Generalist
What changed, and what is the evidence.
Physical AIModelsDataDeploymentsHumanoidsControlHandsFundingActuationFactoryWorld ModelsEnd EffectorsHandsInferenceFleetRuntimeVision
🗞️ Pulse first · 3 news posts
Coverage
- 🗓️ 2026-10-22 · ✍️ RoboterminalEka's Robot Picks 1,500 Things an Hour. Its Gripper Is Rated for About 139 Days of That.
- 🗓️ 2026-10-21 · ✍️ RoboterminalSchaeffler Will Buy Thousands of Humanoids From Its Actuator Customers. Its €400 Million Order Book Doesn't Count Those Deals.
- 🗓️ 2026-10-19 · ✍️ RoboterminalFifteen Humanoids Are Loading Car Parts at 98%. The Number That Decides Their Future Is 15 Seconds.
Related topics
Timeline
What changed, and what is the evidence.
Oct 21, 2026 · Post · posted
| Headline |
|---|
Subtle adjustments and physical reasoning shown here are so cool!! https://t.co/LWOxfsck97 |
Skyfall's launch thread says SKY-001 beats every frontier model tested on CAD and BIM. Skyfall's own research post is more careful: it "matches or exceeds" them. That gap between the tweet and the paper is the whole story here. The real… |
Robotics is entering the API era. ★ Dream Labs | Dream-1 | From $3/hr ★ Google DeepMind | ER 2 | $1 in/$5 out ★ Generalist | Waitlists ★Physical Intelligence | Waitlists Robot manufacturers can focus on hardware instead of building their… |
What if a robot could learn a new manufacturing task in seconds instead of being reprogrammed? 🤖 🔗 https://t.co/KXPBS33K7O #PhysicalAI #ArtificialIntelligence #AI #Robotics #IndustrialRobotics #Manufacturing #Automation… |
Humanoids are starting to look more like delivery workers. In warehouses and on delivery routes, hands alone aren't enough. People brace bulky packages against their torso or slide them along an arm to keep them stable and free a hand for… |
JUST IN: As robots move out of cages and into shared spaces with people, how do we actually test whether they’re safe? That’s the problem @safeworldai is solving. The company just emerged from stealth with $12M led by @shinecapital and… |
A generalist model, not a dedicated robot policy, pulled off physical control by reasoning and using tools. Here's the four-step approach Gemini 4 figured out, given high-level instructions and the overall task: 1. Watched a human throw,… |
I am trying to visit them all and have only gotten a small percent. https://t.co/MCCp6V9nuE |
This is the capital of robotics! 🌁 Silicon Valley is home to so many physical AI companies that you could spend a month visiting them and still not see even 10% of them (trust me, I tried). So, the Bay Area is the world's leading… |
GeneralistのGEN-1.5、数秒の実演(お手本)から追加学習なしで新しい仕事に適応。 現場ごとの作業立ち上げ負担を減らす可能性。 https://t.co/jmKZHtoWEX #PhysicalAI #RobotLearning #foundation #model #ZeroShot #Generalist https://t.co/diw0EvQQU3 |
Generalist physical AI? 1 AI brain, 6 domains ... Odyssey-3 just demonstrated a single world model controlling: 🤖 Humanoids 🦾 Robot arms 🚗 Self-driving cars 🚁 Drones 🎮 Video games 🌎 AI training environments This is pretty huge if it… |
Interesting to think of this on terms of how flexibly one can create opposition. A good read https://t.co/FV3fzg0fMp |
98% success at each step sounds impressive. But across a 100-step task, independent errors can bring a robot’s chance of finishing the whole job down to just 13%. That’s why teaching robots to spot mistakes and recover matters so much.… |
Imagine what it could do with more sophisticated hands instead of just grippers https://t.co/it2HrFbn7X |
Incredible how much the ‘post video’ function sucks! Especially after @nikitabier told us over and over again how important it is to use it, instead is downloading the videos and post them again… Whether you use a link, it’s not working.… |
Crazy how far robot foundation models have come in terms of true fine motor skills: Without task-specific training, no sim-to-real transfer, and... fully autonomous ❗ The dual-arm robot places an O-ring onto a hydraulic rod and removes it… |
