iiinova — where your digital self keeps evolving

Reposted by Will Held

Open Athena @openathena.ai · 25/06/2026

In a new blog, Russell Power explains how the Marin team nearly doubled its sustained TPU usage by creating a custom global scheduler: Iris. Iris searches every region where Marin has compute, places each job wherever capacity appears, and moves data along as needed. 🔗 openathena.ai/blog/cluster...

1102

Will Held @williamheld.com · 11/05/2026

To train better open models, we need predictable scaling. Delphi is Marin’s first step: we pretrained many small models with one recipe, then extrapolated 300× to predict a 25B-param / 600B-token run with just 0.2% error. Getting there took some work 🧵

2389

Will Held @williamheld.com · 29/10/2025

Super interested to what degree this interaction can be fine-tuned into models in a non-reversible fashion! Voice cloning is unfortunately a capability which inherently shows up in pretrained audio models. It would be great to be able to largely limit the capability at the level of model weights!

110

Reposted by Will Held

Dan Jurafsky @jurafsky.bsky.social · 24/08/2025

Now that school is starting for lots of folks, it's time for a new release of Speech and Language Processing! Jim and I added all sorts of material for the August 2025 release! With slides to match! Check it out here: web.stanford.edu/~jurafsky/sl...

web.stanford.edu

Speech and Language Processing

315358

Will Held @williamheld.com · 11/08/2025

"GPT-5 shows scaling laws are coming to an end"

060

Reposted by Will Held

George Pearkes @peark.es · 06/08/2025

We’ve discovered a literal miracle with almost unlimited potential and it’s being scrapped for *no reason whatsoever*. This isn’t even nihilism, it’s outright worship of death and human suffering.

47103173294

Will Held @williamheld.com · 06/08/2025

Really great pointer from Hao Zhang on the other site in relation to GPT OSS use of attention sinks. If I were to guess, the attention sink is what allows them to omit QK-Norm which has become otherwise standard. www.evanmiller.org/attention-is...

evanmiller.org

Attention Is Off By One

Let’s fix these pesky Transformer outliers using Softmax One and QuietAttention.

010

Will Held @williamheld.com · 28/07/2025

The SALT Lab is at #ACL2025 with our genius leader @diyiyang.bsky.social. Come see work from @yanzhe.bsky.social, @dorazhao.bsky.social @oshaikh.bsky.social, @michaelryan207.bsky.social, and myself at any of the talks and posters below!

Alt Text:

Conference schedule for July 28th (Monday) and July 29th (Tuesday), listing talk titles, locations, times, and authors:

July 28th, Monday:

1. Attacking Vision-Language Computer Agents via Pop-ups
Location: Hall 4/5, Time: 11:00–12:30
Authors: Yanzhe Zhang, Tao Yu, Diyi Yang

2. SPHERE: An Evaluation Card for Human-AI Systems
Location: Hall 4/5, Time: 18:00–19:30
Authors: Dora Zhao*, Qianou Ma*, Xinran Zhao, Chenglei Si, Chenyang Yang, Ryan Louie, Ehud Reiter, Diyi Yang*, Tongshuang Wu*
(asterisk denotes equal contribution)

July 29th, Tuesday:

1. SynthesizeMe! Inducing Persona-Guided Prompts for Personalized Reward Models in LLMs
Location: Hall 4/5, Time: 10:30–12:00
Authors: Michael J Ryan, Omar Shaikh, Aditri Bhagirath, Daniel Frees, William Barr Held, Diyi Yang

2. Distilling an End-to-End Voice Assistant Without Instruction Training Data
Location: Room 1.61, Time: 14:12 (Second Talk)
Authors: William Barr Held, Yanzhe Zhang, Weiyan Shi, Minzhi Li, Michael J Ryan, Diyi Yang

3. Mind the Gap: Static and Interactive Evaluations of Large Audio Models
Location: Room 1.61 (implied), follows previous talk
Authors: Minzhi Li*, William Barr Held*, Michael J Ryan, Kunat Pipatanakul, Potsawee Manakul, Hao Zhu, Diyi Yang
(asterisk denotes equal contribution)

4. EgoNormia: Benchmarking Physical Social Norm Understanding
Location: Hall 4/5, Time: 16:00–17:30
Authors: MohammadHossein Rezaei*, Yicheng Fu*, Phil Cuvin*, Caleb Ziems, Yanzhe Zhang, Hao Zhu, Diyi Yang
(asterisk denotes equal contribution)

030

Will Held @williamheld.com · 28/07/2025

I'm in Vienna for #ACL2025! My work is all presented tomorrow, but today you'll find me today at the poster session from 11-12:30 evangelizing my labmate Yanzhe Zhang's work on his behalf. If you're interested in the risks traditional pop-up attacks present for AI agents, come chat!

