[petcat]

This is not "open source" AI.

Photoshop source code + OSI license = open source

Photoshop binary = open weight

Photoshop SAAS web app = closed model like GPT, Opus/Fable etc.

There is nothing "open source" about the Chinese models in question. All they're doing is allowing you to run their binary yourself instead of through their API.

If you want actual open source then you would need to look at like OLMo 3

https://allenai.org/

[thoughtpeddler]

It's true that there isn't currently a single Chinese model family that matches the full OLMo/Nemotron end-to-end 'cookbook' one could rightly call "open-source" (i.e. a full manual on how to reproduce your own foundation model by including the models, data, checkpoints, evals, and code for base/thinking/instruct/RLVR)...

But! the Chinese ecosystem does have most of those components, just that they're split across separate projects: MAP-Neo or YuLan-Mini for transparent pretraining, plus DAPO/verl or Open-Reasoner-Zero for transparent reasoning RL.

My bet is it's just a matter of time until there is a Chinese equivalent to OLMo. It won't just be China either - from the public sector across the globe there are efforts ranging from Apertus in Switzerland to LLM-jp-4 in Japan, and more. They're not OLMo-level but I'm sure by 2027 you'll see fully open-source 'cookbooks' help an enterprising individual reproduce frontier models of at least a 2023-2024 vintage.

[SwellJoe]

Open Source AI is challenging. Even tiny models require tremendous resources to produce, so all the things we know about "open source" for software, like a random person in Nebraska can produce a critical piece of the world's infrastructure in their spare time, don't apply.

Actually open source AI is more like scientific research. It needs public funding and reputable institutions as stewards of that funding. I'm hopeful Allen AI is able to do the things, but I'm not sure the US is moving in the direction it needs to be to get to the point where AI in the public interest is something we invest in.

But, also Photoshop is not at all comparable to open weight models. I can't legally give you a copy of Photoshop, but I can give you a copy of GLM 5.2. There are free proprietary applications out there that fit the bill, Photoshop isn't one of them.

[petcat]

OK so replace the Photoshop example with a compiled GIMP or shareware binary. It doesn't change the fact that it is not open source.

The free software community complains all the time about binary blobs that are otherwise legal to freely distribute. Like firmwares. But somehow this is all overlooked with these "open weight" models.

[est31]

> The free software community complains all the time about binary blobs that are otherwise legal to freely distribute. Like firmwares.

With software there are two things: the source code, and the binary.

Firmware usually has restrictions on modification, so it violates the four freedoms.

Open weights is not perfect, but at least you can finetune as you wish, and for example add more knowledge about the current day to those models, or give it knowledge about company internals so you don't need to put stuff into the context.

Many smaller AI labs finetune chinese models and release the result. Finetuning is very valuable!

[elpocko]

I agree with you in general, it's not "Open Source". But in many cases you're allowed to fork them, and create and distribute derivative works (finetunes) from open models, and the training and inference code is also permissively licensed. That's not shareware.

And working with the binary weights of a huge pretained model is much easier and cheaper than doing it based on the entirety of its humongous source datasets.

[SwellJoe]

I'm not at all disagreeing with you about calling these models "open source". They mostly aren't and we should stop conflating the two things.

[thewebguyd]

I think using the term "open" at all in relation to these models is a misnomer. Much in the same way we've spent a lot of time and effort distinguishing "open source" from "source available"

Perhaps instead of "Open weight model" it should be "Weights available model"

[SwellJoe]

Definitely, but there's also a distinction between "weights available" and "you can do stuff with those weights". There are models you can download but the license prohibits redistribution, fine-tuning, etc. The earlier Gemma models had license encumbrances, for instance, that Gemma 4 do not.

[jrm4]

But again, I think you're dancing around "Stallman was right," which is frankly why I come to the reverse idea.

I think it is fine to call these things "open" because "open" was loosey-goosey already. It wasn't Free.

[DonHopkins]

Lord forbid open source models have as horrible a user interface as GIMP!

[parl_match]

> Open Source AI is challenging. Even tiny models require tremendous resources to produce, so all the things we know about "open source" for software, like a random person in Nebraska can produce a critical piece of the world's infrastructure in their spare time, don't apply.

No, but a small university team with dark time on the state school system's cluster can build a basic, functional LLM.

It will be a few years (or more) until true "open source" LLMs are available and high quality, for sure. But it's not as impenetrable as it seems, IMO.

[SwellJoe]

"No, but a small university team with dark time on the state school system's cluster can build a basic, functional LLM."

Hey, remember when we were just talking about how it "needs public funding and reputable institutions as stewards of that funding"? (Hint: A state school system is a reputable institution running on public funding.)

