frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

France's homegrown open source online office suite

https://github.com/suitenumerique
130•nar001•1h ago•69 comments

Start all of your commands with a comma (2009)

https://rhodesmill.org/brandon/2009/commands-with-comma/
355•theblazehen•2d ago•122 comments

Hoot: Scheme on WebAssembly

https://www.spritely.institute/hoot/
53•AlexeyBrin•3h ago•11 comments

OpenCiv3: Open-source, cross-platform reimagining of Civilization III

https://openciv3.org/
738•klaussilveira•17h ago•232 comments

Google staff call for firm to cut ties with ICE

https://www.bbc.com/news/articles/cvgjg98vmzjo
13•tartoran•6m ago•0 comments

Reinforcement Learning from Human Feedback

https://arxiv.org/abs/2504.12501
30•onurkanbkrc•2h ago•2 comments

Coding agents have replaced every framework I used

https://blog.alaindichiappari.dev/p/software-engineering-is-back
92•alainrk•2h ago•86 comments

The Waymo World Model

https://waymo.com/blog/2026/02/the-waymo-world-model-a-new-frontier-for-autonomous-driving-simula...
994•xnx•23h ago•564 comments

Vocal Guide – belt sing without killing yourself

https://jesperordrup.github.io/vocal-guide/
125•jesperordrup•7h ago•55 comments

Unseen Footage of Atari Battlezone Arcade Cabinet Production

https://arcadeblogger.com/2026/02/02/unseen-footage-of-atari-battlezone-cabinet-production/
86•videotopia•4d ago•18 comments

Ga68, a GNU Algol 68 Compiler

https://fosdem.org/2026/schedule/event/PEXRTN-ga68-intro/
27•matt_d•3d ago•5 comments

Making geo joins faster with H3 indexes

https://floedb.ai/blog/how-we-made-geo-joins-400-faster-with-h3-indexes
144•matheusalmeida•2d ago•39 comments

Show HN: Look Ma, No Linux: Shell, App Installer, Vi, Cc on ESP32-S3 / BreezyBox

https://github.com/valdanylchuk/breezydemo
250•isitcontent•17h ago•27 comments

Monty: A minimal, secure Python interpreter written in Rust for use by AI

https://github.com/pydantic/monty
260•dmpetrov•18h ago•138 comments

Cross-Region MSK Replication: K2K vs. MirrorMaker2

https://medium.com/lensesio/cross-region-msk-replication-a-comprehensive-performance-comparison-o...
6•andmarios•4d ago•1 comments

Show HN: I spent 4 years building a UI design tool with only the features I use

https://vecti.com
351•vecti•19h ago•157 comments

Hackers (1995) Animated Experience

https://hackers-1995.vercel.app/
523•todsacerdoti•1d ago•253 comments

Sheldon Brown's Bicycle Technical Info

https://www.sheldonbrown.com/
402•ostacke•23h ago•104 comments

Show HN: Kappal – CLI to Run Docker Compose YML on Kubernetes for Local Dev

https://github.com/sandys/kappal
7•sandGorgon•2d ago•2 comments

Show HN: If you lose your memory, how to regain access to your computer?

https://eljojo.github.io/rememory/
320•eljojo•20h ago•196 comments

What Is Ruliology?

https://writings.stephenwolfram.com/2026/01/what-is-ruliology/
52•helloplanets•4d ago•52 comments

Microsoft open-sources LiteBox, a security-focused library OS

https://github.com/microsoft/litebox
365•aktau•1d ago•189 comments

An Update on Heroku

https://www.heroku.com/blog/an-update-on-heroku/
446•lstoll•1d ago•294 comments

Dark Alley Mathematics

https://blog.szczepan.org/blog/three-points/
99•quibono•4d ago•26 comments

How to effectively write quality code with AI

https://heidenstedt.org/posts/2026/how-to-effectively-write-quality-code-with-ai/
288•i5heu•20h ago•245 comments

Female Asian Elephant Calf Born at the Smithsonian National Zoo

https://www.si.edu/newsdesk/releases/female-asian-elephant-calf-born-smithsonians-national-zoo-an...
48•gmays•12h ago•22 comments

Was Benoit Mandelbrot a hedgehog or a fox?

https://arxiv.org/abs/2602.01122
26•bikenaga•3d ago•15 comments

I spent 5 years in DevOps – Solutions engineering gave me what I was missing

https://infisical.com/blog/devops-to-solutions-engineering
163•vmatsiiako•22h ago•74 comments

