frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Lego's 0.002 mm Specification and Its Implications for Manufacturing (2025)

https://www.thewave.engineer/articles.html/productivity/legos-0002mm-specification-and-its-implic...
83•scrlk•1h ago•51 comments

Microsoft BitNet: 100B Param 1-Bit model for local CPUs

https://github.com/microsoft/BitNet
114•redm•2h ago•68 comments

The entities enabling scientific fraud at scale are large, resilient and growing

https://doi.org/10.1073/pnas.2420092122
35•peyton•1h ago•7 comments

Whistleblower: DOGE member took Social Security data to new job

https://www.washingtonpost.com/politics/2026/03/10/social-security-data-breach-doge-2/
138•raldi•48m ago•49 comments

PeppyOS: A simpler alternative to ROS 2 (now with containers support)

https://peppy.bot/
34•Ekami•3d ago•7 comments

Building a TB-303 from Scratch

https://loopmaster.xyz/tutorials/tb303-from-scratch
146•stagas•3d ago•51 comments

WA income tax clears House after 24-hour debate

https://www.seattletimes.com/seattle-news/politics/wa-income-tax-passes-house-after-24-hour-debate/
36•garbawarb•32m ago•36 comments

Zig – Type Resolution Redesign and Language Changes

https://ziglang.org/devlog/2026/#2026-03-10
330•Retro_Dev•13h ago•162 comments

AI Agent Hacks McKinsey

https://codewall.ai/blog/how-we-hacked-mckinseys-ai-platform
52•mycroft_4221•4h ago•17 comments

Cloudflare crawl endpoint

https://developers.cloudflare.com/changelog/post/2026-03-10-br-crawl-endpoint/
382•jeffpalmer•16h ago•148 comments

Create value for others and don’t worry about the returns

https://geohot.github.io//blog/jekyll/update/2026/03/11/running-69-agents.html
521•ppew•8h ago•366 comments

U+237C ⍼ Is Azimuth

https://ionathan.ch/2026/02/16/angzarr.html
358•cokernel_hacker•16h ago•61 comments

Yann LeCun raises $1B to build AI that understands the physical world

https://www.wired.com/story/yann-lecun-raises-dollar1-billion-to-build-ai-that-understands-the-ph...
534•helloplanets•1d ago•445 comments

Tony Hoare has died

https://blog.computationalcomplexity.org/2026/03/tony-hoare-1934-2026.html
1894•speckx•23h ago•248 comments

UK MPs give ministers powers to restrict Internet for under 18s

https://www.openrightsgroup.org/press-releases/mps-give-ministers-powers-to-restrict-entire-inter...
21•robtherobber•31m ago•8 comments

TADA: Fast, Reliable Speech Generation Through Text-Acoustic Synchronization

https://www.hume.ai/blog/opensource-tada
74•smusamashah•8h ago•19 comments

Julia Snail – An Emacs Development Environment for Julia Like Clojure's Cider

https://github.com/gcv/julia-snail
120•TheWiggles•3d ago•15 comments

Agents that run while I sleep

https://www.claudecodecamp.com/p/i-m-building-agents-that-run-while-i-sleep
376•aray07•19h ago•426 comments

RISC-V Is Sloooow

https://marcin.juszkiewicz.com.pl/2026/03/10/risc-v-is-sloooow/
280•todsacerdoti•18h ago•297 comments

SSH Secret Menu

https://twitter.com/rebane2001/status/2031037389347406054
277•piccirello•1d ago•126 comments

When the chain becomes the product: Seven years inside a token-funded venture

https://markmhendrickson.com/posts/when-the-chain-becomes-the-product/
32•mhendric•3d ago•14 comments

Writing my own text editor, and daily-driving it

https://blog.jsbarretto.com/post/text-editor
153•todsacerdoti•12h ago•71 comments

Debian decides not to decide on AI-generated contributions

https://lwn.net/SubscriberLink/1061544/125f911834966dd0/
353•jwilk•23h ago•268 comments