140

Will Held @williamheld.com · 03/07/2025

A while ago I mentioned that for marin.community project, this gradient increase led to problematic loss ascent which we patched with Z-loss. I was curious, does AdamC just work? So over the weekend, I ran 4 experiments—130M to 1.4B params—all at ~compute-optimal token counts...🧵

marin.community

Marin

141

Will Held @williamheld.com · 03/07/2025

kyutai.org/next/unmute has built in turn-detection on the ASR and full I/O streaming for the TTS. Solves the latency issues that I think are 90% of why people use end-to-end speech models in the first place! From the details, you can @kyutai-labs.bsky.social is focused on real-world utility.

unmute.sh

Unmute by Kyutai

Make LLMs listen and speak.

010

Reposted by Will Held

Haley L. @haleyhaala.bsky.social · 21/06/2025

Flattered and shocked for our paper to receive the #facct2025 best paper award.

1103

Will Held @williamheld.com · 17/06/2025

I've only seen Veo 3 (or any other video generation model) used to produce viral videos. The fake videos seem to successfully trick the majority of commenters and have no visible watermark or disclosure of AI use.

110

Reposted by Will Held

Brendan Nyhan @brendannyhan.bsky.social · 12/06/2025

What would you say if you saw it in another country? A senator from a coequal branch of government dragged away by security from asking a question of a Cabinet official

26479142

Reposted by Will Held

Yijia Shao @echoshao8899.bsky.social · 12/06/2025

🚨 70 million US workers are about to face their biggest workplace transmission due to AI agents. But nobody’s asking them what they want. While AI R&D races to automate everything, we took a different approach: auditing what workers want vs. what AI can deliver across the US workforce.🧵

1227

Will Held @williamheld.com · 06/06/2025

Really cool to see theory connect to practice! We observed this phenomenon when trying to do deeper WSD cooldowns of our 8B model in the marin.community project! We Z-Lossed our way through the pain, but cool to see some stronger theory: marin.readthedocs.io/en/latest/re...

marin.community

Marin

0101

Will Held @williamheld.com · 02/06/2025

Based on current administration policies, China is about to have an influx of returning talent and a accelerated advantage in research investments. You need to be both sinophobic and irrational to expect the US to continue as the global scientific powerhouse with these policy own-goals.

https://www.nature.com/articles/d41586-020-00084-7

130

Reposted by Will Held

Kate Starbird @katestarbird.bsky.social · 25/05/2025

"“From time-to-time instances will arise in which the society, or segments of it, threaten the very mission of the university & its values... In such a crisis, it becomes the obligation of the university as an institution to oppose such measures & actively to defend its interests and its values.”

427872

Will Held @williamheld.com · 19/05/2025

How much faster would the science of large-scale AI advance if we could open-source the *process* of building a frontier model? Not just the final models/code/data, but also negative results, toy experiments, and even spontaneous discussions. That's what we're trying @ marin.community

194

Will Held @williamheld.com · 15/05/2025

It feels worth conference organizers running a study to see if this significantly impacts reviewer scores. I hope things like this are placebos, but if not we need to seriously consider whether existing peer-review processes for big ML conferences are providing value.

040

Will Held @williamheld.com · 07/05/2025

Introducing CAVA: The Comprehensive Assessment for Voice Assistants A new benchmark for evaluating the capabilities required for speech-in-speech-out voice assistants! - Latency - Instruction following - Function calling - Tone awareness - Turn taking - Audio Safety TalkArena.org/cava

talkarena.org

Comprehensive Assessment for Voice Assistants

CAVA is a new benchmark for assessing how well Large Audio Models support voice assistant capabilities.

101

Reposted by Will Held

Myra Cheng @myra.bsky.social · 02/05/2025

How does the public conceptualize AI? Rather than self-reported measures, we use metaphors to understand the nuance and complexity of people’s mental models. In our #FAccT2025 paper, we analyzed 12,000 metaphors collected over 12 months to track shifts in public perceptions.

34914

Reposted by Will Held

Naomi Saphra @nsaphra.bsky.social · 26/04/2025

I wrote something up for AI people who want to get into bluesky and either couldn't assemble an exciting feed or gave up doomscrolling when their Following feed switched to talking politics 24/7.

nsaphra.net

The AI Researcher's Guide to a Non-Boring Bluesky Feed | Naomi Saphra

How to migrate to bsky without a boring feed.