[parl_match]

I have nothing to say back other than: That's a very snarky and unnecessarily hostile tone to take with someone who is agreeing with you.

[SwellJoe]

Sorry, I took "No, but" as a disagreement. You're right I was unnecessarily snarky about it.

[Onavo]

The challenge is all legal. Every single AI shops is training on pirated data corpuses and grey area scraped data so obviously not a single company is willing to reveal their data.

[embedding-shape]

> Open Source AI is challenging. Even tiny models require tremendous resources to produce, so all the things we know about "open source" for software

Compiling/creating the Linux kernel or Chromium is no child's dance either, doesn't make them more/less FOSS than other things. Open Source AI gets its name from the license, not how easy/difficult it is to run/produce yourself.

[SwellJoe]

"Compiling/creating the Linux kernel or Chromium is no child's dance either"

I can compile the Linux kernel on a modest 15 year old laptop. Sure, it took 35 years and thousands of people to build it into what it is today, but anyone with pretty much any computer can meaningfully participate in Linux kernel development, and many Linux contributors have done so using modest hardware. The same is not true of AI. And, the Linux kernel is the biggest open source project, but plenty of small ones with one or two developers are in use on millions of systems.

I have pretty big hardware for local AI, more than most people have (a Strix Halo and a couple of 32GB GPUs in my desktop), but I can barely train anything useful locally; I can do QLoRAs for small models, or LoRAs for very small models, that's about the extent of it. I would need to rent big GPUs to do anything more than an experiment.

That's not comparable.

[jononor]

Not comparable today. But in 10 or 20 years the situation might be considerably different. Hopefully closer to a everyone-can-build-using-hardware-they-already-have.

[antonvs]

> But, also Photoshop is not at all comparable to open weight models. I can't legally give you a copy of Photoshop

Please tell me this level of obtuseness is deliberate.

[SwellJoe]

I don't know what you mean, so, if I am obtuse it is from a genuine lack of understanding. What is obtuse about that statement?

[shwaj]

Equating an open weight model to a closed source binary is ludicrous. An open weight model can be fine tuned, for example. How easy is it to modify a closed source binary for your purposes?

[throwaway27448]

I see you didn't grow up cracking apps and games, this is quite easy with a hex editor and bintools. Well, until codesigning came along, but that doesn't seem relevant to this analogy.

[shwaj]

I didn’t grow up with that hobby. The point stands anyway.

Cracking a binary circumvents authorization to provide full access to functionality the program already has. That’s very different from modifying the program to add new features.

[throwaway27448]

Not really. It's a social structure that separates these two things, and a social structure nobody really needs to recognize.

[kennywinker]

Cracking games is and was illegal. Fine tuning an open weight model isn’t. Well, not yet anyway

[throwaway27448]

> Cracking games is and was illegal.

Only distribution.

[kennywinker]

Don’t think that’s correct, i think creating the cracked version of a piece of software is infringement of copyright even without distribution.

But even if distribution is required to trigger the law - you are free to distribute fine-tunes of most open weight models.

[TaLiTr]

> creating the cracked version of a piece of software is infringement

Depending where you are it might be "DRM circumvention" all the way up to some bollocks hacking charge.

It's pretty murky in places because lots of copyright laws protect DRM but also have carve-outs for personal backups.

[selfhoster11]

Interoperability clause in the EU might make it legal? Not sure.

[amelius]

The big AI companies have shown us that caring about copyright is apparently for losers.

[otterley]

If the rapid rate at which games in the 1980s-2000s were cracked is any indication, pretty easy indeed.

[Groxx]

If it's not obfuscated, it's often not very hard. Much of the time you can just LD_PRELOAD stuff too, which I think has a rather strong claim at being roughly equivalent to open-weight models.

[m4rtink]

Funny enough, it seems to be surprisingly easy with LLMs from what I've heard. :)

[Groxx]

Especially since some of the arguments in the article seem to hinge on "just fix the training!", yeah, I think this is a completely fair call-out.

Open weights can sometimes get additional training, but you can't remove existing training, so there kinda isn't a fair claim to "just train it to be [nationality]". That would need "real" open source so you can train a realistically-equivalent model from the ground up.

[edflsafoiewq]

The big labs don't generally have the resources to remove stuff and re-run training again either. That's reserved for new models. Existing models are modified by operating on the weights even by the model creators, which exactly satisfies the GPL's definition of source as the "preferred form for making modifications".

[Groxx]

You're still describing the relative dregs we have now and saying "this is Open Source because it is the dominant form that exists". That's like saying "Windows is open source because it has documentation and everyone uses it. You can even modify registry values!"

I don't agree. At all. Existence and utility are not the defining characteristics of "open source" in roughly any reasonable definition. "Open weights" is descriptive and more than good enough, just use that.