Standardizing source maps

https://bloomberg.github.io/js-blog/post/standardizing-source-maps/
62•Timothee•9h ago•5 comments

Levels of Agentic Engineering

https://www.bassimeledath.com/blog/levels-of-agentic-engineering
232•bombastic311•1d ago•109 comments

Launch HN: RunAnywhere (YC W26) – Faster AI Inference on Apple Silicon

https://github.com/RunanywhereAI/rcli
228•sanchitmonga22•21h ago•134 comments

Roblox is minting teen millionaires

https://www.bloomberg.com/news/articles/2026-03-06/roblox-s-teen-millionaires-are-disrupting-the-...
191•petethomas•3d ago•230 comments

Universal vaccine against respiratory infections and allergens

https://med.stanford.edu/news/all-news/2026/02/universal-vaccine.html
311•phony-account•16h ago•114 comments

Where did you think the training data was coming from?

https://idiallo.com/blog/where-did-the-training-data-come-from-meta-ai-rayban-glasses
12•speckx•1h ago•0 comments

JPMorgan marking down loan portfolios of private credit groups

https://www.ft.com/content/389a0003-d8de-4afd-9de9-be6e9fc6888c
4•petethomas•9m ago•0 comments
Open in hackernews

Microsoft BitNet: 100B Param 1-Bit model for local CPUs

https://github.com/microsoft/BitNet
112•redm•2h ago

Comments

QuadmasterXLII•1h ago
headline hundred billion parameter, none of the official models are over 10 billion parameters. Curious.
Tuna-Fish•1h ago
The project is an inference framework which should support 100B parameter model at 5-7tok/s on CPU. No one has quantized a 100B parameter model to 1 trit, but this existing is an incentive for someone to do so.
152334H•1h ago
but there is no trained 100b param model? "can run a 100B BitNet" is about the inference implementation, not about the existence of any such model
syntaxing•1h ago
Misleading title but this is pretty exciting. Interesting how this is based on llama cpp. Its nice to see some momentum since they released the paper in 2023
radarsat1•1h ago
I'm curious if 1-bit params can be compared to 4- or 8-bit params. I imagine that 100B is equivalent to something like a 30B model? I guess only evals can say. Still, being able to run a 30B model at good speed on a CPU would be amazing.
LuxBennu•1h ago
The title is misleading — there's no trained 100B model, just an inference framework that claims to handle one. But the engineering is worth paying attention to. I run quantized 70B models locally (M2 Max 96GB, llama.cpp + LiteLLM), and memory bandwidth is always the bottleneck. The 1.58-bit approach is interesting because ternary weights turn matmuls into additions — a fundamentally different compute profile on commodity CPUs. If 5-7 tok/s on a single CPU for 100B-class models is reproducible, that's a real milestone for on-device inference. Framework is ready. Now we need someone to actually train the model.
rustyhancock•1h ago
Yes. I had to read it over twice, it does strike me as odd that there wasn't a base model to work with.

But it seems the biggest model available is 10B? Somewhat unusual and does make me wonder just how challenging it will be to train any model in the 100B order of magnitude.

wongarsu•1h ago
Approximately as challenging as training a regular 100B model from scratch. Maybe a bit more challenging because there's less experience with it

The key insight of the BitNet paper was that using their custom BitLinear layer instead of normal Linear layers (as well as some more training and architecture changes) lead to much, much better results than quantizing an existing model down to 1.58 bits. So you end up making a full training run in bf16 precision using the specially adapted model architecture

wongarsu•1h ago
I've also always though that it's an interesting opportunity for custom hardware. Two bit addition is incredibly cheap in hardware, especially compared to anything involving floating point. You could make huge vector instructions on the cheap, then connect it to the fastest memory you can buy, and you have a capable inference chip.

