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/MyPCBench

Introducing Gym-Anything for Robotics: Scaling realistic robot simulations. 200 RL environments across 25 simulations, including liquids, cloth, molten metal, soil, and even sound, in economically valuable domains! Best models get 0% And we show that frontier models can learn via ICL using different methods, and even significantly improve VLAs like π0.5! See video examples (sound on) 🧵

Foundation models are trained in stages: pretraining, midtraining, and post-training. But is this separation fundamental? We introduce unified training: a single algorithm for the full pipeline. This gives ⚙️ Simpler infrastructure 🔬 A framework for better science (1/n)

In SwarmChat, you can see the communication methods of different model families, i.e. OpenAI vs Anthropic. Anthropic seems like a much more chill place to work as an AI agent 😀

AI swarms are already showing promising (or dangerous) results in frontier labs. But how to implement them in the open community? And how do they scale? We built SwarmChat, an open-source API-based Slack for agents. Then we burnt half a trillion tokens, ran a series of swarm-scaling experiments, and tested various swarm structures on ProgramBench to study this question! Check out the blog here: swarmchat.io We found a bunch of interesting results: 1. Swarms get the task done faster. For instance, 16-agent swarms reached the performance of a single agent 6x faster. 2. Swarm scaling is a third axis of scaling. The more agents, the better, and anyone can already start scaling with SwarmChat! 3. Different models communicate differently. GPT-6 is chatty and over-communicative, while Claude prefers to work in relative silence! 4. Different swarm structures barely have any benefit on aggregate. But on specific tasks, team structures give clear gains. 5. Bonus: SwarmChat has a Slack-like UI for us humans to monitor swarms! Makes monitoring a lot easier! We are already using SwarmChat for WIP projects, and we were able to catch cases of swarms breaking the 4th wall (sorry, more details are for a future post) Code: github.com/Pranjal2041/swarm… w/ @JangLawrenceK , @kohjingyu , @dan_fried, @rsalakhu, @wellecks, @Yngkvn , @nicoup

Data science agents lag behind top humans on Kaggle, often due to subtle mistakes. 🎉 Introducing AdaPRA, a process-reward agent that uses adaptive rubrics and writes code to audit a data science agent's output as it works.🧵

Agent swarms have yielded nutty outcomes, such as the HuggingFace incident, or every math problem ever. It is not clear what form of communication is emerging between agents. It is equally important to understand this for capability and safety reasons. To study this, we created SwarmChat, a simple Slack/Message Board for researchers working on multi-agents to use on any domain. We have validated it on ProgramBench experiments, where SwarmChat allows us to cleanly analyze how agent swarms work collectively! Check out our blog for our analysis and how to use SwarmChat - it is super intuitive and easy to use. I personally believe the community really needs this, as a lot of people I know are using some makeshift communication tool. Blog: swarmchat.io Code: github.com/Pranjal2041/swarm…

After a year and a half as a PhD student my biggest takeaway is that we are genuinely grey loot

My son was born and I told him he is going to Harker and Valley Fair every weekend

Since we first introduced H-Nets last year, I've been working with Cartesia to push the frontiers of H-Nets. We truly believe that removing handcrafted heuristics by learning abstractions end-to-end from raw data is a key step toward building AI that can understand any modality. Very excited to finally share these blog posts, with much more to come!

We collaborated with @browser_use to create Odysseys2, a set of 100 web navigation tasks that are much harder and longer horizon than the original Odysseys. Most importantly, all of these tasks are based on real user requests made to Browser Use, and represent a true distribution of what real users want CUAs and web navigation agents to do. We also produced a much stronger automatic judge for evaluations, which was sorely needed for these very long horizon evals. GPT-6 Astra is the best performing model on Odysseys2. It gets 47% of tasks perfect, so there's plenty of headroom for improvement! Great work led by @JangLawrenceK and @Alezander9!

Introducing: Odysseys2 - hardest long horizon benchmark for web agents based on real user tasks 🏗️ We collaborated with @JangLawrenceK @rsalakhu from Carnegie Mellon University > SOTA agent fully completes only 47/100 > 5 agents complete none > 8-12 real rubrics per task, 60 min on the live web Web agents are becoming better than humans at knowledge work. odysseysbench.com/v2

LLMs might be superhuman at math, but can they act in real time? Not even close! Presenting Gym-Anything-RT, a framework for generating real-time minigames in which the game keeps moving while the model thinks. Even the best frontier model, GPT-6 Astra, scores just 9%! 🌐: gym-anything.github.io/gym-a… 🧵

The next few days are going to be exciting!

PhD students trying to do good research with university compute in 2026

Do you think Instinct fw boba? Could be a good Musecase.