I usually don’t even understand the lingo they use. “Open-weighted” is the most recent one, then it usually goes down to specific “models” that everybody is supposed to know about.
These are my thoughts (I will stick to the vague “it” for now, but of course therein lies another question: “and how does all this apply to various specialised AIs”):
- Is it really feasible to run it 100% locally? I know there’s plenty of people with very powerful rigs indeed, but still. Or are 99% of these people really saying “it would, in theory, be possible to run that locally, therefore your concerns are invalid”?
- If yes to the previous: the software doesn’t come from nowhere and ultimately still relies on gas-turbine-powered datacenters and stolen IP and stolen personal data, no?
If what I wrote above is true, what exactly are people arguing when they say it’s still possible to use LLMs ethically or true to FOSS philosophy, because … ???
edit
Thanks to all who answered.
I guess it’s my fault for asking several questions in one, but this thread has attracted exactly the type of people I’m writing about; several even used the term “open-weighted models” without explaining it.
Asking to get arguments explained, I got more arguments instead.
Ok seeing your edit, some basic explanations:
A Token is a small unit used by LLMs. Mostly not a full word, more of a word fragment of a few characters.
A Dataset is an accumulation of different data (mostly real-life data to avoid problems) which is used to train a Model.
A Model is basically a huge file, full of Tokens that are interconnected with each other.
Open Weights means that the numbers which describe the relationship between Tokens (this is all basically a ton of matrix multiplication, which is the reason this is using so much power) are available.
If you want to really understand, 3Blue1Brown has a complete series about neural networks!
I’ll go one by one… although I do have to say, I can’t answer the main question of whether it can be 100% ethical
Open-weight model
It’s actually quite literal. LLMs may look like magic, but under the hood are basically very, very big mathematical/statistical functions that transform one type of input (for example, “User: hey ChatGPT what is an open-weight model?”) to an output (example, “ChatGPT: To understand what an open model is…”)… To simplify, these models function by piecing together a ridiculous amount of tiny, nonlinear mathematical functions, apply a weight to each of these tiny functions, and somehow get a coherent result out of it. The theory behind this has been known in math research for decades before OpenAI.
Because LLMs are mathematical models, if you know the weights of those tiny functions and how you chain theme together, you can run an exact replica of it. Open-weight models are just that, models where the entire weight and model architecture are released with open, permissive licenses. Many of these end up being modified by other users/hobbyists, and I do believe this sharing/modifying part of AI/LLM is quite true to FOSS philosophy.
As others have mentioned though, “open-weight” has no bearing on how the model is trained. It just means that the final product is free and open
Is it really feasible to run it 100% locally?
Theoretically yes, practically yes but depends on the model.
- For the really small models designed for mobile phones, absolutely. For example, LFM-2.5 and Ling-3.0-tiny run on just about any potato. Granted these models are… not great. They are probably functional as “search engine wrappers” though, methinks that’s what they were built for.
- For the intermediate models, yes but with an asterisk. They require at least a decent GPU, so either you have a gaming rig, happened to have snagged one of the AI computers, or try to kneecap the model in some way (limit the amount it generates, make the mathematical operations less precise a.k.a. quantization, be okay with it running for a long time, …). I have a local setup for Qwen3.8-27B with 4-bit quantization, but it runs at 2 tokens/second on my laptop (which has a ton of RAM but not a very good GPU)…
- For really large models, you can technically run them, but they are realistically designed for organizations with privacy in mind: hospitals, private companies that handle sensitive data, etc…
The software doesn’t come from nowhere
I mean IMO that’s the heart of the whole debate. Making a modern LLM requires massive amounts of data and computation… both of these do have to come from somewhere. And my understanding is that, current AI/LLM research didn’t create that much technological breakthrough, so building a better model often does indeed rely on either using more data or bigger architecture (which takes longer to compute)… which then leads to a truckload of ethics issues here and there
For private companies, we know that OpenAI is blatantly hoarding data from just about everywhere, see the recent Mathematics Millennium Prize scandal. A lot of the open-weight models are trained by Chinese training houses, and Chinese tech companies are not known for respecting intellectual property… and China uses lots of coal energy still.