Taku is not another walking humanoid. It is a semi-humanoid physical agent built for a laundry-room shift: load, transfer, unload, fold, stack, and interrupt the current job when a machine finishes. Form follows the work. Human-shaped… |
Taku is not another walking humanoid. It is a **semi-humanoid physical agent** built for a laundry-room shift: load, transfer, unload, fold, stack, and interrupt the current job when a machine finishes. Form follows the work. Human-shaped… |
AI learned to talk from the internet. How does it learn to handle the real world? That’s the challenge for physical AI. In one test cited by Humanoid Analytics, training on more physical data raised a robot’s success rate in unfamiliar… |
Chinese robotics firm GigaAI is valued at $3bn Unlike hardware focused firms like Unitree, GigaAI focuses on intelligence A better analogy would be to consider it as the Chinese counterparties of US startups Generalist or Physical Int They… |
The secret to picking up a new task in ~30 mins sure seems to be large scale pretraining on your embodiment based on (this, generalist, sunday, skild) https://t.co/XZ1dE3MXbw |
Simulation only matters if it can predict what happens in the real world. That’s why @XSquareRobot X2Real benchmark is interesting. 44 hierarchical, long-horizon tasks across 10 capability dimensions, built to test how well generalist… |
🚨 Physical AI is having its "lab to production" moment, and the people driving it will be in one room later this month. If you're building or investing in embodied intelligence, Fellows Forum 2026 is where to be. On September 23–24, 500+… |
🚨 Physical AI is having its "lab to production" moment, and the people driving it will be in one room later this month. If you're building or investing in embodied intelligence, @Fellows_Forum 2026 is where to be. On September 23–24, 500+… |
Skild's founders just dunked on the entire demo-industrial complex. Skild just disclosed a $100M annual revenue run rate, only 10 months after their first commercial deployment. 60+ paying customers, including G10 Fulfillment, Mitsui & Co,… |
Robots learning by watching instead of being retrained for every new task could be a major shift. Skild’s S1 is showing the idea in practice, adapting to four different kitchen tasks from video demonstrations without fine-tuning. NVIDIA,… |
Just show it how, and this robot can learn these skills. In @deepakpathak’s clip, Skild’s S1 follows four different video prompts in the same kitchen, changing its behavior without fine-tuning. In just a few weeks, NVIDIA, Skild,… |
We’ve already gathered 500,000+ hours of real-world robot data at @GeneralistAI, but one lab can’t cover every task, tool, or embodiment. So we’re opening data bounties onchain via Robinhood. |
We’ve collected 500,000+ hours of real world robot data at @GeneralistAI, but no single lab can cover every task, tool and robot. We’re bringing data bounties onchain through Robinhood so the wider robotics community can contribute.… |
$CCXI Jeff Bezos just followed Chris Paxton. Paxton is one of the sharpest minds in embodied AI — currently at Agility Robotics (Digit), previously Meta and NVIDIA. He lives and breathes the hard problems: sim-to-real, faster-than-demo… |
Skip to main content Search arXiv Press Enter to search · Advanced search Computer Science > Machine Learning arXiv:2604.15483 (cs) \Submitted on 16 Apr 2026 ( [v1), last revised 24 Apr 2026 (this version, v2)\] Title:π0.7: a Steerable… |
From record humanoid funding to new data engines and factory-scale deployment plans, here’s what happened this week in Physical AI: 1️⃣ @XPENG_Global’s robotics unit raised $900M+ at a $6.3B valuation, the largest private embodied-AI round… |
Naively finetuning a generalist policy on a new task often just turns it into a specialist: it forgets all the tasks it was trained on before. 🥲 This happens regardless of whether the base policy is large or small (e.g., VLA or diffusion… |
