frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model

https://github.com/nex-agi/Nex-N2/issues/4
103•unrvl22•2h ago•57 comments

Show HN: Kage – Shadow any website to a single binary for offline viewing

https://github.com/tamnd/kage
19•tamnd•28m ago•4 comments

Firewood Splitting Simulator

https://screen.toys/firewood/
344•memalign•4d ago•118 comments

The Birth and Death of JavaScript (2014)

https://www.destroyallsoftware.com/talks/the-birth-and-death-of-javascript
158•subset•5h ago•87 comments

Ask HN: What are you working on? (June 2026)

39•david927•1h ago•82 comments

Caddy compatibility for zeroserve: 3x throughput and 70% lower latency

https://su3.io/posts/zeroserve-caddy-compat
83•losfair•4h ago•22 comments

Perlisisms

https://www.cs.yale.edu/homes/perlis-alan/quotes.html
44•tosh•2h ago•16 comments

Lisp's Influence on Ruby

https://blog.tacoda.dev/lisps-influence-on-ruby-6a54f1a7740e
155•tacoda•3d ago•21 comments

No, everyone is not using AI for everything

https://gabrielweinberg.com/p/people-are-consuming-ai-like-they
250•yegg•3h ago•252 comments

FarOutCompany

https://faroutcompany.com/
61•bookofjoe•3h ago•5 comments

Swiss voters reject proposal to cap population at ten million

https://www.swissinfo.ch/eng/swiss-politics/swiss-voters-reject-proposal-to-cap-population-at-ten...
95•FabCH•1h ago•59 comments

I indexed 669 GB of my GoPro videos using my M1 Max computer and local ML models

90•iliashad•2h ago•14 comments

The only scalable delete in Postgres is DROP TABLE

https://planetscale.com/blog/the-only-scalable-delete
65•hollylawly•3d ago•30 comments

Global density and biomass of arbuscular mycorrhizal fungal networks

https://www.science.org/doi/10.1126/science.adu4373
25•zdw•1d ago•1 comments

Rio de Janeiro's city government model Rio3.5 beats Qwen3.7 in recent benchmarks

https://twitter.com/zenmagnets/status/2065796012820848699
111•lucasfcosta•3h ago•28 comments

Formal Methods and the Future of Programming

https://blog.janestreet.com/formal-methods-at-jane-street-index/?from_theconsensus=1
69•eatonphil•5h ago•19 comments

Show HN: Dual YOLOv8n UAV Detection on RK3588S at 42 FPS Using NPU

https://github.com/alebal123bal/khadas_yolov8n_multithread
25•alebal123bal•3h ago•4 comments

How did Atari apply side art to Arcade Cabinets?

https://arcadeblogger.com/2026/06/14/how-did-atari-apply-side-art-to-arcade-cabinets/
51•msephton•4h ago•11 comments

How to Earn a Billion Dollars

https://paulgraham.com/earn.html
253•kingstoned•6h ago•695 comments

Quivers: A year of linear algebra by drawing arrows

https://lisyarus.github.io/blog/posts/quivers-a-year-of-linear-algebra-by-drawing-arrows.html
14•ibobev•4d ago•1 comments

Show HN: 3D print Z reinforcement via injected loops

https://mgunlogson.github.io/magma/
18•mgunlogson•5d ago•7 comments

Linux 7.1

https://lore.kernel.org/lkml/CAHk-=wi4BF4bMhZNZ1tqs+FFV4OuZRe3ZqdWB+LxRLmRweUzQw@mail.gmail.com/T/#u
68•berlianta•1h ago•3 comments

A 'cold blob' in the Atlantic could be a sign of AMOC shutdown

https://www.cnn.com/2026/06/12/climate/cold-blob-atlantic-amoc-ocean-circulation
110•tambourine_man•3h ago•138 comments

Free SQL→ER diagram tool, runs in the browser, nothing uploaded

https://sqltoerdiagram.com/
313•robhati•14h ago•60 comments

Honda Civics and the Evil Valet

https://juniperspring.org/posts/honda-evil-valet/
365•librick•17h ago•86 comments

FTX's former Anthropic stake would be worth about $75B at today's valuation

12•adam_rida•1h ago•10 comments

Dillo directory – Directory of useful sites that work reasonably well on Dillo

https://dir.dillo-browser.org/
13•HotGarbage•52m ago•0 comments

KPMG pulls report on AI usage due to apparent hallucinations

https://techcrunch.com/2026/06/13/kpmg-pulls-report-on-ai-usage-due-to-apparent-hallucinations/
78•Brajeshwar•3h ago•7 comments

Extinction-Level Capitalism

https://matthewbutterick.com/extinction-level-capitalism.html
95•laurex•3h ago•38 comments

Dangerous hormone-disrupting chemicals found in US breast milk samples

https://www.theguardian.com/us-news/2026/jun/14/breast-milk-research-chemicals
55•andsoitis•3h ago•20 comments
Open in hackernews

Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model

https://github.com/nex-agi/Nex-N2/issues/4
103•unrvl22•2h ago

Comments

unrvl22•2h ago
The municipality of Rio de Janeiro (via its IT company IplanRIO) released Rio-3.5-Open-397B, presented as a homegrown Qwen3.5 fine-tune that beats comparable open models on benchmarks. The linked issue argues it's actually a weighted merge of ~60% Nex-N2 Pro + ~40% Qwen3.5-397B-A17B - Nex-N2 having been released about a week earlier.
Lucasoato•39m ago
So the problem isn’t in the missing attribution to Qwen, but with the fact that they didn’t mention Nex-N2 Pro right?
Aurornis•19m ago
The problem is that they claimed to have made a big achievement with their home grown post training, and they expected to receive a lot of praise for it.

Then researchers looked at the weights and there is no post training at all.

They are now attributing both models they merged, but their excuse for the lack of post training is to claim they accidentally uploaded the wrong files.

DonsDiscountGas•19m ago
I didn't know model merging like that was possible. (Obviously possible from a pure software standpoint but I'm surprised it's effective)
AnotherGoodName•1h ago
This is fascinating that it worked though. Can we just merge all the open weight models and get something better?
_3u10•1h ago
No, they need the same arch, but you can distill them into a single model. And yes, if you use the API directly Claude will often say it’s an open weight model (likely the ones it was distilled from)
wds•1h ago
I imagine it'd work the same as merging all the good-tasting foods to get an even tastier one
dindunuf•55m ago
that kinda worked in llama 1/2 era, not between different models but between finetunes of the same model. the briefly legendary Mythomax was IIRC a merge of 5+ tunes, some of which were merges themselves.
avereveard•48m ago
most merge improve a small subset of "feeling" benchmark (too small, too specific, or out of distribution) and tend to show degradation on actual benchmark, with especially punishing result on long chain benchmarks.

also only work on matching architectures (i.e. finetunes/loras of the same model)

AlienRobot•1h ago
The model's webpage at https://huggingface.co/prefeitura-rio/Rio-3.5-Open-397B says it's a merge now. It previously didn't contain this paragraph:

>The model is built via a merge of https://huggingface.co/nex-agi/Nex-N2-Pro and https://huggingface.co/Qwen/Qwen3.5-397B-A17B, proceeded by On-Policy Distillation from a stronger model. We detected an incorrect upload in the previous version, where the base merged version was upload instead of the final distilled model. We are sorry for the confusion and apologize profusely.

Incidentally are people using Github issues as blogs now?

jonchurch_•8m ago
It wasnt framed as an issue which is the norm breakage I think you’re reacting to, as in they didnt ask that the readme be updated etc, but it is common now for folks to use a project’s issue tracker to name and shame them in a place they cant easily ignore.

Whether that’s right, prosocial, or professional is up for debate (as well as if any single definition of etiquette can be expected in 2026 on an issue tracker).

But surely you can see the optics reason why someone would take their complaint to the repo directly? It pressures the maintainers to respond, it allows for a pile on from the internet, and makes any decision to lock down a hostile thread into its own kind of statement.

The maintainers should absolutely post an official response and lock the thread though, its getting ugly in there.

zinodaur•1h ago
Oh no, someone is profiting off of their work without proper attribution!?!?
internet2000•1h ago
Attribution isn't the relevant part. Lying about your lab's capabilities is.
Planktonne•1h ago
That's also something all the AI companies have been doing.
dofm•46m ago
Lying about model capability is right now the lingua franca of the cloud AI business model, almost; they yes-and each other's lies because they are in a position of needing to generate interest, including going as far as needing to trigger regulatory capture.

(It's not news to anyone who has worked in sales-led businesses that salespeople are prone to believing the claims of other salespeople, I guess).

functionmouse•53m ago
leopards ate my face
adrian_b•53m ago
I do not see anyone lying.

The model card says:

> Post-trained from Qwen 3.5 397B

The model card also says that they use an inference framework based on "SwiReasoning: Switch-Thinking in Latent and Explicit for Pareto-Superior Reasoning LLMs" by Shi et al.:

https://arxiv.org/abs/2510.05069

So the sources seem properly attributed.