[jononor]

Even in a fine tuning scenario, the preferred form for making modifications is the weights plus the training pipelines, including evaluation protocols, tooling, etc.

[syrrim]

This is a facile and pointless discussion. The hard work in building source code is paying millions of dollars to software engineers to write it. "Open source" allows others to benefit from this expenditure. The hard work in building an ai model is spending millions of dollars on training. "Open source" models allow others to benefit from this expenditure. Regardless of whether you had access to the complete training data and architecture of a model, you would never retrain from scratch in order to make a modification to the model. You would fine-tune in whatever behaviour you wanted. Just like open source software is when you are allowed to modify it, open source models are when you are allowed to fine tune like this.

[dspillett]

> Regardless of whether you had access to the complete training data and architecture of a model, you would never retrain from scratch in order to make a modification to the model.

Just because the concept is (currently) impractical in a given realm, doesn't mean we should just use the concept's name for something different in that realm, surely? Using the phrase "open source" for something that arguably isn't for essentially marketing purposes feels at best disingenuous.

Also, recreating from scratch is far from always the purpose of being able to see/investigate how something works and was created.

[ece]

Controlling how to train, not over fitting, making sure training data is good, and being able to do distributed training are all reasons why open weight isn't like open source. You can't fork during pre-training with open weight-only models.

[georgemcbay]

> This is not "open source" AI.

I both know what you mean and I agree with you, but attempts like this to maintain some sort of linguistic purity of meaning never work once the discussion escapes small communities into the larger mainstream.

So you should probably get used to the idea of people saying "open source LLM" when they mean "open weights LLM".

[zem]

the distinction is important though, because the OP throws in the point about how "if you think the Chinese models have political propaganda baked in, go ahead and remove it, it's open source!". it is not at all obvious to me that having the model weights available will make this sort of retraining easy or even possible, but framing them as "open source" makes the analogy with open source software an easy one to get misled by.

[georgemcbay]

> the distinction is important though

Only to nerds. Of which I am one, so I'm not trying to be personally dismissive here.

But I'd argue the average person only really cares about the legal aspects of duplication (and not their ability to modify things) both when it comes to open source software and open weight LLMs.

> it is not at all obvious to me that having the model weights available will make this sort of retraining easy or even possible

people already both fine-tune and distill from existing models all the time without the original training data or source and having the weights be openly downloadable makes that a much easier process.

[Capricorn2481]

> But I'd argue the average person only really cares about the legal aspects of duplication (and not their ability to modify things) both when it comes to open source software and open weight LLMs.

The average person does not know the term "open source" to begin with, and is 100% not going to run an open source weight model on their computer. If they are, they will call them "free" models. I think the distinction is only going to enter the periphery of nerds.

[georgemcbay]

> The average person does not know the term "open source" to begin with, and is 100% not going to run an open source weight model on their computer.

I agree that the fully average person isn't going to call them "open source models" (or really anything more specific than "AI") but policy makers and the media having this discussion will call them "open source models". And they will be technically incorrect, but it won't matter, the name will stick.

And yeah, it will be quite a long time before average people run open weight models on their computer, but I think a significant amount of them will run hosted open weight models in the nearish future. It won't matter to them that these models are open weight, it will only matter to them that they cost less than Claude or GPT while doing basically the same stuff just about as well.

[Capricorn2481]

I just don't see how policy makers, media, or the average person would find any utility in calling something "open source" over "free." Maybe an actual legal bill will. But in everday speech, everyone will refer to it as "free AI", because it describes the only part they actually care to know about.

Now I am living in a world where "ChatGPT" was being used in every day conversation so I can't underestimate people

[Otterly99]

Open weights models are trainable. However, I would not count it as "easy" as you generally do not have access to the training material, so there is a substantial risk of model drift if you use weight-based retraining.

However, with RLHF you don't need to have access to the original training set to steer a model towards a preferable behavior.

[nl]

The two aren't comparable.

You can't reconstruct the same LLM weights even with the same training data.

You can fine tune an open weights model.

Much of the value in modern LLMs is in the RL envrionments, not the pre-training data. That is a very different piece of software to an LLM, and the LLM weights are useful without it.

[waterTanuki]

When talking about "open source" models no one is talking about the legal framework and textual doctrine of open source. They simply want to know: "can I run this on hardware I own or no?" That's it. It's simply about control. If you don't think there's a difference between models you can download and run on hardware you own, regardless of how big it is, vs. models you can only talk to through an API, you are seeing the forest for the trees.

You are also indirectly helping the closed SOTA model toll-keepers by taking the focus away from their mass copyright theft and onto some pointless textbook definition.

