Shikhar

@ChShersh I am sorry for their loss.

My talk proposal was rejected.

@elonmusk @doganuraldesign Finally someone said it

@doganuraldesign Me too

Fuck it, man. I’m so sick of these algorithm changes.

@hackSultan Hey Claude, be dumber!

lol. Suffering from success.

@cy1nus How you gonna be in the morning

I don't wake up at 6 am

@Chiebuniem_ Bro has an entire AR system embedded in his helmet.

Tony Stark had a whole 32 inches ultra wide monitor inside his Iron Man helmet. Please let’s be serious.

@iBuild Google is the only lab working on all code, image, video, and music generation as well as world models at the same time.

google really cooked with their new cinematic video overview in notebooklm. > visualizations are so much more powerful now > mostly it hits too well, sometimes can be not so cinematic > overall? brilliant. loved it

@slow_developer It's natural he will hype it because he sells the shovels, but in this case I agree with him. Although I wonder if agentic will stick in the future or will be replaced by some more appropriate term.

Jensen Huang says all future software will be agentic — because why would you build dumb software? Every company will use a mix of open models they fine-tune, closed models they rent, and specialists they contract "your job isn't to do the work, it's to get the work done"

@original_ngv The actual Dhurandhar was in front of us all the time.

Dhurandhar isn't just one person

@curlysaarthak RAG is dead. Long live RAG.

> "rag is dead" explains a simplified versin of graphRAG

@elonmusk @SawyerMerritt @SpaceX The question is when.

@SawyerMerritt @SpaceX SpaceX will ultimately deliver millions of tons to the Moon to build a self-growing city there and same for Mars

NEWS: NASA is planning a bigger @SpaceX Moon mission role using Starship, in a massive blow to Boeing. With the new proposal, Boeing's SLS would no longer be used to boost Orion close to the moon. Instead, Starship and Orion would dock in Earth orbit, giving Starship the pivotal role of propelling the capsule to the moon’s orbit, before taking astronauts down to the surface. bloomberg.com/news/articles/…

@freshlimesofa Einstein would be a Claude user, I'm sure.

Einstein would've loved vibe coding I'm sure.

@harsh567_8 Yeah, that's true

@shikhr_ i think the competition won't decrease if we take stress , ulta it Will distract us to keep working jisse aur stress hoga nd cycle continues

chill out you stress out on bullshit things . someone is at -: 1600 on CF ? 2000 on leetcode? 40% are cheating .... web dev projects? must have used AI not from scratch. people are cheating in OA nd getting rejected . don't stress looking online so much . work hard on yourself, you only need one job , don't be scared by less no of jobs .

@BillyM2k Yeah sure, like the high IQ people are wasting time posting on X

instead of filtering by country, i would like to filter replies by user iq

@Divyansh91565 Yeah, Odd levels and even levels form separate groups.

@shikhr_ I think it's obvious like if we pick any node as root and color nodes by 0 and 1 as basic concept of bipartite. Since a tree has no cycles, no edge can connect same-depth nodes whch means every edge connects opposite colors hence bipartite.

So I was solving a graph problem and the core intuition turned out to be just checking whether the graph is bipartite or not (plus a few extra conditions)... So I thought maybe a article kind of , on this.. If you already understand how BFS/DFS works, here's a simple way to check bipartiteness: Bipartite Graph Check - > Most of you already know BFS/DFS. > But many graph problems reduce to a simple question: > Can we divide nodes into 2 groups such that no edge connects nodes of the same group? If yes then graph is bipartite Let's understand the core Idea: Instead of thinking in groups, think in colors: > Assign each node a color (0 or 1) > For every edge -> colors must be different So the problem becomes: Can we 2-color the graph without conflicts? Basic Working: 1. Start from any node: 2. Assign color = 0 3. Their neighbors = 1 4. Continue using BFS/DFS While traversing: If node is unvisited -> assign opposite color If already visited and same color -> NOT bipartite Why This Works? Bipartite => No odd cycle 1) Even cycle -> works fine 2) Odd cycle -> forces same color conflict First-Principles Thinking behind it: This boils down to these points At any node: 1. You only care about previously visited neighbors 2. Constraint: neighbor must have opposite color 3. So each step is just: -> Check conflict -> Assign color -> Continue That's it.

@dazzBuilds It's a good question which I didn't think of for a very long time when LLMs were new

@shikhr_ 🤯

Modern LLMs can output structured JSON using schemas. But here’s the catch: LLMs are probabilistic models. So how do you ensure at inference time, the response always matches your schema? Hint: Theory of Computation comes to the rescue.

@himanshustwts Good to see more labs are focusing on coding models.

cursor did their first continued pre-training run is a news to me. they have ultimately a fancy research team (especially RL) and bros are coming for any of frontier labs (in coding atleast) exciting times ahead.

@craft_withcode Great! Did you learn how to use Redis in code or how it is used in distributed systems?

Today's log : ~POTD ~ 2 questions on hashing ~backend learn about redis ~ networking CIDR (Celebrate new year with narmada arti.)

@kartikktwt Cool, I gotta visit these lectures as well

done with Lecture 2 this lecture was mostly about golang > goroutines > IO concurrency > Parellelism > Async programming vs goroutines > thread challenges > waitgroups > mutexes > some goroutine examples