Chanukya

Interests: Data Engineering | Distributed Systems | ML | తెలుగువాణ్ణి

[Chanukya @iChanukya]

https://t.co/XAMfVfdwjV

[Chanukya @iChanukya]

Awe Great.. 😇

[Nic Cruz Patane @niccruzpatane]

From concept to reality.

@SpaceX has successfully deployed Starlink V3 satellites into Earth’s orbit and made contact with them for the first time.

At scale, one Starship carries 60 V3 satellites, the same network capacity as about 20 Falcon 9 launches.

[Chanukya @iChanukya]

@mim_djo Nice summary

[Chanukya @iChanukya]

Tiny PR of mine to pyspark-data-sources: merged last August, shipped in v0.1.11 in January, and I only noticed recently.
It adds a JSONPlaceholder source for Spark's Python Data Source API.

Wrote up what the engine does with a custom source:
https://t.co/Dc56Iemq6g

[Chanukya @iChanukya]

@mim_djo @viggy28 it is massive. releases come to Databricks at first, and slowly to OSS. for eg., DLT was on Databricks long time ago and SDP was on OSS later!

[Chanukya @iChanukya]

@akshay_pachaar what was the reasoning for naming typesafe?

[Chanukya @iChanukya]

@GergelyOrosz I understand its complex, but looks like this is the way forward. not super comfortable with agent to deploy... as of yet..

[Gergely Orosz @GergelyOrosz]

Here's what OpenAI's agentic software factory looks like, today.

Details: https://t.co/UUZufLxG2r

(thanks to all the OpenAI folks who explained how it works! And Perf Factory looks especially interesting to me)

[Chanukya @iChanukya]

🫠🫠🫠

[Pritam Chauhan @Im_pritam18]

Dr. Pratosh at IISc Bengaluru tells his students that rapid advances in AI are commoditizing intellectual labour.

He raises an unsettling question: if companies stop recruiting on campus in the next 5–10 years—even at India’s top universities—what will the purpose of a university education be ?

A must watch video for everyone in tech

[Chanukya @iChanukya]

dictionary encoding isnt really something we can force. its more like a broadcast hint, spark tries and quietly gives up when it doesnt fit. 50gb wont broadcast, same thing. row groups are the bigger issue tbh. they're sized in bytes not rows, so theres just no row count setting. may be, duckdb if you actually need N rows per group, order by sorts everything globally, cluster by only within partitions. neither changes the data, both change where it sits. thats enough to change what the writer sees per row group.

[Chanukya @iChanukya]

@VaibhavSisinty Nice! Much needed!

[Chanukya @iChanukya]

@s_shinobi11 @JespernissenSEO ohh!!

[Chanukya @iChanukya]

- rails are the easy part. the real gap is trust: which vendors can an agent actually verify before paying them autonomously? UCP and MCP dont answer that.
- building https://t.co/TYZp6EcJDZ for exactly this: an open, independently governed vendor registry sitting above the protocol layer. spec + reference impl at https://t.co/FBIMFgHZkx

Would love to connect with someone who's working on this to compare and reflect on the ideation of the registry..

[Chanukya @iChanukya]

@araseb_ You need to pitch idea, thought process, gtm fit..! a lot..

[Chanukya @iChanukya]

good work, this is a real synthesis. openclaw's always-on local capture + karpathy's llm wiki pattern for compounding it, fused into one product. mine takes the opposite filter; only what i choose to keep goes in. voice capture parsed to text, long posts as md files, chats; all curated, all lands in the brain. https://t.co/S6l0hzKT6R

[Chanukya @iChanukya]

@thinkymachines This is cool! Is it so that sound travels fast and thats why audio modality is much faster for humans to reciprocate than the visual? Can you drop that analysis?

[Chanukya @iChanukya]

voice mode on claude code, just works fine! instead of text based response, voice based response would be ideal! #ClaudeCode

[Chanukya @iChanukya]

@mustafasuleyman Nice!! with whom Microsoft competes? Google/OpenAI/Anthropic?

[Y Combinator @ycombinator]

"You need to make yourself a big target for luck, and the way to do that is to be curious. Try lots of things, meet lots of people, read lots of books, ask lots of questions."

— Paul Graham, How to Do Great Work

https://t.co/AyIk5fnKZL

[Chanukya @iChanukya]

@ThuHuong2806 @PaulSolt its not that difficult with xcode and claude code. you can develop your own app to your device alone - with basic feature to experiment. once successful, then you can scale with intended features;