They only claim that what they did to "Qwen 3.5 397B" has improved the LLM, including, as expected, with "strong performance in Portuguese".

alfiedotwtf•1h ago
Wasn’t it already obvious given the awfully familiar parameter numbers?
ekjhgkejhgk•1h ago
One funny thing about incompetence is that they don't have the competence to know that their incompetence is straightforward to verify by a competent person.
root-parent•56m ago
You just described every single vibe coder...
carlosjobim•45m ago
Why would they care? They get their salaries and pensions and bonuses, and the tax payer is footing the bill.
thimabi•32m ago
I wouldn’t describe what happened here as incompetence. As a “carioca”, I am pleasantly surprised to know that the government’s IT department is involved in AI work — even without the budget to create its own models from scratch.
arcticfox•25m ago
This seems kind of insane though, every time I go to Rio I think of the potential of AI/technology to solve some problems and leave it even more paradisiacal... But working on their own model? Wtf? There are a million applications of existing ones there that should be followed up on instead.
MadrasTh0rn•1h ago
Not surprised
fkozlowski•58m ago
I'm honestly surprised that they even had the inclination to attempt creating a model. I guess it's bullish that a municipal IT department had the guts to try this?
yieldcrv•55m ago
Didn’t the last thread about this have someone from the lab or an enthusiast in Rio saying exactly that?

Its a fine tune of Qwen

Not a conspiracy

daemonologist•36m ago
The allegation here is that it's not actually a fine-tune of Qwen, but instead an undisclosed mashup (merge) of someone else's fine-tune of Qwen and the original model. Rio subsequently said that the model was in fact a merge, that they did additional fine-tuning after the merge, and that they accidentally uploaded the base merge instead of the version with additional fine-tuning. But this seems like quite an oversight...
jrm4•31m ago
“Well, Steve (Jobs), I think it’s more like we both had this rich neighbor named Xerox, and I broke into his house to steal the TV set, but I found out that you had already stolen it.”

-- Bill Gates

wunderlotus•14m ago
lmao i really hope this is a real quote cuz it’s a banger
ckcheng•3m ago
Apparently:

https://www.folklore.org/A_Rich_Neighbor_Named_Xerox.html

hintymad•2m ago
> Every weight tensor in Rio is, to thousands of standard deviations, the same 0.6/0.4 blend of Nex and Qwen — across all 60 layers and every component of the network. Other finetunes cannot be explained as interpolations.

I find it amazing how robust the current deep learning models are. A simple linear combination of every weight did not degrade the performance of the model, but enhanced it.

00index•25m ago
Are you talking about the credit that was just updated an hour ago? lol
petu•22m ago
That's attribution to Qwen team.

There (is/was) no attribution to Nex team (they've released a model based on Qwen 3.5 397B as well).

As per OP link Nex claims that what Rio team released (so far) is just linear interpolation of weights between Nex and OG Qwen model. With no attribution to Nex and zero signs of Rio doing any training of their own.

outside2344•23m ago
But the whole game is lying and stealing isn't it?
bachmeier•52m ago
"Their work"? First you had the original content creators that did 99.99% of the work. Then you had the US companies bundle it up into a frontier LLM. Then "they" did the "work" of using the US model as a foundation for their own. So in the sense of doing 0.00001% of the actual work that went into their product, sure.

I'd say it's more like someone forking a Linux distro, adding a few themes and fonts, and then complaining when someone else forks their distro and adds another theme.

dghlsakjg•47m ago
That’s the joke.
harikb•45m ago
It is only a problem if you claim it to be an independently developed OS with no attribution to base
bwilliams18•45m ago
That was the joke of the parent comment.
JoshStrobl•42m ago
That joke really went over your head, huh...
idiotsecant•38m ago
Oof this is delete your post level I think. Sorry bud, I been there.
woadwarrior01•48m ago
Are you new to the latest AI hype cycle? /s
carlosjobim•48m ago
This is a pure scam on tax payer money. But what else would be expected?
jrm4•33m ago
Unlike the big companies who do this, which often are merely impure scams on tax payer money a little more downstream.
carlosjobim•21m ago
Great, now we're defending embezzlement and fraud with public funds on HN, because we really really hate big business.

A child caught doing something bad will cry "but my friends also did it!", is that the level of reasoning hackers want to be at?

jrm4•17m ago
What part of that said "defense?"

They can both be bad.

blanched•15m ago
That seems like a bad faith read to me. Nobody is defending it, just pointing out the irony / hypocrisy. Two things can be bad, and they can be related.
Aurornis•21m ago
This is an open weights model based on other open weights models.

The dispute is that they released it with claims about having done some post training that improved the outputs. It was discovered that the model was not post trained like they claimed.

The HF page now says it’s a merge of models, which wasn’t there before. They’re trying to claim they accidentally uploaded the wrong model to HF and that they’ll upload the real one soon.

Basically, they thought they could splice two open weights models together and claim their team had accomplished some amazing post training, but they weren’t smart enough to realize that other researchers would discover that there wasn’t any post training.

moritzwarhier•16m ago
Thanks for the factual clarification. This is so important when everyone already has their trigger finger on politics. Not meaning that politics are irrelevant here, see sister comment by jobim.

But it's impossible to form a nuanced opinion when political association has a higher priority than the facts; which, again, don't look flattering for the implementers.

iknowstuff•5m ago
How do they just splice two models together?