Daniel Fried @dan_fried

All 23 lectures for my CMU Advanced NLP course are now on YouTube. The slides and 20 code examples are also publicly available. - YouTube: youtube.com/playlist?list=… - Course Page (Slides, Schedule): cmu-l3.github.io/anlp-spring202… - Code: github.com/cmu-l3/anlp-sp… The lectures are grouped into 7 themes: fundamentals, architectures, learning & inference, modeling, evaluation, RL & agents, and scaling & efficiency. Check them out if you’re looking for an introduction or refresher on the fundamentals of LLMs, key ideas from recent NLP research, or are just curious to learn more.

Slides and videos will be posted online on the main site, so please follow along if you're interested: cmu-agents.com We're also looking for compute sponsors, so if you can help provide infra support (sandboxes, GPUs, etc) please reach out!

The instructors are me (CMU prof and developer of @OpenHandsDev) and @dan_fried (CMU prof and formerly on Meta's agent team). We have an all star team of TAs too! @yueqi_song @Jiarui_Liu_ @uilydna @apurvasgandhi @saujasv @sunweiwei12 @Aditya_Soni_8 @demisama_

We're creating a new course on AI Agents at CMU this Fall! We’re aiming to give students hands-on experience: from building agentic harnesses and evals to training with RL. Check out our course site for the full schedule: cmu-agents.com

Graham Neubig @gneubig

a day ago

This Fall at CMU we're teaching a new course on AI Agents! The goal is that you learn how to create a scaffold, build evals, and train an agentic LLM using RL. We'll try to balance theory and practice, and introduce modern frameworks and best practices.

38 183 2K 103K 2K

We're officially 3 days away from our #ACL2026 tutorial: "Future of Work in the Age of LLMs"! 🎉 📍 July 2, 9:00-12:30 PT, San Diego @Diyi_Yang kicks us off with an overview of the landscape of work agents: agent reality check, major NLP challenges, and the key research questions ahead @ZhiruoW dives into building agent harnesses and training backbone LMs, as well as future building directions @EchoShao8899 walks through designing datasets and metrics for evaluating AI agents at work 💡We close with a panel discussion on the societal & economic impact of AI, featuring economists and researchers from Stanford, MIT, and OpenAI: @erikbryn, @Alex_M_Richmond, Thomas Malone, and David Nguyen See you in San Diego! 🌊

Zora Wang @ZhiruoW

2 months ago

Excited to announce our tutorial: "Future of Work in the Age of LLMs" at #ACL2026 in San Diego, July 2! 🌴 There's a lot of speculation about AI and the future of human work. Our tutorial unpacks it from four angles: → The landscape of human work → How to build LLMs to augment

2 20 141 25K 60

I'll be at #ACL2026 presenting our work on convention formation! I'm thinking about collaborating and communicating with agents these days, and am excited to chat with others working on these problems! Find us at Poster Session E on Monday, July 6 to check out the work!

Saujas Vaduguru @saujasv

8 months ago

People adapt their language to communicate more efficiently over time. How can we make models do this? In our recent work, we trained models in self-play, and found that using the right incentives can make models adapt to communicate efficiently even without human demonstrations.

1 6 25 6K 9

🚨 Excited to share Pragmatic Reasoning via Self-Training, a method for LLM self-improvent on pragmatic reasoning. PragReST improves by +5.37% and +5.50% for Qwen3-8B/14B across pragmatics benchmarks with no human annotations or teacher models. LLMs still struggle w/ pragmatics: understanding what a speaker means, not just what they literally said. They often default to literal interpretations and miss implicature, intent, or context-dependent meaning. To close this gap, we started with a key question: Can we treat pragmatics as a LLM reasoning task? ➡️ Following a long line of work in pragmatics (e.g. RSA, IBR), PragReST treats pragmatic understanding as counterfactual reasoning. Instead of teaching models to ask “is this interpretation compatible with the words?”, we teach the model to reason about questions like “if the speaker meant something else, what would they have said instead?” ➡️ PragReST is self-improving: it self-generates pragmatic QA data, self-filters noisy examples, learns counterfactual reasoning traces via SFT, and further improves with GRPO using a self-judged correctness reward. ➡️ Error analysis shows that gains correlate with increased counterfactual reasoning. This suggests PragReST’s improvements are tied to reasoning over communicative alternatives, rather than simply more pragmatic data or more training. 🧵👇

As my time in Pittsburgh comes to an end... I'm excited to share that I will be joining UIUC as an Assistant Professor in fall 2027! I will recruit PhD students in both ECE and CS. I’m looking for students and postdocs who are excited to continue building towards collaborative AI systems that learn and adapt through interaction.