You'd still need full GPUs for training, but for inference the hardware would be orders of magnitude simpler than what Nvidia is making

regularfry•1h ago
You only need GPUs if you assume the training is gradient descent. GAs or anything else that can handle nonlinearities would be fine, and possibly fast enough to be interesting.
embedding-shape•1h ago
> Framework is ready. Now we need someone to actually train the model.

If Microslop aren't gonna train the model themselves to prove their own thesis, why would others? They've had 2 years (I think?) to prove BitNet in at least some way, are you really saying they haven't tried so far?

Personally that makes it slightly worrisome to just take what they say at face value, why wouldn't they train and publish a model themselves if this actually led to worthwhile results?

gregman1•1h ago
Cannot agree more!
throwaw12•54m ago
Because this is Microsoft, experimenting and failing is not encouraged, taking less risky bets and getting promoted is. Also no customer asked them to have 1-bit model, hence PM didn't prioritize it.

But it doesn't mean, idea is worthless.

You could have said same about Transformers, Google released it, but didn't move forward, turns out it was a great idea.

embedding-shape•47m ago
> You could have said same about Transformers, Google released it, but didn't move forward,

I don't think you can, Google looked at the research results, and continued researching Transformers and related technologies, because they saw the value for it particularly in translations. It's part of the original paper, what direction to take, give it a read, it's relatively approachable for being a machine learning paper :)

Sure, it took OpenAI to make it into an "assistant" that answered questions, but it's not like Google was completely sleeping on the Transformer, they just had other research directions to go into first.

> But it doesn't mean, idea is worthless.

I agree, they aren't, hope that wasn't what my message read as :) But, ideas that don't actually pan out in reality are slightly less useful than ideas that do pan out once put to practice. Root commentator seems to try to say "This is a great idea, it's all ready, only missing piece is for someone to do the training and it'll pan out!" which I'm a bit skeptical about, since it's been two years since they introduced the idea.

wongarsu•33m ago
On the one hand, not publishing any new models for an architecture in almost a year seems like forever given how things are moving right now. On the other hand I don't think that's very conclusive on whether they've given up on it or have other higher priority research directions to go into first either
zozbot234•22m ago
What OpenAI did was train increasingly large transformer model instances. which was sensible because transformers allowed for a scaling up of training compared to earlier models. The resulting instances (GPT) showed good understanding of natural language syntax and generation of mostly sensible text (which was unprecedented at the time) so they made ChatGPT by adding new stages of supervised fine tuning and RLHF to their pretrained text-prediction models.
GorbachevyChase•45m ago
The most benign answer would be that they don’t want to further support an emerging competitor to OpenAI, which they have significant business ties to. I think the more likely answer which you hinted at is that the utility of the model falls apart as scale increases. They see the approach as a dead end so they are throwing the scraps out to the stray dogs.
riskable•25m ago
Not to mention Microsoft's investments in Nvidia and other GPU-adjacent/dependent companies!

A successful ternary model would basically erase all that value overnight. In fact, the entire stock market could crash!

Think about it: This is Microsoft we're talking about! They're a convicted monopolist that has a history of manipulating the market for IT goods and services. I wouldn't put it past them to refuse to invest in training a ternary model or going so far as to buy up ternary startups just to shut them down.

Want to make some easy money: Start a business training a ternary model and make an offer to Microsoft. I bet they'll buy you out for at least a few million even if you don't have a product yet!

cubefox•1h ago
LLM account
orbital-decay•1h ago
Funny enough I now involuntarily take RTFA as a slight slop signal, because all these accounts dutifully read the article before commenting, unlike most HNers who often respond to headlines.
yorwba•1h ago
Not all of them do: https://news.ycombinator.com/item?id=47335156 There are evidently lots of people experimenting with different botting setups. Some do better at blending in than others.
PeterHolzwarth•50m ago
Interesting - the account you mention, and the GP, are both doing replies that are themselves all about the same length, and also the same length between the two accounts. I get what you mean.
cubefox•1h ago
Yeah. It correctly pointed out that the editorialized HN title is wrong, there is no 100B model.
vova_hn2•57m ago
First they claimed that if you use em dashes you are not human

And I did not speak out

Because I was not using em dashes

Then they claimed that if you're crammar is to gud you r not hmuan

And I did not spek aut

Because mi gramar sukcs

Then they claimed that if you actually read the article that you are trying to discuss you are not human...

