A S M Iftekhar

@yoshitomo_cs ACs should just discard poor quality reviews.

I feel #ICLR2026 PCs Determining whether or not reviews in question were generated by some LMs is very tricky and it's hard to "prove" People often refer to the results of "GPTZero" or other tools, but it should be tough for PCs to use those as direct evidence...

and “Jailbreak Distillation: Renewable Safety Benchmarking” by @jackjingyuzhang, @ben_vandurme, @DanielKhashabi, @aagohary, @ASMIftekhar1, & more proposed a novel framework that “distills” jailbreak attacks into high-quality safety 🦺 benchmarks: aclanthology.org/2025.findings-… (7/7)

@soumithchintala Thanks a lot for pytorch!

Leaving Meta and PyTorch I'm stepping down from PyTorch and leaving Meta on November 17th. tl;dr: Didn't want to be doing PyTorch forever, seemed like the perfect time to transition right after I got back from a long leave and the project built itself around me. Eleven years at Meta. Nearly all my professional life. Making many friends for life. Almost eight years leading PyTorch, taking it from nothing to 90%+ adoption in AI. Walking away from this was one of the hardest things I've ever done. But I'm leaving with a full heart. PyTorch handles exascale training now. It powers foundation models that are redefining intelligence. It's in production at virtually every major AI company. It's taught in classrooms from MIT to rural India. The tools I dreamed about making accessible? They are. The barrier to entry I wanted to lower? It's almost gone. To be clear, there’s so much more to do. As long as AI evolves at a breakneck pace, PyTorch will continue to play catch up. Obsessing over the yet-to-come sometimes makes us forget how much we’ve already done. To everyone who built this with me—who believed research should be joyful, that tools should be elegant, that open source changes everything—thank you. This wasn't my journey. It was ours. What's next for me? Something small. Something new. Something I don't fully understand yet. Something uncomfortable. I could have moved to something else inside Meta. But I needed to know what's out there. I needed to do something small again. I couldn't live with the counterfactual regret of never trying something outside Meta. It's very hard to leave. I probably have one of the AI industry’s most leveraged seats, I lead the software layer that powers the entire AI industry. Every major AI company and hardware vendor are on a speed dial. This kind of power is really hard to give up. But curiosity ultimately won out in my head. Keep making AI delicious and accessible. I'll be watching. Probably filing issues. Definitely staying involved. Is PyTorch going to be okay? I don't want to be doing PyTorch forever. I don't want to be like Guido or Linus— bound to a single thing for decades. Last November, coinciding with the birth of my daughter, I started planning my exit with Aparna. My goal was to leave PyTorch in a good and stable place. By this August, during the second half of my parental leave, I knew: Edward, Suo, Alban, Greg, John, Joe and Jana were ready. The team faced hard people, product, technical and organizational problems and didn’t feel the need to lean back on me to solve these for them (unlike in the past). The product story they crafted for the PyTorch Conference was coherent—really coherent. The things I'd flagged red were turning healthy. The project didn't need me anymore. Unlike 2020-2022 (when I stepped down to go do robotics and came back when Lin, Dima and Dwarak left), I have strong confidence that this time PyTorch is truly resilient. The most aligned culture carriers of PyTorch – Greg, Alban, Ed, Jason and Joe are at the decision table now, and people with strong value alignment – Suo, John and Jana have joined them at the table. And there’s a long list of equally value-aligned people willing to sit at the table should any of these people leave. There are many little things that make up my confidence on the people – John worked on Julia and open-source for a very long time (in fact we hacked a Torch.jl in 2015), Suo has been the strongest systems builder and strategic partner I’ve had for the past two years, and Jana worked on resilient core systems for a very long time, I’ve had long technical and organizational discussions with her over the past few months that give me confidence. And the product lineup and execution in 2025 should be sufficient evidence for any remaining doubt. I’m confident that this band of PyTorchers are going to do exceptionally well. PyTorch might change in flavor because I no longer impose my own taste from the top, but I’m confident that the values are going to stay intact and the product is going to be awesome. My time at Meta The early years of FAIR were absolutely magical. I was part of a small family of absolutely brilliant people building state-of-the-art AI out in the open. From working on GANs with Emily Denton, Rob Fergus, Leon Bottou, Martin Arjovsky and the (now legendary) Alec Radford to building Starcraft bots with Gabriel Synnaeve, to building the first FAIR Cluster with Howard Mansell, to working on object detection with Adam Lerer and Piotr Dollar, to building PyTorch. It was more fun than I can describe in words. 