[joe_the_user]

It's a matter of degree. If open source means "everything you need to reproduce this" then maybe an open source application should include a detailed architecture document and course in the c language. Oppositely, you can certainly use models to reproduce model - the Chinese models themselves were supposedly created by interactions with the American models so you can use them/

[swiftcoder]

> This is not "open source" AI.

The OP actually acknowledges this in the footnotes. Perhaps we don't need to re-litigate this on every post?

[austin-cheney]

Does that then make the title deceptive click bait?

[swiftcoder]

That really depends how anal-retentive one is being about semantics. My own feeling is that open-source is itself pretty meaningless ever since corporations figured out how to capture all the value in the ecosystem...

[kakwa_]

And at the risk of being a downer, I'm not sure Open Source LLMs is actually viable.

Training a model requires two or three orders of magnitude more investment than the typical OSS/Creative Common contributor can reasonably afford.

It reminds me a bit of Open Source hardware (CPU, GPU).

[trinsic2]

Anyone got a starters guide for setting up an open source LLM? I'm pretty new to AI, but technical. Im looking to setup my own open source LLM on my workstation to help me code if its possible. Been using Claude and I tried opencode recently.

[kennywinker]

Unless you have a LOT of VRAM (>16gb), you’ll want to use a 2, 3, or 4 bit quantization of qwen3.6-35b-a3b. Ollama is a quick way to try this out and see what perf you can get, tho you’ll want to move to a better tool (vLLM, llama.cpp, etc) if you end up using it seriously.

Use pi.dev as your agent harness, and point it at ollama or whatever you’re running your model in.

Since qwen3.6-35b-a3b is a “MoE” model (indicated by the a3b part), you can get away with a lot less VRAM and a slower GPU. For me, 35b-a3b performs about as well as a 12b “dense” model. You’ll still need enough RAM+VRAM to hold the entire model - otherwise it’ll be super slow as it loads pieces of the model off disk for every token. Loading from RAM to VRAM is slower than having the whole thing in VRAM but better than involving your ssd.

[antonvs]

It’s a good analogy. One difference is that it would be much more difficult to “fine tune” Photoshop, so there’s a sense in which free weights are more useful than a free binary.

[gorgoiler]

I sort of agree with your analogy but in another sense, think of Firefox.

Could I have made it myself? No!

Am I grateful that I can have the output of Mozilla’s internal efforts — the browser source code — so that I can at least tweak it and rebuild it myself, and even ship the results myself? Yes!

A open weight model is both an incomprehensible binary blob like Photoshop, but also patchable and adjustable piece of source code, like Firefox.

[40four]

This is a great distinction that I don’t think is getting talked about enough. “Open source” is probably the wrong term to use. Open “weights”, sure.

If your only concern is how good is it at coding, then I don’t have an issue with using the Chinese models. Especially if you want to run it locally, they are kind of the only choice.

For any other use than coding, it’s going to have to be a hard pass from me.

[TurdF3rguson]

Well I don't see how it even could be open source, since the source is distilling closed weights models, lol.

Also, who cares about having the source? Are you going to tweak it + spend a billion dollars training to get slightly different weights? Who would do that?

[_carbyau_]

Analysing the source can provide information about the model without ever running it. Does it have an inbuilt aim/censorship/bias/certain-capabilities?

[TurdF3rguson]

There's a much easier way to get that information though. Just ask it some questions. Would you rather do that or analyze TBs of training data?

[kennywinker]

Since you can’t analyze the source for biases in any models - chinese open weight or US closed weight - how is that a dealbreaker for some but not all?

[BoingBoomTschak]

Let's reuse a good ol' word: shareware LLM!

[prometheus1992]

>>There is nothing "open source" about the Chinese models in question.

hard disagree here. openness can have a scale and on that scale Open AI is less open than zai. olmo will be very very open on that same scale.

[anon373839]

Chinese models are open-source for purposes of inference: you know exactly what the model architecture is and the code to run it is released under an open source license. The “binary blob” is a useful artifact that has many techniques available to use in the creation of future binary blobs. And it is fixed and reliable.

Contrast this with crap like Claude, where you have no idea how many/what size/shape models you are interacting with. They can silently inject outputs from other systems, apply steering vectors to sabotage a particular user’s work, etc.

I agree that 100% free-range organic open source models are lovely, but they also are unlikely ever to attain parity with the frontier because the training data sources are open and expose them to legal liability that closed model providers can thwart.

[jrm4]

The more I see this argument, the sillier it becomes to me. The point of "open source" in software is something like "the ability to deconstruct the thing line by line for purposes of predicting what it will do."

But you can't fully do that for generative AI anyway? What's the point?

Now, if you wanted something more robust, then you probably should have supported Stallman's ideas and insisted on Free Software in the first place.