenizers improve? At the. End of the day tokenizers for sota models are largely trained on English vocab. Check out sutra and papers on it

@maanas_tyagi 🥱🥱🥱🥱 It won’t matter soon as costs come down and tokenizers improve. Next subject.

@coderaw_ Pythonn and C

Be honest, which programming languages can you code without using AI?

@SakshiSugandhi Just a degree is useless. DUH

Be honest. If your younger brother asked you today "should i do CS degree" what would you actually tell him? 💀

A LLM trained on data before 1930 check it out talkie-lm.com/chat

@Sakshi50038 You need to understand,f or many years now engineering has been the path to success as it was easy to find job, when it used to be a respectable proff. Things like music and lit. are only critised if you are not earning from it. At the end of the day money matters.

India Has One Religion CSE India didn’t just favor one career path it practically declared CSE the only acceptable way to exist. Everything else? Mocked, dismissed, treated like failure. Arts are “useless.” Philosophy is “unemployable.” Music is a “side hobby.” And if you pick anything outside the script, suddenly it’s a family crisis. So now you’ve got millions of engineers who can build systems, write code, and follow instructions but struggle to question, imagine, or create something original. Because the system never valued that. It trained execution, not thinking. The truth is harsh: innovation doesn’t come from just building it comes from knowing what to build, why it matters, and how it connects with people. That’s where philosophy, design, and storytelling come in. And India didn’t just ignore them it actively discouraged them. This isn’t about capability. It’s about a culture that crushes curiosity at 16, forces everyone into the same mold, and then wonders why originality is rare.

@BLUECOW009 I WANNA JOINNNNN

i wanna start a lab, reply to be a candidate

This is it guys. People who are not needed will be wiped off.

@CodeWithAmann Remove the app

If you were X CEO today, what’s the first change you’d make?

@sahill_og Learn new shit brother. New things do come, but the foundations remains the same. Make your foundations strong.

Real question: How do you stay relevant as a developer when the tools change every 3 months?

@TechByTaraa I think Fast API if its a heavy application. But for regular use I think Flask is good.

As a dev, what do you prefer for backend?

@YashHustle_22 CODEX AND ANTIGRAVITY

What are you using now? Claude or Codex

Starting Coding for the first time Which IDE to use?

@md_kasif_uddin @obsdmd ANY BLOOODY DAY. I am in LOVE with that Software.

What’s the best note taking system for developers?

@PratikSinhatwt Where to keychain from?

What should I install first in my ThinkPad

@0xPrajwal_ TRUE. I USED TO BE AN AVID USER OF IT. BRO WILL BE MISSED

Bro disappeared like it never existed.

@Heymaxi01 Bruh Anyone with basic RAG knowledge can ans that

ai intern interview question

@omarvvvr No Windowss

What's the biggest problem with MacBooks?