| ▲ | petcat 6 hours ago |
| > long live open source models There are no open source models, at least not useful ones (yet) [0]. Open weight is not the same as open source. The current "open weight" models are just opaque binary blobs you can run on your own computer instead of through a web API. [0] https://allenai.org/ |
|
| ▲ | adastra22 6 hours ago | parent | next [-] |
| This viewpoint doesn't make any sense to me. The weights + inference code are the "source code" for AI. I literally don't know what else you are demanding for the "open source" label. |
| |
| ▲ | kingstnap 5 hours ago | parent | next [-] | | If you think of LLMs as programs. The weights and inference code are very much a binary. While the training code and data are the true source. Since if you want to robustly modify the LLM that's actually what you need. But since "compilation" (training) is extremely compute intensive this isn't something accessible to anyone without an entire datacenter. Anyway semantics aside having the binary is still infinitely better than dealing with an api as far as privacy and control go. | | |
| ▲ | adastra22 2 hours ago | parent | next [-] | | That’s not how it works though. Two training runs on the same data don’t produce the same weights. And if you want to modify the AI, you do so by fine tuning the weights not rerunning training. In every respect that matters, the weights are both the binary and the source code together. | | |
| ▲ | stavros an hour ago | parent [-] | | If I want to remove censorship from your open-weights model, how do I do that? | | |
| |
| ▲ | Vetch 2 hours ago | parent | prev | next [-] | | The weights + the architecture are already 100% of the code, the transformer is just a mathematical expression + helper programs whose sources are provided. The transformer itself is not even a stateful program, so a it is no more a binary than Piet or Tromp's BLC are. It's merely incomprehensible. Training isn't compilation either, since training a model is closer to program induction and the data are samples defining the solution space. | |
| ▲ | kzrdude 4 hours ago | parent | prev [-] | | I agree. But models in difference to compiled binaries, are useful as just weights and can be further refined and post-trained, at least. I don't know LLM theory well enough to say if there's some secret sauce they can hold back that makes training ineffective. Less effective I'm sure, we don't have access to their smart training schemes, but post-training should always be possible IIUC. | | |
| ▲ | MeetingsBrowser 2 hours ago | parent [-] | | At the risk of taking the analogy too far, I would treat refining like modifying a dynamic library. You can technically modify behavior, but only in a very coarse way. post-training is like writing a wrapper around the binary. It is closer to building on top of than truly modifying, in that you can tailor things to your needs slightly but cannot make fundamental changes to the underlying thing. |
|
| |
| ▲ | chasd00 5 hours ago | parent | prev | next [-] | | > I literally don't know what else you are demanding for the "open source" label. you need to "literally" go read the definition of open source software or even ask an LLM to define it for you. Weights + inference code are not the source code they're more like the compiled binary. Making modifications to the behavior of a model with additional training is like writing a mod for minecraft. Sure, you can change things but it doesn't make it open source. Calling these models "open source" is an old trap that software companies use to use. Free to download but then, once you're fully comitted, the trap snaps shut and you must pay up to continue. | | |
| ▲ | jmiskovic 3 hours ago | parent [-] | | The vendor lock-in trap does not apply here. Users don't actually even run the original weights, they mostly use open source inference engines with open source agent harness. If AI provider decided to start charging, a) released models would not be affected, and b) people would drop it instantly and move to other models. The actual problem is that we know nothing about the training set of any open-weights model. They could be intentionally biased to influence users, from political censorship to brand advertising, or general shaping of cultural norms. You run the model on own hardware not knowing if it is designed to act against you. Having whole chain open source would allow audit and reproducing the results. |
| |
| ▲ | keketi 5 hours ago | parent | prev [-] | | > I literally don't know what else you are demanding for the "open source" label Training data and code. |
|
|
| ▲ | wasfgwp 6 hours ago | parent | prev | next [-] |
| Nemotron Super is sort of open source in the sense that Nvidia provides almost everything you need to replicate it from scratch. Of course it’s performance is not exactly stellar but it could be a good starting point for other research teams. |
| |
|
| ▲ | drusepth 5 hours ago | parent | prev | next [-] |
| Can you explain what's missing for the "open source" label that open-weight models like DeepSeek/Quen/GLM/etc don't release? Is it just the supplementary data/code for how they were trained, not just the final product? |
| |
| ▲ | petcat 4 hours ago | parent [-] | | The open weight model providers don't provide the training data or the build tooling. You cannot reproduce the model yourself, or even really know what the model contains. They don't even provide high-level catalogues/descriptions of the training data. An improvement would be something like "trained on the entire WWW up to Aug 1 2026". Or "trained on a Wikipedia archive + Anna's Archive and everything we were able to scrape from Github". They don't provide any of this stuff. I don't mind open-weight models, but they are not open source. It's like bringing home a dog from the rescue and just hoping that it doesn't have a history of biting kids in the face. You just can't know, because you don't know the full history. You can try to add new training (fine tune) to tell it not to bite kids, but that's it. | | |
| ▲ | drusepth 4 hours ago | parent [-] | | Thank you for the detail/explanation. Hope my comment came off correctly as curiosity and a desire to understand, which this helped with. |
|
|
|
| ▲ | Matumio 6 hours ago | parent | prev | next [-] |
| There is also https://apertus-ai.org/ but yeah, "not useful yet" if you were looking to replace your coding agent. Very useful if you are doing LLM research. |
|
| ▲ | prometheus1992 5 hours ago | parent | prev | next [-] |
| I really like molmo 2 |
|
| ▲ | sleepgummies 6 hours ago | parent | prev [-] |
| [dead] |