In-Context Learning Results Hint at a “GPT Moment” for Robotics For general-purpose robots to be useful and economical, you need to be able to teach them new skills on the fly. And now we can see how this will work, with in-context long… |
Post Log in Sign up Post The Rundown AI on X: "Top stories in robotics today: \- Cyborg roaches become first responders \- SoftBank may add NEO to its robot stack \- Amazon builds robot-run delivery station \- Why Generalist is suddenly… |
Post Log in Sign up Post Rhys on X: "Cool to see a bunch of new robotics foundation model releases: π0.7 from @physical\int GEN-1.5 from @generalistAI S1 from @SkildAI Feels like the GPT-3 moment for robotics. Robot Models are Few-Shot… |
SubscribeSign in Discover more from a16z Jobs New jobs every day. Subscribe By subscribing, you agree Substack's Terms of Use, and acknowledge its Information Collection Notice and Privacy Policy. Already have an account? Sign in Open… |
scaling laws do not apply to physical AI yet. i'll say it again. scaling laws do not apply to physical AI. chelsea finn from physical intelligence showed this during her YC talk. they trained pi0.7 to make coffee. the robot couldn't learn… |
$14B valuation. Pre-revenue. A handful of customers. That is Skild AI today. And it is not the outlier. Physical Intelligence: $11B Generalist: $2B to $3B in months, $600M raised Skild AI: $14B Three companies. Billions in capital. Almost… |
- Home - Posts - Markets surge, track records fall, and in-context learning hits physical AI Markets surge, track records fall, and in-context learning hits physical AI Unitree’s market whiplash, an 8.86s 100m sprint, and how in-context… |
This is the mark of a general purpose robot foundation model -- that it can be taught tasks from examples without training. We saw an early version of this from rhoda and a solid one last week from generalist, and now a really impressive,… |
Skild AI's S1 takes a video demonstration as its prompt. No language instruction, no fine-tuning. Show it a task it has never seen, and it gets 66% of the steps right. Describe that same task in words instead, and you get 9%. Prompting a… |
Generalist AI just released GEN-1.5. It might be robotics' GPT-3 moment. What is one-shot learning via in-context prompting? In the case of language models like GPT-3, including a Q&A example in the prompt before the actual question… |
Post Log in Sign up Post Will Knight on X: "I'm usually wary about talk of "the ChatGPT moment for robots" because language is less complex than the physical world. After visiting @GeneralistAI, though, where I saw their robots perform… |
Post Log in Sign up Post Anirudha Majumdar on X: "Very cool work, Pete! I'm curious if you've tried an extreme version of this: no in-context example at all; just place objects in front of the robot and see if it can infer the task. I… |
Seeing a hype wave around GEN-1.5, and rightfully so. Lots of respect to Pete & Andy for executing so well. The secret is in the naturally repetitive motions in human-collected data. There're 2 main sources for such repetitions: (1)… |
Post Log in Sign up Post Generalist on X: "Or when fine-tuned to place a block into a bowl, it can clear obstacles (like a piece of paper covering the bowl) to complete the task, despite that not being in the demonstrations." - Generalist… |
Post Log in Sign up Post Generalist on X: "GEN-1.5, our latest embodied foundation model, can learn new tasks prompted with 3 - 12 seconds of a single demonstration, no gradient updates or fine-tuning. It generalizes prompts to new… |
Overview video Figure 1: Action flow as a shared control interface. Top: Hydra-0 learns from diverse interaction videos featuring egocentric human demonstrations, handheld UMI grippers, bimanual robot arms, and single-arm robots. Middle:… |
De-An Huang, Danfei Xu" /> Leveraging Compositionality for One-Shot Imitation Learning \| The Stanford AI Lab Blog Leveraging Compositionality for One-Shot Imitation Learning De-An Huang, Danfei Xu May 6, 2020 How do you teach a robot to… |