I now assume that all ads on Apple news are scams

https://kirkville.com/i-now-assume-that-all-ads-on-apple-news-are-scams/
1100•cdrnsf•1d ago•483 comments

PC Floppy Copy Protection: Vault Prolok

https://martypc.blogspot.com/2024/09/pc-floppy-copy-protection-vault-prolok.html
79•kmm•5d ago•13 comments
Open in hackernews

The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks (2018)

https://arxiv.org/abs/1803.03635
121•felineflock•1mo ago

Comments

laughingcurve•1mo ago
Article from 2018/19 and this hypothesis remains just that afaik with plenty of evidence going against it
yorwba•1mo ago
What evidence against it do you have in mind? I think it's a result of little practical relevance without a way to identify winning tickets that doesn't require buying lots of tickets until you hit the jackpot (i.e. training a large, dense model to completion) but that doesn't make the observation itself incorrect.
kingstnap•1mo ago
The observation itself is also partially incorrect. This is a video I watched a few months ago that went further into the whole how do you deal with subnetworks thing.

https://youtu.be/WW1ksk-O5c0?list=PLCq6a7gpFdPgldPSBWqd2THZh... (timestamped)

At the timestamp they discuss how actually the original ICLR results only worked on these extremely tiny models and larger ones didn't work. The adaptation you need to sort of fix it is to train densely first for a few epochs, only then you can start increasing sparsity.

paulsutter•1mo ago
Watched the video - thanks

Ioannu is saying the paper's idea for training a dense network doesn't work in non-toy networks (the paper's method for selecting promising weights early doesn't improve the network)

BUT the term "lottery ticket" refers to the true observation that a small subset of weights drive functionality (see all pruning papers). It's great terminology because they truly are coincidences based on random numbers.

All that's been disproven is that paper's specific method to create a dense network based on this observation

swyx•1mo ago
i intereviewed Jon (lead author on this paper) and yeah he pretty much disowns it now https://www.latent.space/p/mosaic-mpt-7b
gwern•1mo ago
Could you explain why you think that? I'm looking at the lottery ticket section and it seems like he doesn't disown it; the reason he gives, via Abhinav, for not pursuing it at his commercial job is just that that kind of sparsity is not hardware friendly (except with Cerebras). "It doesn't provide a speedup for normal commercial workloads on normal commercial GPUs and that's why I'm not following it up at my commercial job and don't want to talk about it" seems pretty far from "disowning the lottery ticket hypothesis [as wrong or false]".
oofbey•1mo ago
I think that was pretty clear even when this paper came out - even if you could find these sub networks they wouldn’t be faster on real hardware. Never thought much of this paper, but it sure did get a lot of people excited.
gwern•1mo ago
(Cerebras is real hardware.)
oofbey•1mo ago
It is real in that it exists. It is not real in the sense that almost nobody has access to them. Unless you work at one of the handful of organizations with their hardware, it’s not a practical reality.
aaronblohowiak•1mo ago
how long will that be the case?
IshKebab•1mo ago
At least for the foreseeable future (next 50 years say).
oofbey•1mo ago
They have a strange business model. Their chips are massive. So they necessarily only sell them to large customers. Also because of the way they’re built (entire wafer is a single chip) no two chips will be the same. Normally imperfections in the manufacturing result in some parts of the wafer being rejected and other binned as fast or slow chips. If you use the whole wafer you get what you get. So it’s necessarily a strange platform to work with - every device is slightly different.
sailingparrot•1mo ago
It was exciting because of what it means regarding how a model learns, regardless on whether or not its commercially applicable.
laughingcurve•1mo ago
i saw how it nerdsniped an extremely capable faculty member
swyx•1mo ago
he pretty much always says it offline haha but i maay have mixed it up with the subsequent convo we had at neurips https://www.latent.space/p/neurips-2023-startups
laughingcurve•1mo ago
cool beans, thanks for this -- I think it's easier to hear it directly from the authors. I was hesitant to start researchposting and come off like a dick.

also; note to self: If I publish and disown my papers, shawn will interview me :)

observationist•1mo ago
Neural networks are effectively gauge invariant, and you have a huge space of valid isomorphisms as far as possible "valid" layer orderings go, and if your network is overparameterized, the space of "good enough" approximations gets correspondingly larger. The good enough sets are a sort of fuzzy gauge quotient approximating some "ideal" function per layer or cluster or block (depending on your optimizer and architecture.)

https://arxiv.org/html/2506.13018v2 - Here's an interesting paper that can help inform how you might look at networks, especially in the context of lottery tickets, gauge quotients, permutations, and what gradient descent looks like in practice.