Computer-use evals like OSWorld still don’t really test personal assistant use cases: logged-in accounts, user data, personalized workflows, or realistic desktop/web environments. so we made MyPCBench: A Benchmark for Personally Intelligent Computer-Use Agents, with 184 tasks across 17 popular website clones, seeded with realistic user data and centered on Michael Scott’s hypothetical desktop. It’s easy to adopt if you already use OSWorld-style runners - I view it as an personalization-focused, more realistic upgrade for CUA evals. Website: mypcbench.com Paper: arxiv.org/abs/2606.16748 Code: github.com/ljang0/MyPCBen…

@lltjuatja Congrats Lindia, and UT!!

In case you’ve been wondering what I’ve been up to these days… So excited to (re)join an amazing community of linguists and NLP researchers at UT :)

UT Linguistics Dept @UT_Linguistics

2 weeks ago

We are excited to announce that Lindia Tjuatja (@lltjuatja) will be joining us as an Assistant Professor, starting in Fall 2027! Lindia is an alum of UT Linguistics and Electrical and Computer Engineering, and is currently finishing her PhD at CMU. Welcome back to UT, Lindia!

1 4 66 51K 4

New @fulcrum_inc research: Inverse Rubric Optimization (IRO), a testbed for agent science. Long-horizon tasks are often noisy, making them hard to study. In IRO, an agent learns a hidden judge's preferences under a label budget. We observe rich agent behavior and smooth scaling.

We are building AI technologies to empower humans, and this requires awareness of human reliance. Our latest work measures human cognitive offload using our workflow induction toolkit. Beyond showing the accuracy of our measure, we find that high reliance isn't inherently harmful. When users bring intentional engagement and genuine task understanding, AIs can facilitate human learning ✨

Vishakh Padmakumar @vishakh_pk

a month ago

People are increasingly worried that AI tools make us overreliant. But how do we actually measure this? We introduce Offloading Score, a measure of reliance based on the fraction of cognitive effort offloaded to AI while completing a task. In a controlled user study, Offloading

keyboard_arrow_left Previous keyboard_arrow_right Next

7 75 214 78K 100

New work on Multi-Agent Computer Use (MACU). The future of computer-use agents lies in multi-agent systems that combine planning, coordination, and parallel execution. Paper: arxiv.org/abs/2606.01533 Webside + Code: jykoh.com/macu MACU introduces a manager agent that decomposes tasks into a dynamic directed acyclic graph (DAG) of subtasks, dispatches parallel subagents, and continuously updates the plan as new information arrives. Across OSWorld, Online-Mind2Web, WebTrailBench, and Odysseys, we see performance improvement by 4.7–25.5%, achieving better test-time scaling, and solving long-horizon tasks that single-agent systems often fail to complete. On Odysseys, MACU reduces task completion time by 1.5×, showing that multi-agent coordination is a powerful path toward more capable and efficient computer-use agents. See a more detail thread by @kohjingyu.

Jing Yu Koh @kohjingyu

a month ago

Computer use agents are slow and brittle. The fix isn’t just stronger models, but also deploying them as multi-agent systems. MACU is a general Multi-Agent Computer Use framework that consistently lifts success rates by 3.4-25.5% and is up to 1.5x faster on long-horizon tasks.🧵

2 25 84 28K 35

Computer use agents are slow and brittle. The fix isn’t just stronger models, but also deploying them as multi-agent systems. MACU is a general Multi-Agent Computer Use framework that consistently lifts success rates by 3.4-25.5% and is up to 1.5x faster on long-horizon tasks.🧵

Work by @kohjingyu , with @rsalakhu Paper: arxiv.org/abs/2606.01533 Website: jykoh.com/multi-agent-co…

New work: a simple and general multi-agent computer use framework. It uses a manager to plan and re-plan by creating a task DAG, with subagents for parallel execution. It improves success rate across benchmarks, and substantially improves efficiency on long-horizon tasks.

Jing Yu Koh @kohjingyu

a month ago

Computer use agents are slow and brittle. The fix isn’t just stronger models, but also deploying them as multi-agent systems. MACU is a general Multi-Agent Computer Use framework that consistently lifts success rates by 3.4-25.5% and is up to 1.5x faster on long-horizon tasks.🧵

2 25 84 28K 35

LLMs refuse ambiguous queries that look harmful but aren't. Can they recover once users clarify, while staying safe? Our new interactive multi-turn benchmark measures both. 🚨 Turns out: not both at once.