Game Development, AI
New York
Joined August 2013
- Tweets581
- Following238
- Followers402
- Likes87
Hart Laboratories retweeted
hello there the jacobian conjecture is false thanx to my close friend akhil for asking about it and my other close friend fable for working during the world cup final
((1+xy)^3 z + y^2 (1+xy) (4+3xy), y + 3 x (1+xy)^2 z + 3 x y^2 (4+3xy), 2 x - 3 x^2 y - x^3 z): \C^3\to \C^3, has jacobian determinant -2, and sends (0, 0, -1/4), (1, -3/2, 13/2), and (-1, 3/2, 13/2) to (-1/4, 0, 0)
Hart Laboratories retweeted
Following up on the suggestion from Will Sawin, here is an illustration of the new configurations that disprove Erdos' unit distance conjecture (made with the help of ChatGPT 5.5 Thinking).
Hart Laboratories retweeted
Project Genie is a great way to experiment and get a feel for world models and this exciting research frontier.
I've been having fun exploring interactions that are usually quite hard to simulate with physics engines 😀
Hart Laboratories retweeted
Triple inverted pendulum in transition control
A classic control problem, done in real time.
This setup moves smoothly between all eight equilibrium points of a triple inverted pendulum.
The system reacts every 1 millisecond, which shows how fast modern control loops can be.
• Real-time control at 1 ms sampling
• Stable transitions between multiple balance points
• Built with Simulink and LW-RCP02 hardware
This is a beautiful example of how theory meets practice in advanced control engineering.
Video: youtube.com/watch?v=Rh7JuL3P…
Credit: Embedded Control Lab, Inha University
—-
Weekly robotics and AI insights.
Subscribe free: scalingdeep.tech
Hart Laboratories retweeted
Here's some footage from my Tesla's dashcam.
Just kidding. Did you believe me?
The truth is, this video is completely fake. It was generated by Tesla's world simulator, a generative AI model that not only allows Tesla to generate realistic synthetic training data, but also validate new versions of their self-driving model.
Hart Laboratories retweeted
Yes, pixels are in the direction of photons.
Hart Laboratories retweeted
I added this interesting radial magnetic bound state setup realized in 2021 to my ResearchGate page as researchgate.net/publication… where full description of the setup can be found.
Hart Laboratories retweeted
This is what it looks like in Genie 3.
It is 2025. You type the world you want. The rain comes. The lights burn. A train glides away, straight and cold. People stand close. They wait and do not speak. Above them fish turn slow and bright.
You type more. The city answers. It gives you color and noise and steam. It gives you the long street and the wet neon. It gives you the old hunger too. You feel it in the crowd. You feel it in the rail as it hums past.
You try again. You ask for quiet. The rain softens. The crowd breathes. The koi slow and dim. It is almost right.
It's Genie 3.
Hart Laboratories retweeted
New paper: World models + Program synthesis by @topwasu
1. World modeling on-the-fly by synthesizing programs w/ 4000+ lines of code
2. Learns new environments from minutes of experience
3. Positive score on Montezuma's Revenge
4. Compositional generalization to new environments
topwasu.github.io/poe-world
[1/n]
Hart Laboratories retweeted
We must not piss away our credibility. Credibility is all we have.
davidbau.com/archives/2025/0…
Hart Laboratories retweeted
Ken Thompson, upon receiving the Turing award, wrote a terrifying paper.
“Reflections on Trusting Trust” illustrates a scenario of original sin.
Because the C compiler is written in C itself, a compromised compiler can self-replicate with no trace in source code.
Hart Laboratories retweeted
Introducing Oasis: the first playable AI-generated game.
We partnered with @DecartAI to build a real-time, interactive world model that runs >10x faster on Sohu. We're open-sourcing the model architecture, weights, and research.
Here's how it works (and a demo you can play!):
Hart Laboratories retweeted
We hid a pure gold treasure worth $25,000 in a shrinking circle. Over the next three weeks, someone is going to find it. projectskydrop.com
Hart Laboratories retweeted
Diffusion Models Are Real-Time Game Engines
abs: arxiv.org/abs/2408.14837
project page: gamengen.github.io/
Google presents GameNGen, the first game engine powered entirely by a neural model that enables real-time interaction with a complex environment over long trajectories at high quality. GameNGen can interactively simulate the classic game DOOM at over 20 frames per second on a single TPU. Next frame prediction achieves a PSNR of 29.4, comparable to lossy JPEG compression. Human raters are only slightly better than random chance at distinguishing short clips of the game from clips of the simulation. GameNGen is trained in two phases: (1) an RL-agent learns to play the game and the training sessions are recorded, and (2) a diffusion model is trained to produce the next frame, conditioned on the sequence of past frames and actions.
Hart Laboratories retweeted
What's really going on in machine learning? Just finished a deep dive using (new) minimal models. Seems like ML is basically about fitting together lumps of computational irreducibility ... with important potential implications for science of ML, and future tech...
writings.stephenwolfram.com/…
Hart Laboratories retweeted
Have you ever done a dense grid search over neural network hyperparameters? Like a *really dense* grid search? It looks like this (!!). Blueish colors correspond to hyperparameters for which training converges, redish colors to hyperparameters for which training diverges.
Hart Laboratories retweeted
The countdown begins...three days to go until the Future of Go Summit! Stay up to date on our new #AlphaGo17 webpage goo.gl/mqauIM
Hart Laboratories retweeted
#AlphaGo is playing again in May, in a festival designed to explore the mysteries of Go! More on our blog: deepmind.com/blog/exploring-…
Hart Laboratories retweeted
A look at the Technology behind the 4D Game Miegakure (ft. 4D Crystals!)
youtube.com/watch?v=vZp0ETdD…