So robots can climb ladders now? Umm.. That doesn't scary at all!
Plenty of human work happens one ladder off the ground.
Rooftops, scaffolding, the top of the warehouse racks...
They are coming. They just figured out the way up.
This is LadderMan.
A Unitree G1 humanoid that climbs ordinary ladders on its own, using only what it sees.
Different angles, different rung spacing, ladders it never trained on.
Learned entirely in simulation, then run on the real robot with zero fine-tuning.
It starts from a single reference climb.
That becomes several experts, one per ladder shape.
Then all of them distill into one policy that runs off raw depth.
Thin rungs usually break depth sim-to-real.
So instead of brute-forcing it with randomization, they use a vision foundation model to clean up what the robot sees.
And climbing is only the warm-up.
Once it's up there, it doesn't just cling on.
It straightens a painting. Swaps a light bulb. Takes a box handed up to it.
All while balancing on the rungs.
Ladder climbing was a DARPA Challenge task, back when humanoids mostly faceplanted.
The robots aren't coming for the stairs. They're taking the ladder.
Congrats to the LadderMan team, wild work. @SihengZhao, @Yuanhang__Zhang, @520xyxyzq, @pabbeel
The SpaceX IPO is less than a week away, and 3 sectors are already flashing warning signs.
The SOX is at its most overbought RSI reading ever.
Financials are deteriorating under the surface.
Regional banks are sending signals that could matter before Friday’s debut.
Matt Maley goes live tomorrow at 1 PM ET to show how he’s positioned.
Reserve your spot: https://t.co/e33xHNoCv2
.@tylercowen on why AI creates more jobs than it destroys:
"One of the neatest properties of current AI models is they allow a small number of individuals working with AI to really do a lot more work than was possible previously."
"This will mean more companies, more projects, more nonprofits, just more ventures."
"One area is generally energy, electricity, the grid... It's completely screwed up. It will take twenty years, thirty years, forty years to fix... The AIs cannot do that on their own."
"The biomedical sector and medical trials, there will be many, many, many more ideas to test. AIs will help with the testing, but I don't think pure testing by simulation will be possible anytime soon."
"Simply care for the elderly. There will be robots, personal companions. We have this already. But the elderly also will want human care. It wouldn't surprise me if in the future, fifteen, twenty percent of all jobs were elderly care."
"Luis Garicano had an excellent online essay. He referred to what he called 'messy jobs': jobs where it's hard to explain exactly what the job is, but on a given day you're doing eleven different things, and it requires coordination and figuring out what you ought to do next and getting other people to help you... There's a real future in messy jobs."
Tyler Cowen with @dataWyatt
Generalist AI has raised a $400M funding round at a ~$2B valuation, bringing the total amount raised since its founding to over $500M .
Led by Radical Ventures, with new participation from 8VC, Union Square Ventures, Hanabi Capital, and Norwest. Existing backers NVIDIA's NVentures, Boldstart, Spark, Bezos Expeditions, and NFDG followed on. Notable angel additions: Fei-Fei Li, Naval Ravikant, and Xiaomi co-founder Bin Lin.
Founded by alumni of OpenAI, Google DeepMind, and Boston Dynamics, the team is building foundation models that span humanoids, robot arms, mobile warehouse robots, and autonomous space systems.
GEN-1 (April 2026) hit 99% reliability on diverse tasks and 3x faster execution than prior models. The new funding will help scale models, the data engine, compute, and commercial deployments..
GEN-1 model in action, fully autonomous:
Day 113, orbit 1753 — I love robotic operations! 🦾 In this timelapse, controllers on the ground are repositioning Dextre, our robotic handyman currently installed at the end of the Canadarm2. They used Dextre to unload equipment from the unpressurised Dragon trunk. Such a beautiful choreography to watch with Earth in the background!
Dextre is designed for delicate tasks and is attached to one end of Canadarm2 (the same robotic arm that we use to capture visiting cargo vehicles). Dextre is very often used to place scientific experiments outside of the ISS.
Go science! Go robotics!
Congratulations to the @csa_asc 🇨🇦 for enabling these robotic systems!
🎥 @esa / @NASA
#εpsilon • @esaspaceflight • @Space_Station • @NASAJohnson •
MotionDisco discovers long-horizon humanoid behaviors from scratch. Instead of relying on teleoperation or motion retargeting, it uses an LLM-guided evolutionary search to generate contact plans and a kinodynamic optimizer to verify whether those plans are physically feasible.