2015 and 2016 were probably the most productive and professionally enjoyable years of my life. I’ll probably romanticize this period of my life forever. When I joined FAIR, I had massive impostor syndrome, and the first 3 months were very very difficult. I can’t credit Andrew Tulloch enough for being the most thoughtful, kind and welcoming mentor, without whom I wouldn’t have made it. I’m so damn bullish for Meta just from the fact that he’s back. --- My time on PyTorch was special. I loved every part of building it—designing it, managing it, being the PM, TL, comms lead, doc engineer, release engineer, squashing bugs, growth hacking, turning it into a coherent product with hundreds of people, transitioning it to industry stakeholdership – the whole nine yards. To the core PyTorch team at Meta: the engineers, researchers, open-source maintainers, docs writers, CI infrastructure folks, hardware partners, the community builders. To the hundreds more inside and outside Meta—thank you. You turned a library into a movement. There are too many people to credit and thank, but I can't not mention Adam Paszke, Sam Gross, Greg Chanan, Joe Spisak, Alban Desmaison, Edward Yang, Richard Zou, Tongzhou Wang, Francisco Massa, Luca Antiga, Andreas Köpf, Zach DeVito, Zeming Lin, Adam Lerer, Howard Mansell and Natalia Gimelshein. And Schrep. They made the launch happen. And so many more people became centrally important later: Lu Fang, Xiaodong Wang, Junjie Bai, Nikita Shulga, Horace He, Mark Saroufim, Jason Ansel, Dmytro Dzhulgakov, Yangqing Jia, Geeta Chauhan, Will Constable, Briah Hirsh, Jane Xu, Mario Lezcano, Piotr Balecki, Yinghai Lu, Less Wright, Andrew Tulloch, Bruce Lin, Woo Kim, Helen Suk, Chris Gottbrath, Peng Wu, Joe Isaacson, Eli Uriegas, Tristan Rice, Yanan Cao, Elias Ellison, Animesh Jain, Peter Noordhuis, Tianyu Liu, Yifu Wang, Lin Qiao and hundreds more. It’s criminal of me to not take the space to list out everyone else I should be mentioning here. PyTorch is nothing without its people ❤️. The most joyful moments of building PyTorch was meeting users eager to share their happiness, love and feedback. I remember a grad student coming to me at Neurips 2017, in a slurring emotional voice he said he’d been trying to make progress on his research for 3 years but within 3 months of using PyTorch he made so much progress that he was ready to graduate. That moment made it tangible that what we do matters, a lot, to a lot of people, even if you don't constantly hear from them. I do miss the intimacy of the PyTorch community, with a 300 person conference that felt like an extended family gathering, but I feel that’s a small price to pay considering the scale of impact PyTorch is truly having today – yes the Conference is now 3,000 people where market-moving deals get brokered, but it’s helping orders of magnitude more people to do their best AI work. I miss the intimacy, but I'm proud of that growth. --- To Mark Zuckerberg and Mike Schroepfer, who believed that open-sourcing is fundamentally important and is a sound business strategy. This is so hard to understand for most people within the course of business, but we’ve run lock-step on this strategy without ever having to discuss it. Without you two, neither FAIR nor PyTorch would’ve happened. And those mean so much to me. To Yann LeCun and Rob Fergus, for building the magical early FAIR that I so revere. To Aparna Ramani, a leader that I find so rare at Meta in her ability to hold a really high bar for the org, technically brilliant with the span to discuss deep infra systems and industry-strategy within the same conversation and for being an absolute execution-machine! I’ve learned so much from you. To Santosh, Kaushik, Delia, Oldham and Ben for being so welcoming to Infra. For someone coming over from FAIR with a wildly different culture, you all made me feel at home and made me part of the family, and thank you for that. To all my managers who've championed me through the PSC video game – Serkan, Howard, Jerome, Abhijit, Yoram, Joelle, Aparna and Damien – I owe you a lifetime of drinks. --- Signing off for now. —Soumith

@xwang_lk you should just send this.

