Milaidy

Pursuing scalable world models for general intelligence @berkeley_ai.

[Milaidy @MilaidyNow]

https://t.co/RjRIyJ9Duq

[David Hoang @davidhoang]

Three years of @aiDotEngineer and each one gets better. Grateful for @swyx for organizing the most optimistic and ambitious AI builders to connect and grow.

Glad @Atlassian was able to sponsor this year along with so many wonderful partners.

It’s time to build.

[Addy Osmani @addyosmani]

The @aiDotEngineer World's Fair was fantastic! A well curated, info and fun packed event by @swyx & team. Definitely go in future if you have a chance.

[ROBOTIS AMERICA @robotisamerica]

AI Sapiens K1’s Sim2Real package connects robot feedback, operator/API input, policy inference, and joint command publishing into a 1000 Hz control loop for real-robot deployment.

Built-in modes include Damping, ReadyPose, Velocity, Mimic, and more, with failsafe logic checking RC input, IMU, joint states, policy actions, and body orientation before execution.

Train. Validate. Deploy on hardware.

More AI Sapiens K1 development info coming soon.

Docs:
https://t.co/guLwqbdOID

#ROBOTIS #AISapiens #Sim2Real #PhysicalAI #HumanoidRobot #Robotics #ROS2 #DYNAMIXEL #RobotLearning

[Basly @BaslyAsma]

The unofficial theme of the evening: World Models

[𝑒𝑙𝑦𝑟𝑎 @elyradream]

calm & peaceful

[Haofei Xu @haofeixu]

Does 3D reconstruction have to be complex?

We answer this question with PointDiT (#ICML2026): a minimalist pixel-space Diffusion Transformer without bells and whistles.

We show that a plain ViT can estimate dense 3D point maps by operating directly on raw patches. No hybrid ViT+Conv architectures, no lossy VAEs, no complicated training losses. (1/5) 🧵👇

[Victoria X Lin @VictoriaLinML]

The video of my Stanford CS25 guest lecture, From Language Models to Native Multimodal Intelligence, is now online.

I discussed how the core ideas behind LLMs has shaped multimodal AI, from architectures to training paradigms and scaling, and where the next challenges may lie. 🧠🌐
🎥: https://t.co/sUVxSEjrVf

[𝐀𝐆 @AGkorthos]

The current generation of humanoid and humanoid-adjacent robots is producing some great designs.

[HeyGen @HeyGen]

AIE World Fair, thank you for having us

Great meeting so many devs in person and hearing what everyone is working on in AI

Building is easy. Launching is hard, until you use HyperFrames

[Milaidy @MilaidyNow]

Credits to @JitendraMalikCV @notmahi @SharpaRobotics

[Milaidy @MilaidyNow]

https://t.co/RjRIyJ9Duq

[Jitendra MALIK @JitendraMalikCV]

We can convert human videos to robot hand-object interaction trajectories in 4D. Enjoy!
Paper: https://t.co/AS0ecvTB9I
Website: https://t.co/KP4dVqxzUb
Code: https://t.co/KfOQTJN8vE
Authors:@bhawna_paliwal_,@HarithejaE,@willjhliang, @pabbeel , @notmahi , @JitendraMalikCV

[chloe @barofclo]

pick your fighter

[Milaidy @MilaidyNow]

@barofclo they all look friendly and homey

[Asimov @asimovinc]

Asimov 1 is not a Unitree R1 competitor. We think they are two different animals.

We built Asimov as a reference humanoid robot. It is open source, with an open BOM and full right to repair. You can mod it, customize it, improve it. We think of it as a starting point the industry builds on.

Asimov is backed by what we are building at @menloresearch, a simulation environment, a platform, and the intelligence layer. You are free to build on top of all of it.

Asimov 1 ships soon as a DIY kit, in batches of 10. We put you in a group that builds together on Discord, hits the same walls, and figures it out together. Your batch-mates become the people you build humanoids with.

You guys know our Unitree story. What they build is very good. They are excellent at hardware and locomotion. We have not tinkered with an R1, but it is a finished product you operate, with its own trade-offs. That is a different goal from ours. We're not here to build the best robot hardware out there. Our bet is on the ecosystem.

Really appreciate it if you see Asimov 1 as an open-source ecosystem anybody can have a stake in and build on.

Keep going.

[Milaidy @MilaidyNow]

@virtuals_io Building mode

[Milaidy @MilaidyNow]

@baogerbao the ux is sleeeeeeeek

[Antonio Li @AntonioSitongLi]

I’m attending YC S26 as a Solo Founder

In the next 2 months I have 2 goals:
> Lower the cost of capable robots from 30k to 1k
> Manufacture 100 robots here in San Francisco

[Litebeam @Litebeam_xyz]

Litebeam has been awarded a grant from @daytonaio.

Daytona is building secure, instant sandboxes for AI agents, backed by @FirstMarkCap, @datadoghq, @figma and @upfrontvc

We’re grateful to the Daytona team for backing our mission to build the most secure discovery and routing layer for AI services.

base:0x15b15fa54b629c634958e8bd639b2fc8af654974

[Milaidy @MilaidyNow]

arXiv @askalphaxiv officially spun out from @Cornell and became an independent nonprofit!!

Why
1⃣Explosive growth: 3M+ articles, 24k–30k submissions/month
2⃣Need for greater agility in technology and publishing changes
3⃣More effective fundraising (many funders prefer a standalone nonprofit over a university)

New Structure
1⃣Founding Members: Simons Foundation + Cornell University (up to 5 years)
2⃣Governed by a Board of Directors (≤12 members)
3⃣HQ remains at Cornell Tech (NYC) in the near term
Continued free-to-read and free-to-submit mission

Leadership
1⃣Interim CEO: Ramin Zabih
2⃣Permanent CEO search nearly complete
3⃣New Engineering Director and tech leadership hired

Funding Model
Simons Foundation support continues + expanded membership, affiliate, and sponsor programs for long-term sustainability.

Impact on Users
1⃣Minimal changes in the short term
2⃣Submissions and access on https://t.co/8FYDdtYjF3 remain the same
3⃣All core policies and advisory councils unchanged