Since arriving at its destination five years ago, our Perseverance Mars rover has collected data that hints at a history of past life on the Red Planet.
Catch up on Percy’s biggest discoveries in this week’s episode of our Curious Universe podcast: https://t.co/J5dh8FhHjw
Elon Musk on Starship:
“Starship is really the key to making life multiplanetary and preserving the light of consciousness. That's what it's all about. And it may end up being the most important thing that we ever do.”
Video: SpaceX
Starship Booster 20 has made it to Masseys after rolling out from the production site overnight, for its cryo testing campaign, ahead of Flight 13.
Interestingly the roll took much longer then normal, as the SPMTs broke down a few times, forcing a extended stop, and the booster did a little loop of Massey's following its arrival.
@NASASpaceflight | https://t.co/RxPTBtuyLj
One of the most underrated things about Starship isn't just that it's bigger
It's that it's designed to completely rewrite the economics of spaceflight
SpaceX, Falcon's reusability already removed roughly 85% of the historic cost of launching to orbit
Starship aims to deliver another order-of-magnitude improvement on top of that
And it's not just about cost.....It's also about throughput
Falcon Heavy can deliver up to 64 metric tons to orbit
Starship Version 3 is targeting roughly 100 metric tons to orbit
Future versions could potentially push that to around 200 metric tons
The philosophy behind all of this comes from the Elon's "The Algorithm":
1. Question the requirements
2. Delete parts and process steps
3. Simplify
4. Accelerate
5. Automate
You can see this philosophy perfectly in the evolution of the Raptor engine
What began as a complex engine evolved into Raptor 3 dramatically simplified, more powerful, more reliable, and easier to manufacture
This is how SpaceX keeps pulling ahead
Not by adding complexity......By relentlessly removing it
The Moon's South Pole is one of the coldest, harshest places in our solar system. Lunar Outpost's Pegasus LTV is built to go there.
Built to operate through 500ºF temperature swings and tackle 20º slopes, Lunar Outpost's rovers will prospect and eventually extract resources at scale in these regions, including Water Ice and critical volatiles. Our robotic workforce unlocks the resources that will be crucial to enabling a truly sustainable human presence on the Moon.
#LunarOutpost #DrivingArtemis #TheNextLeap #SpaceTech #Innovation
🚨 $SPCX SpaceX kostet beim IPO $135 pro Aktie.
Es werden $74,4 Milliarden eingesammelt bei einer Bewertung von $1,75 Billionen.
Der größte Börsengang in der Geschichte der Menschheit steht bevor.
Bist du beim IPO dabei oder wartest du ab?
Today I'm launching a new project called SynthTraces 🔥
It is a minimal codebase to generate synthetic coding agent session traces using Pi (from @badlogicgames)
I wanted a large number of coding-agent traces, so I built a tiny harness where two models talk to each other:
- an open model (served via HF Inference Providers) plays the coding agent. It gets read + bash access to a real open source codebase (the huggingface OSS projects)
- a small local model (llama.cpp) plays the human user, asking simple questions like "how do I run this?" or "how is CI set up?"
The result is more than 2,000 Pi session traces which can be used to train or fine-tune LLMs, and optimize them for Pi 🤯
And ofc everything is published on @huggingface ✅
Nuclear innovation isn't just about energy.
From next-generation reactors to advanced technologies for fuel storage systems, researchers at the @CockrellSchool are rethinking nuclear safety technology from the ground up: https://t.co/3hdKdqc1UY
NEWS: @SpaceX received FAA clearance for Starfall, a circular uncrewed reentry capsule to return up to 1,000 kg payloads from orbit via parachute splashdown
It enables safe recovery of microgravity-manufactured materials, completing the logistics loop for orbital industry
Microsoft unveiled Majorana 2, a next-generation topological quantum chip developed with help from Microsoft Discovery’s agentic AI.
The announcement connects Microsoft’s quantum computing roadmap with its broader push into AI-assisted scientific discovery.
Most general-purpose robots are stuck in the lab, stationary, or automating dummy tasks that provide little to no value.
Yondu is the first company to deploy humanoids for dexterous, mobile tasks like bin-picking for 3PLs. This is normal for us. We are operating every single day, picking hundreds of orders without making mistakes. We integrate directly with ShipHero.
Of all the applications in warehousing, this is one of the most labor-intensive and costly. We are the only company in the world that can automate this task with general-purpose robots.