Receiving a cold email asking me to cite their 03/2025 paper in our 09/2024 paper. I triple checked the dates to make sure I got it right.

Now some of the downsides of my experience with "vibe-coding for serious work". First, recap the good sides I mentioned previously: it's amazing at quick prototyping things, making tiny throw-away hyper-specialized demos, taking care of boilerplate, setup, and writing tests for me. Now the bad part: the code it generates is insanely verbose, overly defensive, bloated, and sometimes plain dumb. The models (I tried Claude code 4 and Codex Gpt5) have two big issues: 1) The model fully trusts you and takes what you say to the extreme. If you mention a requirement, it applies it to everything like a pedant, even if that forces quite insane contortion. A real good human coder would be like "ok wait, but this will make things extremely convoluted for XYZ, do you really mean this to apply here too?" and the answer is most likely "no, I didn't intend that" 2) The model never takes a step back and reconsiders/refactors things. It loves piling shit on top of more shit. A good human programmer would suddenly go "ok, that's a lot, let's simplify/unify things here for a bit". Even if you ask the model to do this, it usually sucks at simplifying. Two concrete real-life examples I had: 1) I had some pytorch distributed issue where some gathers in a library of mine would sometimes hang or die out of sync. Claude correctly identified that the process group was not always correctly initialized. So it started writing hundreds of lines of bookkeeping boilerplate to my library to try fixing this (and eventually did fix). After I looked at its fix, I immediately notice that the real fix was just moving my library's init call after torch distributed init, not before🤦‍♂️ So the real fix involved not a single new line of code, but Claude loves writing more lines! 2) In another library I made rapid iterations with Codex on the design. The core of the library boils down to a kind of graph where you need to walk through the nodes and do work on a node, while stopping on loops. Codex did correctly implement it, and it works; however, it wrote very convoluted code for the core logic, about 200 lines of code with two functions recursing into each other, and a few stacks and queues for traversal bookkeeping. After looking at it and taking a step back, I rewrote the whole thing from scratch in maybe 40 clear lines of code. It was great having Codex's extensive unit-tests to see that my rewrite is correct. So, in conclusion, the current state of vibe-coding is good for boilerplate, rapid iteration/prototyping, or one-off throwaway tools. For code that you intend to use, keep, extend, maintain for a while, you're always better off (re)writing it by hand. Maybe only after the LLM-assisted exploration and unit-test writing, though!

@suchenzang Literally me in half my code reviews lately. "Did you vibe code this?!" Is a meme over here now

Many thanks to my amazing collaborators @aagohary, Xiawei Wang, @ASMIftekhar1, Ahmed Magooda, @ben_vandurme, @DanielKhashabi, and Kyle Jackson! Work done at @Microsoft Responsible AI @MSFTResearch and @jhuclsp @JHUCompSci.

🚀 Introducing Hummingbird 🐦: High-Fidelity Image Generation via Multimodal Context Alignment! 🎨✨ We’re excited to announce Hummingbird’s acceptance to ICLR 2025! 🎉 Developed during my Microsoft internship, it explores new approaches to high-fidelity image generation. Hummingbird, a diffusion-based image generator, creates high-quality images using multimodal contexts including reference images and text. It ensures scene-aware generation, preserving scene attributes while maintaining diversity. 🌐 Project: roar-ai.github.io/hummingbird/ 📄 Paper: openreview.net/forum?id=6kPBT… Thanks to my amazing collaborators @mitts1910, @MengTianjian, @ASMIftekhar1, Vishwas, Barun, Dimitris, and Mei for their insights and dedication! 🙌 We welcome your feedback! 🏆 #ICLR2025

Openly injuring a civilian unprovoked.. #Bangladesh #StudentsUnderAttack #StepDownHasina

