srvmshr's profile picture. ML University of Tokyo. Prev: Microsoft Research RF, @virginia_tech. Personal opinions. Coasting life with @jnchrltte

Sourav

@srvmshr

ML University of Tokyo. Prev: Microsoft Research RF, @virginia_tech. Personal opinions. Coasting life with @jnchrltte

I jumped into the @windsurf camp today. Some of the features they integrate are so cool - like deepwiki and codemaps. Only feedback: please integrate other models via Openrouter for e.g... BYOK is great but limited only to Anthropic


Sourav 已轉發

Claude Code's native installer is now generally available. It's simpler, more stable, and doesn't require Node.js. We recommend this as the default installation method for all Claude Code users going forward.

claudeai's tweet image. Claude Code's native installer is now generally available.

It's simpler, more stable, and doesn't require Node.js. We recommend this as the default installation method for all Claude Code users going forward.

Sourav 已轉發

Hot take: DAgger (Ross 2011) should be the first paper you read to get into RL, instead of Sutton's book. Maybe also read scheduled sampling (Bengio 2015). And before RL, study supervised learning thoroughly.

shaneguML's tweet image. Hot take: DAgger (Ross 2011) should be the first paper you read to get into RL, instead of Sutton's book. Maybe also read scheduled sampling (Bengio 2015). And before RL, study supervised learning thoroughly.
shaneguML's tweet image. Hot take: DAgger (Ross 2011) should be the first paper you read to get into RL, instead of Sutton's book. Maybe also read scheduled sampling (Bengio 2015). And before RL, study supervised learning thoroughly.

Sourav 已轉發

Ever wondered about graph learning? Watch Ameya Velingker (@ameya_pa) and Haggai Maron (@HaggaiMaron) give a masterful introduction at the Simons Institute's workshop on Graph Learning Meets Theoretical Computer Science. Video: simons.berkeley.edu/talks/ameya-ve…

SimonsInstitute's tweet image. Ever wondered about graph learning? Watch Ameya Velingker (@ameya_pa) and Haggai Maron (@HaggaiMaron) give a masterful introduction at the Simons Institute's workshop on Graph Learning Meets Theoretical Computer Science. Video: simons.berkeley.edu/talks/ameya-ve…
SimonsInstitute's tweet image. Ever wondered about graph learning? Watch Ameya Velingker (@ameya_pa) and Haggai Maron (@HaggaiMaron) give a masterful introduction at the Simons Institute's workshop on Graph Learning Meets Theoretical Computer Science. Video: simons.berkeley.edu/talks/ameya-ve…
SimonsInstitute's tweet image. Ever wondered about graph learning? Watch Ameya Velingker (@ameya_pa) and Haggai Maron (@HaggaiMaron) give a masterful introduction at the Simons Institute's workshop on Graph Learning Meets Theoretical Computer Science. Video: simons.berkeley.edu/talks/ameya-ve…
SimonsInstitute's tweet image. Ever wondered about graph learning? Watch Ameya Velingker (@ameya_pa) and Haggai Maron (@HaggaiMaron) give a masterful introduction at the Simons Institute's workshop on Graph Learning Meets Theoretical Computer Science. Video: simons.berkeley.edu/talks/ameya-ve…

Tired to go back to the original papers again and again? Our monograph: a systematic and fundamental recipe you can rely on! 📘 We’re excited to release 《The Principles of Diffusion Models》— with @DrYangSong, @gimdong58085414, @mittu1204, and @StefanoErmon. It traces the core…

JCJesseLai's tweet image. Tired to go back to the original papers again and again? Our monograph: a systematic and fundamental recipe you can rely on!

📘 We’re excited to release 《The Principles of Diffusion Models》— with @DrYangSong, @gimdong58085414, @mittu1204, and @StefanoErmon.

It traces the core…


Different places, different systems 'Tenure' makes no sense in S Asia, but has a lot of weight in N-Am univ. Similarly, one could be a Lecturer/Asst Professor with just a Masters in Asia, and only seek job progression after a doctorate. Best not to have preconceived biases

Received a PhD application email from an Assistant Professor at another county (for themselves). I am so confused at multiple levels.



