Will Laufer (@wlaufer720) profile picture

Founder x2 (Trace, Prelude), dad x2

Brooklyn, NY
Joined February 2014
Photo by Will Laufer (@wlaufer720)Photo by Will Laufer (@wlaufer720)Photo by Will Laufer (@wlaufer720)Photo by Will Laufer (@wlaufer720)Photo by Will Laufer (@wlaufer720)Photo by Will Laufer (@wlaufer720)Photo by Will Laufer (@wlaufer720)Photo by Will Laufer (@wlaufer720)Photo by Will Laufer (@wlaufer720)Photo by Will Laufer (@wlaufer720)
Congrats @yaru_niu! Strong work
Introducing EgoX: a cross-embodiment learning framework that trains manipulation policies across diverse embodiments using data from humans, physical robots, and simulated robots. We cover highly heterogeneous embodiments: humans, physical and simulated humanoids, wheeled bimanual robots, tabletop arms, and quadrupeds. Their end effectors include three- to five-fingered dexterous hands, two-fingered and parallel-jaw grippers, and a fixed tool. One recipe, any embodiment. egox-dream.github.io/
1
2
196
Wild contribution! Congrats @NVIDIARobotics and the GROOT team – this is a watershed moment for robotics research.
A finer touch for humanoid robots. 🤖 At #CoRL2026, NVIDIA is introducing GR00T 2, an open world action model for humanoids, alongside new open industrial benchmarks from @RLWRLD_ai and @UCBerkeley built on Isaac Lab-Arena. Explore how NVIDIA’s open world action models, scalable simulation tools and agentic workflows help researchers train, test and refine robot skills, from connecting cables to assembling small components. Discover more ➡️ nvda.ws/4ixvdZV
4
325
Congrats @danfei_xu @simar_kareer @ryan_punamiya and the Egoverse team! Excited for Trace to contribute to the consortium and enable more progress in robot learning.
A few big updates from EgoVerse.ai: 1. EgoVerse is coming to RSS'26! @ryan_punamiya will present the work, and @simar_kareer will give a keynote at the Data-Centric Robotics workshop. 2. EgoVerse is expanding with more industry partners, including @microagi, @LightwheelAI, and Trace Labs, with a growing pool of training-validated, permissively licensed data (hitting 10k hours soon). 3. We are building the ecosystem around egocentric intelligence: Through a new partnership with the @nvidia Inception Program, we are connecting data vendors with researchers who can put this data to work.
1
7
1,496
Congrats on the result, Orin team!! Excited about what you're cooking up 🧑‍🍳
Introducing Horizon from @0rinlabs: the first long-horizon learning benchmark made from real agent logs - SOTA is 21% on the hardest section - 7-35M tokens of real agent history per task - Models are hardly getting better on the hardest tasks - Humans can score 100% (1/7)
1
11
1,617
fantastic talk by @danfei_xu , and so impressive to see the consistent progress out of @ICatGT human data is a research problem, but many want to treat it as an engineering problem. still, such an exciting time for robot learning!
Gave a talk on Robot Learning from Human Data at Stanford. It was great to be back! Some opinionated points: 1. Human data collection capacity is outpacing the research. 2. We still don't have the "science" for scaling robot capability with human data. 3. We are far from being able to model naturalistic human behaviors. youtube.com/watch?v=NUtaN10c…
1
6
470
Let's go!!! I've gotten to see @lalkaka 's early musings on the future of observability while at cloudflare turn into... the future of observability @usefiretiger ! Congrats to @__achille__ and the whole team!
Today we're introducing @usefiretiger. You and your AI agents write code. Firetiger makes sure it works. Our team and I have plenty of incident war stories building @Cloudflare, @segment, @Twitch. In the agentic coding era, the volume of code changes + quality issues in prod is ever increasing, but observability vendors aren't incentivized to close the gap. They make money when you write more data to them, not when your software actually works. Firetiger is the agentic operations layer for the agentic coding era. We combine production observability data, codebase understanding, and knowledge of your business to find problems before your customers do and fix them before they notice. We've raised $7.6 million led by @sequoia with participation from angels who believe in better software, including @eastdakota, @calvinfo, @NicoRosberg, @dok2001, @jeffawilke, and @alanaagoyal. You can sign up for @usefiretiger today, self serve. We charge for agents that directly make your software better and more reliable, not for observability data ingested, with plans starting at $599/month. Observability is dead. Long live outcome engineering.
1
6
401
Impressive!
Clarification Regarding Unitree's 2025 Sales Data
1
299
Storytime: Corduroy A tale of longing and belonging. Corduroy searches for the missing button to make himself worthy, and finds a friend who loves him unconditionally. A classic. Score: 8/10
1
112
Storytime: Caps For Sale A small, well illustrated story with a cute and memorable gimmick. Light fare, nevertheless a classic. Score: 7/10
1
114
Storytime: Pierre A delightful (cautionary!) tale of a boy who learns to care the hard way. Really hits when kids are going through a ‘no’ phase. Score: 8.5/10
91
Kids book review: The Bear and the Piano Fun coming-of-age story about curiosity, ambition, music, and needing to leave to appreciate what home is. Score: 7.5/10
1
91
Kids book review: Knight Owl’s Little Christmas Big fan of the original Knight Owl, but this installation adds nothing worthy. Like a Christmas cookie: sweet but not filling. Score: 3/10
1
87
Kids book review: Batman! Comprehensive overview of Batman’s abilities, technology, and tactics. All exposition, no plot. Score: 5/10
1
1
80
Fantastic work by @physical_int and @simar_kareer ! Unlike robot data, human data is abundant - exciting times ahead :)
We discovered an emergent property of VLAs like π0/π0.5/π0.6: as we scale up pre-training, the model learns to align human videos and robot data! This gives us a simple way to leverage human videos. Once π0.5 knows how to control robots, it can naturally learn from human video.
2
119
We discovered an emergent property of VLAs like π0/π0.5/π0.6: as we scale up pre-training, the model learns to align human videos and robot data! This gives us a simple way to leverage human videos. Once π0.5 knows how to control robots, it can naturally learn from human video.
79
335
160
2,671
1,190,058
Congrats @GeneralistAI team, inspiring work! The data infrastructure alone is wildly impressive
Introducing GEN-0, our latest 10B+ foundation model for robots ⏱️ built on Harmonic Reasoning, new architecture that can think & act seamlessly 📈 strong scaling laws: more pretraining & model size = better 🌍 unprecedented corpus of 270,000+ hrs of dexterous data Read more 👇
1
179
Will Laufer retweeted
I'm thrilled to share that @onkernel has raised $22M in Seed + Series A funding from @Accel with participation from @ycombinator @cintrifuse @vercel @svangel @RefineryVC @paulg @zeeg @charliermarsh & more 🚀 @rfgarcia and I started Kernel because LLMs can now automate any computer work — yet cloud infrastructure remains the biggest challenge holding developers back from adopting it at scale. We’re already powering workflows for co’s including @CashApp @rye + countless start-ups. And we’re just getting started 🧵
63
44
9
450
93,906
is there a vc focused on backing new sensor tech (hardware)?
80
why aren't @GLG, @AlphaSights etc. doing expert data for AI labs? they are literally networks of experts that have agreed to get paid for their expertise. seems like this the core asset of @mercor, which was founded in 2023 and is generating more revenue than any of them.
1
1
390