Bangladesh's Minister of State for Information and Broadcasting @MAarafat71, stated, "𝐓𝐡𝐞 𝐞𝐯𝐢𝐝𝐞𝐧𝐜𝐞 𝐢𝐬 𝐚𝐥𝐫𝐞𝐚𝐝𝐲 𝐭𝐡𝐞𝐫𝐞 𝐭𝐡𝐚𝐭 𝐦𝐚𝐧𝐲 𝐨𝐟 𝐭𝐡𝐞 𝐚𝐜𝐭𝐢𝐯𝐢𝐬𝐭𝐬 𝐡𝐚𝐯𝐞 𝐛𝐞𝐞𝐧 𝐝𝐫𝐮𝐠𝐠𝐞𝐝 𝐬𝐨 𝐭𝐡𝐚𝐭 𝐭𝐡𝐞𝐲 𝐜𝐚𝐧 𝐛𝐫𝐚𝐯𝐞𝐥𝐲 𝐜𝐨𝐧𝐟𝐫𝐨𝐧𝐭 𝐭𝐡𝐞 𝐩𝐨𝐥𝐢𝐜𝐞, 𝐨𝐩𝐞𝐧𝐢𝐧𝐠 𝐮𝐩 𝐭𝐡𝐞𝐢𝐫 𝐜𝐡𝐞𝐬𝐭𝐬." Does he even understand what he is saying? How disrespectful it is to speak so callously about the innocent young men, who lost their lives due to police brutality. It is fortunate that the government itself provided subtitles for his statement. I kindly request members of the international press to take note of this obnoxious and unacceptable comment by Arafat, as well as his attempt to malign opposition political parties. It is unbelievable that someone can stoop so low.

Bangladesh’s PM Hasina is using the tyrant’s handbook with armed troops, digital blackout, and “shoot-on-sight” orders to crush student protestors. Bangladeshi Americans are blocked from determining safety of loved ones. Hasina must face consequences for the “busloads” of the dead on her hands. amp.theguardian.com/world/article/…

Dozens of protestors have been killed by Bangladeshi authorities in recent days. My constituents cannot reach their loved ones due to a government implemented communications blackout. I call for an end to the blackout and de-escalation of violence against protestors.

Hey @elonmusk is it possible to provide starlink connections to Bangladesh? Bangladeshi Govt has shutdown all communications with the outside world after a day of violent protest.

Pedro, doesn't seem to understand there's a real emergency here. If we figure 250M users of LLMs, conservatively, and the shocking incidence rate tracked by Doomers last year of 121 "incidents" which includes such society shattering events as the Pope in a puffer jacket, then we have a 0.000048% incident rate, roughly below the rate of injury from water guns, but above the rate of injury from garden snails. Something must be done now before it's too late! It's gotten out of control! It's gotten so out of control that we should put the bills that prohibit time travel and transdimensional vampire invasions on hold to address this more pressing issue immediately. As for folks out there working to make a difference, we encourage everyone in college looking to volunteer or donate time and money to focus on magical, time traveling superintelligence from the future as your primary concern. Smaller concerns like poverty, inequality, and political repression can wait! We're also encouraging all college campuses and high schools to make AI Fear Mental Health support a top priority, and to ensure it's included with full coverage by today's greedy insurance companies. If you're having hallucinations about magical future AI, you shouldn't have to wonder who you can turn to in your time of need! There is hope!

@giffmana @savvyRL hmm, so just like the crystal clear definition of embeddings!

@savvyRL I think it's about as clearly defined as "a regularizing effect" or "implicit bias"

What the hell is "agent"? Is it just a software system where you replace lots of code with "let's ask an LLM"?

OpenAI released GPT-3.5 and then GPT-4 just a couple of months later (even though the latter had been in development for a while). This historical accident had the unintended effect of giving people a greatly exaggerated sense of the pace of LLM improvements, and led to a thousand overreactions ranging from influencer bros to x-risk panic (remember the "pause" letter?). It's taken more than a year for the discourse to cool down a bit and start to look more like a regular tech cycle.

📢🔥#HIRING Summer Research Interns 2024! Responsible and #OpenAI Research (ROAR) team at @Microsoft is hiring Research Interns in Multimodal AI/Video Understanding to publish at top-tier venues like CVPR/ICCV/ICLR/NeurIPS. Please apply here: jobs.careers.microsoft.com/global/en/job/…

In AI, the ratio of attention on hypothetical, future, forms of harm to actual, current, realized forms of harm seems out of whack. Many of the hypothetical forms of harm, like AI "taking over", are based on highly questionable hypotheses about what technology that does not currently exist might do. Every field should examine both future and current problems. But is there any other engineering discipline where this much attention is on hypothetical problems rather than actual problems?

v excited to finally announce our new work that formalizes one of the most effective practices for training LLMs—something that many industry leaders have conspicuously avoided discussing