Piotr Sokólski

CTO @ Third Dimension AI

[Piotr Sokólski @pyetras]

Excited for Dr. Christian Rupprecht to join @3rddimensionai as Chief Researcher.

Christian is a leading figure in the computer vision world, and I’m looking forward to working together to apply the latest 3D research towards unlocking more possibilities in spatial generation.

[Third Dimension AI @3rddimensionai]

Big news: Dr. Christian Rupprecht (@Oxford_VGG, @CVPR 2025 Best Paper) is joining Third Dimension as Chief Researcher!

Christian will lead our efforts to advance our world modelling and spatial generation technology into unparalleled simulations and worlds.

Here’s to turning research into reality 🔥

[Piotr Sokólski @pyetras]

7/ This is just the beginning. We're unlocking a new dimension (pun intended) of what's possible in robotics simulation and world modelling. If you’re interested in building the next big thing in robotics and entertainment, join our team: https://t.co/CXFP1A07dC

[Piotr Sokólski @pyetras]

1/7 Today we're launching SuperSim, a neural simulator that reconstructs the real world for better training & validation for robotics companies. The Third Dimension team has been working hard for many months, and I’m excited to finally share SuperSim with the world.

[Piotr Sokólski @pyetras]

6/ SuperSim is built to solve the problems I lived through, making it dramatically easier to create training environments that actually work. No more spending years building synthetic worlds that still fall short.

[Tolga Kart @tolgakart]

I was at @foxglove Actuate this week and it really felt like a who’s who of physical AI. I truly enjoyed meeting some of the people I had been looking up to, and definitely made new friends. Congratulations to @adrianmacneil and team for bringing together such an impressive group of people.

Some takeaways inspired by this week’s conversations:

Robots are having a renaissance
Compared to even a year ago, I see more new companies working on different aspects of robotics, from food delivery to factory operations, to construction. And then there are robots that want to do it all.

Neural Simulation vs. Rendering Engine Simulation
As a founder whose company is developing a neural simulator I’m admittedly biased, but I have high conviction that combining diffusion and reconstruction is the right approach for simulation. A key advantage: Our use of Gaussian Splatting solves the domain gap problem - ie. Third Dimension’s simulator looks like you're stepping into an actual video. Not only does your simulation look like a video, but all other actors in the scene behave exactly like they did in real life - this is just not possible with classical rendering engines.

On the other hand, video game-based simulators are good at providing ground truth, such as the automatic labelling of datasets, which is important for autonomy and robotics training. The downside lies in how expensive they are to build and that they just don’t look real.

Both…For Now
The realization for me is that there is room for both, for now. You want to get pixel perfect ground truth annotations? Tap into your classical sim engine. Want to make better use of your ever growing real world data? Neural simulation is your friend.

[Familiar @familiar_io]

Presenting MetaVoice-1B, a 1.2B parameter base model for TTS (text-to-speech).
* Emotional speech in English
* Voice cloning with fine-tuning
* Zero-shot cloning for American & British voices
* Support for long-form synthesis

https://t.co/35dTjbUvU2
https://t.co/mKpUYtd5qo

[Vaibhav (VB) Srivastav @reach_vb]

Let's go! MetaVoice 1B 🔉

> 1.2B parameter model.
> Trained on 100K hours of data.
> Supports zero-shot voice cloning.
> Short & long-form synthesis.
> Emotional speech.
> Best part: Apache 2.0 licensed. 🔥

Powered by a simple yet robust architecture:
> Encodec (Multi-Band Diffusion) and GPT + Encoder Transformer LM.
> DeepFilterNet to clear up MBD artefacts.

Synthesised: "Have you heard about this new TTS model called MetaVoice."

[Wayve @wayve_ai]

ICYMI: In March, Wayve presented how we are building a Neural Fleet to support autonomous driving.
In this thread, find out how we are leveraging NeRFs at city scale to accelerate the development of our AI Driver

[Wayve @wayve_ai]

How can Neural Radience Fields (NeRFs) be used to train and evaluate AVs? How can this be done at scale?

To find out how we are leveraging the latest method to build city-scale NeRFs at Wayve, tune into our session at #GTC2023 on March 22nd https://t.co/Wl4BS1lmfg

[Piotr Sokólski @pyetras]

Alfa Romeo in the MAXXI, Rome @LumaLabsAI

[Piotr Sokólski @pyetras]

#InstantNeRF Wilson St, London

[Piotr Sokólski @pyetras]

#nerf of The Hoxton Pony, Curtain Road, London. Definitely time for a GPU upgrade, the 1080ti does not cut it anymore

[Piotr Sokólski @pyetras]

@strongapp after the last iOS update I cannot view exercise details (history etc) - the app just freezes. Any workarounds?

[Piotr Sokólski @pyetras]

@CafeMuranoStJ Your voice mail says you're closed, so which is it? ;)

[Piotr Sokólski @pyetras]

Dreaming of mocking juice
The test was born an android
Real-time winter lag

[Piotr Sokólski @pyetras]

@tiarkrompf is there an IDE you'd recommend for developing for Scala-Virtualized? I find both Eclipse and IntelliJ lacking