K0balt•38m ago
I’ve been rounded up for things I wrote two decades ago because of my em dashes lol. The pitchfork mentality gives me little hope for how things are going to go once we have hive mind AGI robots pervasive in society.
vova_hn2•27m ago
If I was operating a bot farm, at this point I would probably add some bots that go around and accuse legit human users (or just random users) of being bots.

Created confusion and frustration will make it much harder to separate signal from the noise for most people.

hrmtst93837•54m ago
I browsed through the history of the user and confirm this statement. I know that there are users who say they used em-dashes even before the rise of ChatGPT and HN statistics support that. For example, one prominent example is dang.

However this user uses — in almost all his posts and he had a speed of 1 comment per minute or so on multiple different topics.

Springtime•54m ago
Hmm, the user joined in 2019 but had no submissions or comments until just 40 minutes ago (at least judging by the lack of a second page?) and all the comments are on AI related submissions. Benefit of doubt is it'd have to be a very dedicated lurker or dormant account they remembered they had.

Edit: oh, just recalled dang restricted Show HNs the other day to only non-new users (possibly with some other thresholds). I wonder if word got out and some are filling accounts with activity.

verdverm•2m ago
There has been a shift to the Ai accounts, they use Show HN less now. This started before dang's comment, I assume because they saw the earlier posts about the increase in quantity / decrease in quality.

I suspect that they are trying to fake engagement prior to making their first "show" post as well.

nkohari•37m ago
I would love to understand the thought process behind this. I'm sure it's a fun experiment, to see if it's possible and so on... but what tangible benefit could there be to burning tokens to spam comments on every post?
butILoveLife•1h ago
>. I run quantized 70B models locally (M2 Max 96GB, llama.cpp + LiteLLM), and memory bandwidth is always the bottleneck.

I imagine you got 96gb because you thought you'd be running models locally? Did you not know the phrase Unified Memory is marketing speak?

WithinReason•1h ago
> a fundamentally different compute profile on commodity CPU

In what way? On modern processors, a Fused Multiply-Add (FMA) instruction generally has the exact same execution throughput as a basic addition instruction

actionfromafar•57m ago
Bitnet encoding more information dense per byte perhaps? CPUs have slow buses so would eke out more use of bandwidth?
ismailmaj•2m ago
if you upscale the precision of 1 bit to whatever FP precision the FMA uses then yeah, no diff.

But if you have an optimized implementation with SIMD-style loading and compute (i.e. packed representation), you drop the memory and compute requirements by an order of magnitude on CPU.

august11•1h ago
In their demo they're running 3B model.
webXL•23m ago
It comes from (intentionally?) misleading docs: https://github.com/microsoft/BitNet/issues/391

(only suggesting that it's intentional because it's been there so long)

verdverm•5m ago
That issue appears to be the one that's wrong. From the technical report

> We evaluated bitnet.cpp in terms of both inference speed and energy cost. Comprehensive tests were conducted on models with various parameter sizes, ranging from 125M to 100B. specific configurations for each model are detailed in the Appendix A.

cyanydeez•3m ago
Check out the new QWEN coder model.

Also, isnt there different affinities to 8bit vs 4bit for inferences

nickcw•1h ago
> bitnet.cpp is the official inference framework for 1-bit LLMs (e.g., BitNet b1.58). It offers a suite of optimized kernels, that support fast and lossless inference of 1.58-bit models on CPU and GPU (NPU support will coming next).

One bit or one trit? I am confused!

drsopp•1h ago
"1-bit LLMs" is just marketing. The Shannon entropy of one letter with a 3 symbol alphabet (-1, 0, 1) is 1.58.
Dwedit•1h ago
Log Base 2 of 3 = ~1.5849625, so that's the limit to how well you can pack three-state values into bits of data.

