HNHacker News
TopNewBestAskShowJobs

netvarun

2,770 karma · joined September 11, 2009

Varun Sivamani

Founder, Keygraph https://keygraph.io/ We build AI pentesting software.

Check out our open source AI pentester, Shannon (47k+ stars): https://github.com/KeygraphHQ/shannon

submissionscomments
netvarun··on Ember-1
Technically kimi k-3 weights license is not open weight (it has a lot of restrictions). I would classify it as ‘weight open’ similar to the bsl and fsl ’source open’ licenses.
netvarun··on Ember-1
Off topic:With sol pricing drop tbh kimi k3’s value prop has not been that great. For our internal use case/testing/benchmarks sol come out with way better quality and much cheaper costs. Kimi really needs to drop their pricing (I heard it’s set by them across all the neoclouds) Sol is at 2/10 vs kimi’s 3/15
netvarun··on Xorshift Generators
https://archive.is/zElcb
netvarun··on Xorshift Generators
Still not loading
netvarun··on Show HN: Scroll through all 43252003274489856000 Rubik's Cube states
Wow, this was really fun to play around with! Amazing execution and launch!

Quick feedback: it’d be great if the page updated when the URL changes, so you can explore different positions without reloading.

Right now, if I change the URL/hash of the cube position (eg: incrementing by 1 https://everycube.alen.is/?view=3d#5319001606686939024 ), the app doesn’t update and I have to open the new URL in a separate tab to see the new position. [Chrome 150 on Mac OS X 10.15.7]

netvarun··on Fields Medals 2026
https://x.com/Qiaoqiao2001/status/2080003441821163958

While there was a bunch of human effort - it is not that hard to imagine this whole pipeline becoming fully autonomous in the coming months.

netvarun··on Fields Medals 2026
I posted a similar comment when the winners were leaked: https://news.ycombinator.com/item?id=48906573

i’d like to revise my earlier comment: 2022 may have been the last time we had pure humans win a Fields Medal.

I’m fairly certain this batch's winners used LLMs for research, lit-revews, reviewing work, and calculations... perhaps not enough to count as a co-author, but still enough to handle a lot of the grunt work.

Who would have imagined the pace of progress in LLM-powered math..

netvarun··on A digestion of the Jacobian conjecture counterexample
https://news.ycombinator.com/item?id=48986943
netvarun··on Codex scraped the ICM website and discovered 2026 Fields Medal winner list
First of all congrats to the winners.

Second, fitting that codex enters the picture.

The last time the fields medals were announced llms were still very nascent :)

And I am convinced this is the last time pure human fields medalists will be announced.

The next batch’s winners are all going to have llms as coauthors.

netvarun··on Show HN: I RL-trained an agent that trains models with RL (for –$1.3k)
“The Anthropic team for their incredible coding models (Fable-5 wrote every line of code in this project), and the Claude Code harness.” Source: the repo
netvarun··on NumClass – a Python CLI classifying integers into 200 number-theory properties
This is an amazing project - thank you for building/sharing it!
netvarun··on Detecting and countering misuse of AI
Shameless plug: We're building this. Our goal is to provide AI pentesting agents that run continuously, because the reality is that companies (eg: those doing SOC 2) typically get a point-in-time pentest once a year while furiously shipping code via Cursor/Claude Code and changing infrastructure daily.

I like how Terence Tao framed this [0]: blue teams (builders aka 'vibe-coders') and red teams (attackers) are dual to each other. AI is often better suited for the red team role, critiquing, probing, and surfacing weaknesses, rather than just generating code (In this case, I feel hallucinations are more of a feature than a bug).

We have an early version and are looking for companies to try it out. If you'd like to chat, I'm at varun@keygraph.io.

[0] https://mathstodon.xyz/@tao/114915606467203078

netvarun··on Ask HN: Why hasn't x86 caught up with Apple M series?
s/x84/x86/
netvarun··on Launch HN: Channel3 (YC S25) – A database of every product on the internet
Hey Shawn!
netvarun··on Launch HN: Channel3 (YC S25) – A database of every product on the internet
Hi - we should chat :)
netvarun··on Find Your People
I think they refer to: IB - investment banking PE - private equity HF - hedge funds or High Frequency trading (?)
netvarun··on OpenAI to buy AI startup from Jony Ive
Tangent: Did Windsurf actually get acquired by OpenAI? I would have imagined some sort of announcement from OpenAI at the very least? Bloomberg was the one to break that news too, but haven't seen any follow up.
netvarun··on Databricks and Neon
Does anyone have insight into Neon's financials - specifically their revenue, COGS, and gross margins? I'm trying to understand what made Databricks value them at $1B. Was it strong unit economics, rapid growth, or mostly strategic/tech value?
netvarun··on Canadian math prodigy allegedly stole $65M in crypto
https://archive.is/kVsvc
netvarun··on The Pain That Is GitHub Actions
Dagger (https://dagger.io) recently seems to have reinvented/rebranded itself as some llm agent platform.
netvarun··on Popular GitHub Action tj-actions/changed-files is compromised
@dang: The original URL (from Step Security, the company that discovered this flaw) is a better source for this:

https://www.stepsecurity.io/blog/harden-runner-detection-tj-...

netvarun··on A Second Search for Bash Scripting Alternatives
Agree - that’s what we use too.
netvarun··on Show HN: Retry a command with exponential backoff and jitter (+ Starlark exprs)
Wow that had been quite a journey - thanks for the detailed response. We’ve been using cel internally in a golang codebase and been pretty happy with it. I’ve only know about starlark in the Bazel context - I’ve learned a couple of things from your post. Thanks :)
netvarun··on Show HN: Retry a command with exponential backoff and jitter (+ Starlark exprs)
This looks cool - will give it a try (hah!) Curious on why you picked starlark instead of cel for the conditional scripting part?
netvarun··on OpenAI Acquires Rockset
Congrats to the team. IIRC their CTO was the creator of RocksDB.
netvarun··on Show HN: Primeval Numbers
Very slick interface and implementation! This is why I love coming to HN!
netvarun··on Show HN: Restate – Low-latency durable workflows for JavaScript/Java, in Rust
Thanks for the detailed answer - please turn it into a blog post! Excited to see competition and different architectural approaches to tackle durable execution. Wishing you all the very best!
netvarun··on Show HN: Restate – Low-latency durable workflows for JavaScript/Java, in Rust
Feedback: everybody’s question is going to be on why this over temporal? I’ve noticed you answered a little bit of that below. My advice would be to write a detailed blog post maybe on how both the systems compare from installation to use cases and administration, etc - I’ve been following your blog and while I think y’all are doing interesting stuff I still haven’t wrapped my head around how exactly is restate different from temporal which is a lot more funded, has almost every unicorn using them and are fully permissively licensed.
netvarun··on The man who killed Google Search?
"Be kind. Don't be snarky. Converse curiously; don't cross-examine. Edit out swipes."

"Please don't fulminate. Please don't sneer, including at the rest of the community."

https://news.ycombinator.com/newsguidelines.html

netvarun··on The man who killed Google Search?
I agree - and flagged it as a result. It is clearly a hit piece and has promptly even made it to Prabhakar's wikipedia page: https://en.wikipedia.org/wiki/Prabhakar_Raghavan
Page 1 of 10Next →