The best Hacker News stories from All from the past day
Latest posts:
Does our “need for speed” make our wi-fi suck?
Superpowers: How I'm using coding agents in October 2025
AMD and Sony's PS6 chipset aims to rethink the current graphics pipeline
The <output> Tag
The <output> Tag
Microsoft only lets you opt out of AI photo scanning 3x a year
Microsoft only lets you opt out of AI photo scanning 3x a year
Daniel Kahneman opted for assisted suicide in Switzerland
I built physical album cards with NFC tags to teach my son music discovery
I built physical album cards with NFC tags to teach my son music discovery
I switched from Htmx to Datastar
Show HN: I invented a new generative model and got accepted to ICLR
I invented Discrete Distribution Networks, a novel generative model with simple principles and unique properties, and the paper has been accepted to ICLR2025!<p>Modeling data distribution is challenging; DDN adopts a simple yet fundamentally different approach compared to mainstream generative models (Diffusion, GAN, VAE, autoregressive model):<p>1. The model generates multiple outputs simultaneously in a single forward pass, rather than just one output.
2. It uses these multiple outputs to approximate the target distribution of the training data.
3. These outputs together represent a discrete distribution. This is why we named it "Discrete Distribution Networks".<p>Every generative model has its unique properties, and DDN is no exception. Here, we highlight three characteristics of DDN:<p>- Zero-Shot Conditional Generation (ZSCG).
- One-dimensional discrete latent representation organized in a tree structure.
- Fully end-to-end differentiable.<p>Reviews from ICLR:<p>> I find the method novel and elegant. The novelty is very strong, and this should not be overlooked. This is a whole new method, very different from any of the existing generative models.
> This is a very good paper that can open a door to new directions in generative modeling.
Show HN: I invented a new generative model and got accepted to ICLR
I invented Discrete Distribution Networks, a novel generative model with simple principles and unique properties, and the paper has been accepted to ICLR2025!<p>Modeling data distribution is challenging; DDN adopts a simple yet fundamentally different approach compared to mainstream generative models (Diffusion, GAN, VAE, autoregressive model):<p>1. The model generates multiple outputs simultaneously in a single forward pass, rather than just one output.
2. It uses these multiple outputs to approximate the target distribution of the training data.
3. These outputs together represent a discrete distribution. This is why we named it "Discrete Distribution Networks".<p>Every generative model has its unique properties, and DDN is no exception. Here, we highlight three characteristics of DDN:<p>- Zero-Shot Conditional Generation (ZSCG).
- One-dimensional discrete latent representation organized in a tree structure.
- Fully end-to-end differentiable.<p>Reviews from ICLR:<p>> I find the method novel and elegant. The novelty is very strong, and this should not be overlooked. This is a whole new method, very different from any of the existing generative models.
> This is a very good paper that can open a door to new directions in generative modeling.
Liquid Glass Is Cracked, and Usability Suffers in iOS 26
Ryanair flight landed at Manchester airport with six minutes of fuel left
Nobel Peace Prize 2025: María Corina Machado
My first contribution to Linux
My first contribution to Linux
OpenAI, Nvidia fuel $1T AI market with web of circular deals
See also <a href="https://www.bloomberg.com/news/articles/2025-10-08/the-circular-openai-nvidia-and-amd-deals-raising-fears-of-a-new-tech-bubble" rel="nofollow">https://www.bloomberg.com/news/articles/2025-10-08/the-circu...</a> (<a href="https://archive.ph/E7nGC" rel="nofollow">https://archive.ph/E7nGC</a>)
Figure 03, our 3rd generation humanoid robot