2335994

Reposted by Will Held

Jameel Jaffer @jameeljaffer.bsky.social · 21/04/2025

Worth noting that a number of universities have now sued over withheld and canceled grants, but no university has yet sued over the arrest, detention, and threatened deportation of its foreign students. www.nytimes.com/2025/04/19/o...

nytimes.com

Opinion | Our Foreign Students Are Terrified, and They’re Right to Be

The immigration crackdown has come to America’s campuses.

182002627

Reposted by Will Held

Ethan Zuckerman @ethanz.bsky.social · 20/04/2025

Mahmoud Khalil writes movingly about what his detention by ICE means for America: www.washingtonpost.com/opinions/202...

washingtonpost.com

Opinion | Mahmoud Khalil: What does my detention by ICE say about America?

A democracy for some is no democracy at all.

04416

Reposted by Will Held

Prasad Jallepalli, MD, PhD @prasad.bsky.social · 14/04/2025

aside: a stunning comment from David Baker, UW professor who won the Nobel Prize in 2024. Now 15 lab members are looking for positions overseas. “There’s so many amazing people who want to come in, & we can’t take them. The Nobel Prize was just a little blip. But things have gotten quite bleak.”

372253938

Reposted by Will Held

Sean Brodrick @seanbrodrick.bsky.social · 14/04/2025

Financial Times: "Since 1990, America has lost over 5 million manufacturing jobs. In that time, it has gained 11.8 million roles in professional and business services, and 3.3 million in transportation and logistical activities, linked to multinational supply chains." #EconSky

620682

Will Held @williamheld.com · 10/04/2025

The Model Context Protocol is cool because it gives external developers a way to add meaningful functionality on top of LLM platforms. To limit test this, I made a "Realtime Voice" MCP using free STT, VAD, and TTS systems. The result is a janky, but makes me me excited about the ecosystem to come!

141

Reposted by Will Held

ashley fairbanks @ziibiing.com · 05/04/2025

people being in the streets means something. never let your cynicism convince you otherwise.

89164332825

Reposted by Will Held

Jameel Jaffer @jameeljaffer.bsky.social · 31/03/2025

The Trump administration’s roundup of students who protested Israel’s bombardment of Gaza marks an astonishing, radical break with what one might justifiably think of as the central American idea. I wrote about it for @theguardian.com. www.theguardian.com/commentisfre...

7347141

Reposted by Will Held

Joshua Weitz @joshuasweitz.bsky.social · 28/03/2025

Working with an interdisciplinary team, we have developed a website to communicate how the White House's proposed cuts to health research would cause losses of $16B and 68,500 jobs. Find out how your community may be impacted. Explore more at SCIMaP: scienceimpacts.org a 🧵

US map via scienceimpacts.org visualization of economic loss due to IDC cuts to 15% as part of Feb 7, 2025 executive order, with shading denoting intensity of cuts.

19664613511

Reposted by Will Held

Michael Hobbes @michaelhobbes.bsky.social · 29/03/2025

We have had nearly two decades of panic about Free Speech on Campus and not a single case, not even the ones they made up, were as bad as what's happening now

168176376349

Reposted by Will Held

Michael Bernstein @mbernst.bsky.social · 18/03/2025

Step 1) Install the #chi2025 module to your Claude/ChatGPT: knollapp.com/add/ZlRKvCmB... Step 2) Ask the LLM, "Given my interests, what are some CHI 2025 papers I should check out?" (If the model doesn't already know your interests, you might need to state them.)

172

Reposted by Will Held

Jameel Jaffer @jameeljaffer.bsky.social · 09/03/2025

Arresting and threatening to deport students because of their participation in political protest is the kind of action one ordinarily associates with the world’s most repressive regimes. It’s genuinely shocking that this appears to be what’s going on right here. 1/

363058737

Reposted by Will Held

Hao Zhu 朱昊 @zhuhao.me · 04/03/2025

As always, we open source everything. Even our nicely made website: egonormia.org Please check out the leaderboard, the blog (w/Bibtex support), the code, data, as well as a data viewer.

egonormia.org

EgoNormia: A Benchmark for Visual Frontier Models' Normative Reasoning

A large scale video dataset and a benchmark for evaluating frontier models' understanding of physical social norms through videos.

131

Reposted by Will Held

Hao Zhu 朱昊 @zhuhao.me · 04/03/2025

We are getting closer to have agents operating in the real physical world. However, can we trust frontier models to make embodied decisions 🎮 aligned with human norms 👩‍⚖️ ? With EgoNormia, a 1.8k ego-centric video 🥽 QA benchmark, we show that this is surprisingly challenging!

