I hoped that someone would fork OpenWebUI before the license change an maintain a OSS Version. Like Valkey¹ is the OSS Fork of Redis and OpenSearch² is the OSS Fork of ElasticSearch.
That might still be an option, all code before a license change falls under the license at that time.
A new license legally becomes active once it has been made public (i.e. it is committed to the source tree) and is not retroactive (for legal reasons).
And because of this, unless the owner/maintainer is willing to nuke a huge chunk of the history within their public repository (with all repercussions which might come with it) it is still possible to make a safe fork of a open source / open core product suddenly moving into a different direction. Time travel to the point before the new license was added to the repository, and fork from that point in time.
The owner/maintainer involved might absolutely not like that but legally they have no choice to allow it, given the license and its permissions at that particular moment in time in the repo.
> That might still be an option, all code before a license change falls under the license at that time.
This is technically true, but as more time passes, it will be more and more unlikely that someone is willing to fork an old codebase, especially since the newer version has so many bugs already fixed and the old version gets outdated in many aspects fast.
So if there is not any active fork right now that is backed by a community, then it's more likely that we would see alternatives coming up like the one OP mentioned here.
OpenWebUI and similar ChatGPT-lite interfaces are a bad UI/UX for LLMs. "Normies" should not be using local models and those who want local models should be exposed to the settings/complexity that projects like oobabooga/sillytavern/comfyUI force a user to educate themselves about (i.e. LLM sampler settings, chat templates, etc). It's telling that the best AI generated content is made by people using these prosumer style projects, and not using simplified trash like OpenWebUI.
The proliferation of bad UI in consumption of AI systems has done unfathomable damage to discoverability of LLM features, especially new innovations in LLMs like far better sampling algorithms (i.e. min_p). Users are massively harmed by the features they don't know about and wish they were using but don't/can't because they are shepherded into the most brain-dead simple UI possible.
If we had a society of actual engineers and hackers, sure.
However, LLMs for the vast majority of people are simple chatbot oracles. The people paying exorbinant wealth in investments are aiming for essentially the Apple of AI where it magically just works and creates a new market to redefine the paradigm.
LLMs are yet again another cyclical cycle where ideas influence material reality and vice versa. Magic is seen to be worth more than the wizard and his tools behind the curtain. The market is hoping that the masses don't find out about the wizard and his tools, so the illusion can continue to live and provide the basis for dreams.
Open WebUI also lets you tweak those parameters, BTW.
Been using both, like Chatbox for how snappy it is, but is local only, vs Lobechat which allows you to setup centralized host to have shared host across clients but feels a bit clunkier.
I've always thought it was incredibly rude to call this project something as incredibly generic as "Open WebUI". For one it's impossible to find documentation because it's so generic as tl be utterly meaningless, and also you're intentionally confusing any and all open web UI projects past and future.
Not to mention it's just janky and hard to work with. And of course every bit of the official documentation is LLM slop. The only way I've been able to navigate is to use another LLM to tell me what these settings are for. Said LLM, having ingested the LLM slop documentation, also produces garbage that's barely correct.
13 comments
[ 2.5 ms ] story [ 43.5 ms ] threadI'm leaning towards this one myself but I'm still early in researching alternatives.
[1]: https://github.com/valkey-io/valkey
[2]: https://github.com/opensearch-project/OpenSearch
A new license legally becomes active once it has been made public (i.e. it is committed to the source tree) and is not retroactive (for legal reasons).
And because of this, unless the owner/maintainer is willing to nuke a huge chunk of the history within their public repository (with all repercussions which might come with it) it is still possible to make a safe fork of a open source / open core product suddenly moving into a different direction. Time travel to the point before the new license was added to the repository, and fork from that point in time.
The owner/maintainer involved might absolutely not like that but legally they have no choice to allow it, given the license and its permissions at that particular moment in time in the repo.
This is technically true, but as more time passes, it will be more and more unlikely that someone is willing to fork an old codebase, especially since the newer version has so many bugs already fixed and the old version gets outdated in many aspects fast.
So if there is not any active fork right now that is backed by a community, then it's more likely that we would see alternatives coming up like the one OP mentioned here.
The proliferation of bad UI in consumption of AI systems has done unfathomable damage to discoverability of LLM features, especially new innovations in LLMs like far better sampling algorithms (i.e. min_p). Users are massively harmed by the features they don't know about and wish they were using but don't/can't because they are shepherded into the most brain-dead simple UI possible.
However, LLMs for the vast majority of people are simple chatbot oracles. The people paying exorbinant wealth in investments are aiming for essentially the Apple of AI where it magically just works and creates a new market to redefine the paradigm.
LLMs are yet again another cyclical cycle where ideas influence material reality and vice versa. Magic is seen to be worth more than the wizard and his tools behind the curtain. The market is hoping that the masses don't find out about the wizard and his tools, so the illusion can continue to live and provide the basis for dreams.
Open WebUI also lets you tweak those parameters, BTW.
Been using both, like Chatbox for how snappy it is, but is local only, vs Lobechat which allows you to setup centralized host to have shared host across clients but feels a bit clunkier.
Not to mention it's just janky and hard to work with. And of course every bit of the official documentation is LLM slop. The only way I've been able to navigate is to use another LLM to tell me what these settings are for. Said LLM, having ingested the LLM slop documentation, also produces garbage that's barely correct.
It's a bad project.