For something more practical, you can pack five three-state values within a byte because 3^5 = 243, which is smaller than 256. To unpack, you divide and modulo by 3 five separate times. This encodes data in bytes at 1.6 bits per symbol.

But the packing of 5 symbols into a byte was not done here. Instead, they packed 4 symbols into a byte to reduce computational complexity (no unpacking needed)

rasz•57m ago
>1-bit model

>packed 4 symbols into a byte

microslop, typical bunch of two-bit frauds!

cubefox•1h ago
Yeah, "1.58 bit" is 1 trit with three states, since log2(3)≈1.58.

So it's not a inference framework for 1-bit models (two states per parameter) but for 1.58 bit models (three states per parameter). Annoying that they try to mix up the two.

silon42•32m ago
I always hope for "just a bunch of if statements" ... this is not it.
himata4113•13m ago
it's if {} else if {} else {}
itsthecourier•1h ago
https://github-production-user-asset-6210df.s3.amazonaws.com...

demo shows a huge love for water, this AI knows its home

_fw•1h ago
Also, very influenced by the literature of Jenkins (2010).
giancarlostoro•1h ago
One of the things I often wonder is "what will be the minimally viable LLM" that can work from just enough information that if it googles the rest it can provide reasonable answers? I'm surprised something like Encyclopedia Britanica hasn't yet (afaik) tried to capitalize on AI by selling their data to LLMs and validating outputs for LLM companies, it would make a night and day difference in some areas I would think. Wikipedia is nice, but there's so much room for human error and bias there.
embedding-shape•1h ago
Your worry about Wikipedia is that there is "much room for human error and bias", yet earlier you seem to imply that a LLM that has access to the www somehow would have less human error and bias? Personally, I'd see it the other way around.
giancarlostoro•29m ago
When GPT 3.5 became a thing, it had crawled a very nuanced set of websites, this is what I mean. You basically curate where it sources data from.
utopiah•1h ago
> validating outputs for LLM companies

How? They can validate thousands if not millions of queries but nothing prevent the millions-th-and-one from being a hallucination. People who would then pay extra for a "Encyclopedia Britanica validated LLM" would then, rightfully so IMHO, complain that "it" suggested them to cook with a dangerous mushroom.

uniq7•1h ago
Since Google Search already includes an AI summary, your minimally viable "LLM" can be just an HTTP GET call
intrasight•1h ago
It's not so much a "minimally viable LLM" but rather an LLM that knows natural language well but knows nothing else. Like me - as an engineer who knows how to troubleshoot in general but doesn't know about a specific device like my furnace (recent example).

And I don't think that LLM could just Google or check Wikipedia.

But I do agree that this architecture makes a lot of sense. I assume it will become the norm to use such edge LLMs.

giancarlostoro•28m ago
Correct! I know RAG is a thing, but I wish we could have "DLCs" for LLMs like image generation has LoRa's which are cheaper to train for than retraining the entire model, and provide more output like what you want. I would love to pop in the CS "LoRa or DLC" and ask it about functional programming in Elixir, or whatever.

Maybe not crawl the web, but hit a service with pre-hosted, precurated content it can digest (and cache) that doesn't necessarily change often enough. You aren't using it for the latest news necessarily, but programming is mostly static knowledge a a good example.

thinkingtoilet•56m ago
Wikipedia has proven to be as accurate as encyclopedias for decades now. Also, I'm betting AI companies have illegally trained their models on the Encyclopedia Britanica's data by now.
bee_rider•42m ago
Isn’t that sort of what a RAG is? You’d need an LLM “smart” enough to turn natural-user prompts into searches, then some kind of search, then an LLM “smart” though to summarize the results.
giancarlostoro•30m ago
Yeah, I think RAG is the idea that will lead us there, though its a little complicated, because for some subjects, say Computer Science, you need a little more than just "This is Hello World in Go" you might need to understand not just Go syntax on the fly, but more CS nuances that are not covered in one single simple document. The idea being having a model that runs fully locally on a phone or laptop with minimal resources. On the other hand, I can also see smaller models talking to larger models that are cheaper to run in the cloud. I am wondering if this is the approach Apple might take with Siri, specifically in order to retain user privacy as much as possible.
Arcuru•1h ago
It's good to see this getting some continued development. I looked into it last year[1] and I thought it showed a lot of promise so I've been very disappointed that I never saw a newer model.