Sourav 已轉發

Our Huxley-Gödel Machine learns to rewrite its own code, estimating its own long-term self-improvement potential. It generalizes on new tasks (SWE-Bench Lite), matching the best officially checked human-engineered agents. Arxiv 2510.21614  With @Wenyi_AI_Wang, @PiotrPiekosAI,…

SchmidhuberAI's tweet image. Our Huxley-Gödel Machine learns to rewrite its own code, estimating its own long-term self-improvement potential. It generalizes on new tasks (SWE-Bench Lite), matching the best officially checked human-engineered agents. Arxiv 2510.21614  With @Wenyi_AI_Wang, @PiotrPiekosAI,…

Guide-coded a high throughput document pipeline using @allen_ai OLMO ocr + LayoutLM3 today. Combining these two proved sticky - especially for runs in parallel. LayoutLM has natural proclivity to use easyocr tesseract type of OCR engine. Too many rough edges to patch over


Slightly non-technical take as recounted by my late landlady Ginny 50s~70s had strong hopes. Post war, there was only one way - UP! People worked more towards common good. Individualism was less. Things got done easier because the path of least resistance also was the fastest.

What’s the best thing written about why the remarkably vigorous and inventive France of the 70s and 80s (TGV, Minitel, Ariane, Rafale, Concorde, the world’s preeminent nuclear grid…) has not been nearly as visible in the 21st century? What went wrong?



There is no moat when you have open source players like @MiniMax__AI coming in hot 🔥 Congratulations on the nice release

We’re open-sourcing MiniMax M2 — Agent & Code Native, at 8% Claude Sonnet price, ~2x faster ⚡ Global FREE for a limited time via MiniMax Agent & API - Advanced Coding Capability: Engineered for end-to-end developer workflows. Strong capability on a wide-range of applications…

MiniMax__AI's tweet image. We’re open-sourcing MiniMax M2 — Agent & Code Native, at 8% Claude Sonnet price, ~2x faster
⚡ Global FREE for a limited time via MiniMax Agent & API
- Advanced Coding Capability: Engineered for end-to-end developer workflows. Strong capability on a wide-range of applications…


Sourav 已轉發

DeepSeek finally released a new model and paper. And because this DeepSeek-OCR release is a bit different from what everyone expected, and DeepSeek releases are generally a big deal, I wanted to do a brief explainer of what it is all about. In short, they explore how vision…

rasbt's tweet image. DeepSeek finally released a new model and paper. And because this DeepSeek-OCR release is a bit different from what everyone expected, and DeepSeek releases are generally a big deal, I wanted to do a brief explainer of what it is all about.

In short, they explore how vision…

I wonder what does it take for @SonyAlpha to spot amateur photographers outside of Instagram (ughh!). Do they even notice people on other channels? Genuinely curious

srvmshr's tweet image. I wonder what does it take for @SonyAlpha to spot amateur photographers outside of Instagram (ughh!). 

Do they even notice people on other channels? Genuinely curious

Weekend hikes be like

srvmshr's tweet image. Weekend hikes be like

Sourav 已轉發

This is an excellent history of LLMs, doesn't miss seminal papers I know. Reminds you we're standing on the shoulders of giants, and giants are still being born today. gregorygundersen.com/blog/2025/10/0…


Sourav 已轉發

Really interesting comparison between recent equilibrium flow matching and equilibrium matching papers arxiv.org/abs/2507.16521

I just read the Equilibrium Matching (EqM) paper; it’s excellent and insightful work! Interestingly, we recently published a related method called Adaptive Equilibrium Flow Matching (AEFM). Leaving out “adaptive” reveals strong conceptual parallels between the two approaches.



Tried shooting & digitizing on film camera (Fujifilm). The render is so beautiful & something DSLRs can only mimic by film recipes. OG film stock has a different appeal

srvmshr's tweet image. Tried shooting & digitizing on film camera (Fujifilm). The render is so beautiful & something DSLRs can only mimic by film recipes. OG film stock has a different appeal

Loading...

Something went wrong.


Something went wrong.