Kolmogorov Arnold Networks are better about exposing gauge symmetry and operating in that space, but aren't optimized for the hardware we have - mechinterp and other reasons might inspire new hardware, though. If you know what your layer function should look like, if it were ordered such that it resembled a smooth spline, you could initialize and freeze the weights of that layer, and force the rest of the network to learn within the context of your chosen ordering.

The number of "valid" configurations for a layer is large, especially if you have more neurons in the layer than you need, and the number of subsequent layer configurations is much larger than you'd think. The lottery ticket hypothesis is just circling that phenomenon without formalizing it - some surprisingly large percentage of possible configurations will approximate the function you want a network to learn. It doesn't necessarily gain you advantages in achieving the last 10% , and there could be counterproductive configurations that collapse before reaching an optimal configuration.

There are probably optimizer strategies that can exploit initializations of certain types, for different classes of activation functions, and achieve better performance for architectures - and all of those things are probably open to formalized methods based on existing number theory around gauge invariant systems and gauge quotients, with different layer configurations existing as points in gauge orbits in hyperdimensional spaces.

It'd be really cool if you could throw twice as many neurons as you need into a model, randomly initialize a bunch of times until you get a winning ticket, then distill the remainder down to your intended parameter count, and train from there as normal.

It's more complex with architectures like transformers, but you're not dealing with a combinatorial explosion with the LTH - more like a little combinatorial flash flood, and if you engineer around it, it can actually be exploited.

pizza•1mo ago
Yes to this. Furthermore:

- you can solve neural networks in analytic form with a hodge star approach* [0]

- if you use a picture to set your initial weights for your nn, you can see visually how close or far your choice of optimizer is actually moving the weights - eg non-dualized optimizers look like they barely change things whereas dualized Muon changes the weights much more to the point you cannot recognize the originals [1]

*unfortunately, this is exponential in memory

[0] M. Pilanci — From Complexity to Clarity: Analytical Expressions of Deep Neural Network Weights via Clifford's Geometric Algebra and Convexity https://arxiv.org/abs/2309.16512

[1] https://docs.modula.systems/examples/weight-erasure/

eru•1mo ago
Thanks for the explanations and the great links!
srean•1mo ago
Wouldn't such local invariance tie in with flatness or shallowness of the minima ?

This would tie in with the observation that flat/shallow minimas are easier to find with stochastic gradient descent and such weights generalise better.

choult•1mo ago
_Fewer_
eru•1mo ago
https://en.wikipedia.org/wiki/Fewer_versus_less

Compare also http://fine.me.uk/Emonds/wholetext.xml

tomhow•1mo ago
Indeed, the original title didn't make that mistake, so we've restored the original title as per the guidelines.
rob_c•1mo ago
This is basically just a rehash of "trained" DNN are a function which is strongly dependent on the initialization parameters. (Easily provable)

It would be awesome to have a way of finding them in advance but this is also just a case of avoid pure DNNs due to their strong reliance on initialization parameters.

Looking at transformers by comparison you see a much much weaker dependence of the model on the input initial parameters. Does this mean the model is better or worse at learning or just more stable?

snaking0776•1mo ago
This is an interesting insight I hadn’t thought much about before. Reminds me a bit of some of the mechanistic interpretability work that looked at branch specialization in CNNs and found that architectures which had built in branches tended to have those branches specialize in a way that was consistent across multiple training runs [1]. Maybe the multi-headed and branching nature of transformers adds and inductive bias that is useful for stable training over larger scales.

[1] https://distill.pub/2020/circuits/branch-specialization/

mceachen•1mo ago
@dang please retitle with (2018)
sbinnee•1mo ago
I was referring to this paper a lot when it was hyped, when people cared about architectural decisions of neural networks. It was also the year I started studying neural networks.

I think the idea still holds. Although the interest has been shifted towards test-time scaling and thinking, researcher still care about architectures like nemotron 3, recently published.

Can anyone give more updates on this direction of research, more recent papers?