[1] - https://jackson.dev/post/dont-sleep-on-bitnet/

algoth1•1h ago
Headline: 100B. Falcon 3 family: 10B. An order of magnitude off
bee_rider•55m ago
What’s the lower limit on the number of bits per parameter? If you use CSR-style sparse matrices to store the weights can it be less than 1?
simonw•54m ago
Anyone know how hard it would be to create a 1-bit variant of one of the recent Qwen 3.5 models?
nikhizzle•48m ago
Almost trivial using open source tools, the question is how it performs without calibration/fine tuning.
wongarsu•39m ago
The results would probably be underwhelming. The bitnet paper doesn't give great baselines to compare to, but in their tests a 2B network trained for 1.58bits using their architecture was better than Llama 3 8B quantized to 1.58bits. Though that 2B network was about on par with a 1.5B qwen2.5.

If you have an existing network, making an int4 quant is the better tradeoff. 1.58b quants only become interesting when you train the model specifically for it

On the other hand maybe it works much better than expected because llama3 is just a terrible baseline

philvas•45m ago
steve jobs would have loved the microsoft repo with demo on mac
rarisma•43m ago
No 100b model.

My disappointment is immeasurable and my day is ruined.

devnotes77•37m ago
The compute throughput question (whether FMA equals ADD on modern CPUs) is accurate — that's not where the gain is. The real win is memory footprint.

A 100B ternary model packs to roughly 20-25GB (100B params at ~1.58 bits each). FP16 would be ~200GB, INT4 ~50GB. That difference is what moves the "doesn't fit" threshold. You go from needing HBM or multi-GPU NVLink to running on a workstation with 32GB DDR5.

DDR5 at ~100 GB/s is still much slower than HBM at ~3 TB/s, so memory bandwidth is still the inference bottleneck — but bandwidth is only a problem once the model actually fits. For 100B-class models, capacity was the harder constraint. That's what 1.58-bit actually solves.

WhitneyLand•34m ago
If they had a big result like, native 1.58-bit quality clearly matches top peers, they would be saying that prominently in the repo.

The engineering/optimization work is nice, but this is not what people have been waiting for, as much as, can’t the Bitnet idea that seemed promise really deliver in a competitive way.

StilesCrisis•3m ago
The output from this model is horrible! It's GPT-2 level babble and repeats entire paragraphs verbatim. It also reuses the same fake citation `(Jenkins, 2010)` over and over again. From the start of their video (which scrolls by fast enough that you don't see the slop clearly...)

``` Ecosystem Services and their impact on the Ecosystem

Ecosystem services refer to the services provided by ecosystems to the human society. These services include water, air, energy, nutrients, and soil (Jenkins, 2010). For instance, water is the most important service provided by an ecosystem and it helps in the conservation of water, irrigation and sanitation (Jenkins, 2010). On the other hand, air provides the oxygen needed for life.

The water cycle is a significant ecosystem service because it involves the cycling of water among the different parts of an ecosystem. It also involves the movement of water through the atmosphere, from one place to another. It is also the process of evaporation and condensation of water from the atmosphere. It also involves the movement of water from the air to the soil and water into the oceans.

The water cycle is a significant ecosystem service because it involves the cycling of water among the different parts of an ecosystem. It also involves the movement of water through the atmosphere, from one place to another. It is also the process of evaporation and condensation of water from the atmosphere. It also involves the movement of water from the air to the soil and water into the oceans. ```