jegp

10 months agodiscuss

What's your stack? After reading this, self hosting suddenly appeals to me.

Great. If they can make some money out of this, why not?

I wonder what happens if they lock down some features as premium-only. The competition is tight in this space and there are tons of alternatives to Hyprland, like Sway, River, etc. Monetizing open-source code sounds like a dangerous path...

jegp

last yeardiscuss

Is the code for the "feeling lucky" selection mechanism open? Or, do you know how they select papers at random?

jegp

last yeardiscuss

That's a great point. I thought about something similar, but I also realized the arXiv numbers are growing like crazy, so I wonder how long it'll take for the (hardcoded) numbers to be deprecated. One could of course add some kind of cronjob to update the numbers, but that sounds like a lot of work...

jegp

last yeardiscuss

Thanks for this! If I knew this existed, I wouldn't have built the page myself.

jegp

last yeardiscuss

Conflicting motivations sounds like a very reasonable thing in large populations, in fact. But I have a hard time believing why workers wouldn't unanimously want better pay, better conditions, and more power. I would be curious to see any counter examples!

jegp

last yeardiscuss

Isn't the question then about the lesser evil? It's wrong to deny unions on the basis that some people are corrupt. Some people in companies are corrupt too and the US lies squarely in the middle of the corruption index https://en.wikipedia.org/wiki/Corruption_Perceptions_Index?w... I don't see why America!=Norway is a relevant argument.

jegp

last yeardiscuss

As an outsider, I'm astounded why workers aren't unionizing to a much higher degree. It's been proven to work [1] against the misinformation, discord, and wealth inequality that companies will, inevitably, cause. Despite the small union fee, the individual clearly stands to benefit[1]. Is it because people are cheap? Or not familiar with history? You'd think that tech workers were quite informed.

[1]: https://nordics.info/show/artikel/trade-unions-in-the-nordic...

jegp

2 years agodiscuss

Hey HackerNews community,

I wanted to share my excitement around our work on a "Neuromorphic Intermediate Representation" (NIR), published in Nature Communications: "NIR defines a set of computational and composable model primitives as hybrid systems combining continuous-time dynamics and discrete events. ... NIR decouples the development of neuromorphic hardware and software, enabling interoperability between platforms and improving accessibility to multiple neuromorphic technologies."

Why This Matters:

Neuromorphic computing, inspired by the architecture and processes of biological nervous systems, has long held promise for achieving more efficient, scalable, and intelligent systems. But so far, there hasn't been any unifying computational model, which has (1) scattered the scientific efforts and (2) hindered reproducibility.

Enter NIR, which acts as a kind of neural assembly language — a common ground where diverse architectures of neuromorphic systems can communicate seamlessly. NIR models are directly interpretable by neuromorphic platforms, much like the earliest compilers that bridged the gap between assembly languages and digital processors.

Key Highlights:

    Seamless Translation Between Computational Realms: The NIR creates a natural mapping between continuous-time neural dynamics and discrete computational models, allowing for unprecedented interoperability between neuromorphic and conventional hardware.

    Enhanced Efficiency and Scalability: By utilizing principles from both analog and digital realms, this approach optimizes data processing, reducing power consumption while maintaining speed and accuracy. This could pave the way for more efficient AI models that run on edge devices and within IoT ecosystems.

    Inspired by Biology, Refined for Technology: NIR embraces the adaptive, decentralized nature of biological neural networks, bringing us closer to hardware that can support general intelligence rather than narrowly focused algorithms.

Why Now is the Time to Pay Attention:

This isn't just another incremental step in AI or computing. NIR could be the catalyst that helps us unlock the next generation of intelligent systems — ones that don't just mimic intelligence but are built to understand, learn, and grow from the complexity of the world around them. This is a chance for developers, researchers, and innovators to join a movement that takes computing closer to the elegance of the human brain.

I'm curious to hear your opinions and feedback!

Disclaimer: I'm the first author.

jegp

2 years agodiscuss

It's still possible to train a network that's aware of the physics and then transfer that to physical devices. One approach to this from the neuromorphic community (that's been working on this for a long time) is called the Neuromorphic Intermediate Representation (NIR) and already lets you transfer models to several hardware platforms [1]. This is pretty cool because we can use the same model across systems, similar to a digital instruction set. Ofc, this doesn't fix the problem of sensitivity. But biology fixed that with plasticity, so we can probably learn to circumvent that.

[1]: https://github.com/neuromorphs/nir (disclaimer: I'm one of the authors)

Submissions

Show HN: Grab a Random ArXiv Paper

An assembly language for brain-inspired computations

comments