1239

Reposted by Will Held

Valentin Hofmann @valentinhofmann.bsky.social · 31/01/2025

Great to see the International AI Safety Report highlight research on dialect prejudice, including our work on covert racism in LLMs! www.nature.com/articles/s41...

161

Reposted by Will Held

Kyle Mahowald @kmahowald.bsky.social · 29/01/2025

Now most *urgently*: we review the history of these models. A straight line can be traced to modern AI from basic science. Not in engineering but in the cognitive science of language. Much of it funded by NSF, whose funding has now been paused. www.goldengooseaward.org/01awardees/pdp

goldengooseaward.org

How We Think: Brain-Inspired Models of Human Cognition Contribute to the Foundations of Today’s Artificial Intelligence — The Golden Goose Award

AWARDEES : Geoffrey Hinton, James L. McClelland, David E. Rumelhart FEDERAL FUNDING AGENCIES: Department of Defense, National Institutes of Health, National Science Foundation Decades before ar...

074

Will Held @williamheld.com · 22/01/2025

Balancing data across domains is key to training the best generalist LLMs! In my summer work on the Meta Llama team, we introduce UtiliMax and MEDU, new methods to estimate data utility and optimize data mixes efficiently. HF Blog: huggingface.co/blog/WillHel... ArXiv: arxiv.org/abs/2501.11747

160

Reposted by Will Held

Yijia Shao @echoshao8899.bsky.social · 17/01/2025

LM agents today primarily aim to automate tasks. Can we turn them into collaborative teammates? 🤖➕👤 Introducing Collaborative Gym (Co-Gym), a framework for enabling & evaluating human-agent collaboration! I now get used to agents proactively seeking confirmations or my deep thinking.(🧵 with video)

12210

Reposted by Will Held

Dan Jurafsky @jurafsky.bsky.social · 12/01/2025

Happy New Year everyone! Jim and I just put up our January 2025 release of Speech and Language Processing! Check it out here: web.stanford.edu/~jurafsky/sl...

web.stanford.edu

Speech and Language Processing

115250

Reposted by Will Held

Nizar Habash @nyhabash.bsky.social · 09/01/2025

NYU Abu Dhabi Opening: Computer Science Professor in Artificial Intelligence, Machine Learning, T/TT - Open Rank! Deadline Feb 28, 2025! apply.interfolio.com/161449 #AI #ML #MachineLearning #NLP #NLProc #AIjobs #MLjobs #BigData #DataScience #AIcareers #DataJobs #JobSearch

apply.interfolio.com

Apply - Interfolio {{$ctrl.$state.data.pageTitle}} - Apply - Interfolio

074

Reposted by Will Held

Jameel Jaffer @jameeljaffer.bsky.social · 19/12/2024

Doctors Without Borders: What our medical teams have witnessed is "consistent with the descriptions provided by an increasing number of legal experts and organizations concluding that genocide is taking place in Gaza." www.doctorswithoutborders.org/latest/gaza-...

doctorswithoutborders.org

Gaza death trap: MSF report exposes Israel’s campaign of total destruction

Amid continued attacks, siege, and blockade, Israel is destroying conditions of life in Gaza.

25927

Will Held @williamheld.com · 17/12/2024

Update: Gemini 2.0 Flash now supported in Talk Arena! Come try the new Gemini and determine how strong it is at Speech & Audio compared to DiVA Llama 3, Qwen 2 Audio, and GPT 4o Advanced Voice at talkarena.org

Artificial Intelligence/Tech/Google
Google launched Gemini 2.0, its new AI model for practically everything

012

Reposted by Will Held

Michael Ryan @michaelryan207.bsky.social · 10/12/2024

Introducing Talk Arena🔊: the platform for interactively evaluating Large Audio Models! With the release of speech AI models like GPT4o, Gemini, Qwen-Audio, etc. which is the best? Cast your votes and help us decide🔥

132

Will Held @williamheld.com · 10/12/2024

With an increasing number of Large *Audio* Models 🔊, which one do users like the most? Introducing talkarena.org — an open platform where users speak to LAMs and receive text responses. Through open interaction, we focus on rankings based on user preferences rather than static benchmarks. 🧵 (1/5)

Talk Arena: Interactive Evaluation of Large Audio Models

3318

Will Held @williamheld.com · 08/12/2024

New Moving Sofa Problem proof dropped!! arxiv.org/abs/2411.19826

120