Maybe Mistral is more ethical, given that 1) they are based in France and care about GDPR and data regulations a lot more, and 2) France uses lots of nuclear energy, but I can’t say for sure. An European company operating an AI/LLM training house from Iceland using their excess geothermal energy is probably the most ethical we can get on this
In theory (and take this with a whole jug of salt) someone can run an off-grid farm and train a small LLM all on their own, which would be very ethical… but I doubt any serious organizations are doing it this way. Most ppl just want better performance and/or more optimized models for now
For a complete FOSS purist with an environmentally conscious angle, I don’t know if there is an AI/LLM that can be considered 100% ethical… but then again Gentoo is a thing. So yeah, I don’t know
There’s a lot of info that you need to know to explore this space, so I’ll take my own shot at answering. Let me know if anything needs further explaining!
An “open weight” model is an AI model where you can download the data needed to run the model on your own hardware for free. Contrast this with proprietary models-as-a-service like ChatGPT and Claude where you have no access to the data needed to run the model yourself – you can only use it through the services provided, usually for a fee, and which can be taken away from you or changed at any time with no recourse.
The mapping to traditional open source terms does not work well since what you get is a binary artifact.
Those artifacts are released with a license – and many of the models are licensed permissively (e.g. MIT or Apache license terms). You can take those weights, modify them, and then release them as new models – and people do actually do this in practice!
Is it really feasible to run it 100% locally?
Yes. I run models on my own computers and have tried a number of configurations to figure out what works well. The Qwen family of open weight models (from Alibaba) are the ones I’ve found most useful so far. Gemma4 models (from Google) are also useful.
I prefer models that have been modified by the community to remove corporate censorship – i.e. stripping that “As a large language model…” cover-your-ass crap and evasiveness on topics like Tiananmen Square. If that means the model is technically capable of telling me to go kill myself too, so be it; I’ve spent 25+ years dealing with assholes on the internet and can handle abuse from a stupid robot if I have to. (In practice though, they’re usually pretty nice still unless I deliberately tell them to act like an asshole – and then Qwen, at least, starts to sound like a snarky redditor; it’s quite funny most of the time, actually.)
If yes to the previous: the software doesn’t come from nowhere and ultimately still relies on gas-turbine-powered datacenters and stolen IP and stolen personal data, no?
Models do require training to create, yes. It’s generally not clear where the training physically happened IRL – so, yes, some of them probably used power from gas turbines, but others may be drawing power from the Three Gorges Dam in China or solar plants or nuclear plants or whatever else is hooked up to the electric grid where the training happened. Most of them are also not very open about the data sets they were trained on. (There are exceptions to this though!) The Chinese models in particular are almost certainly trained heavily on logs extracted from Western models in addition to using whatever other data they could get ahold of. Whether you think that’s ethical or not is a matter of perspective; how do you feel about Robin Hood?
Once a model has been trained though, it can run on a normal GPU. The power requirements to run an LLM are basically the same as running a video game, or, equivalently, about the same as turning on a few incandescent lightbulbs. (The iGPU in one of my systems uses 100W; the discrete GPU in another system I’ve tried uses 215W under load with appropriate tuning – or 300W if you run it naively.)
If you want to run a model yourself, I recommend using llama.cpp – there are instructions on how to get started with it here: https://llama.app/
These are the models I’ve found most useful:
- Qwen3.8-27B (official version): https://huggingface.co/Qwen/Qwen3.8-27B
- Qwen3.8-27B (uncensored): https://huggingface.co/MuXodious/Qwen3.8-27B-absolute-heresy
- Qwen3.6-35B-A3B (official version): https://huggingface.co/Qwen/Qwen3.6-35B-A3B
- Qwen3.6-35B-A3B (uncensored – ⚠️ mildly NSFW graphics on page): https://huggingface.co/llmfan46/Qwen3.6-35B-A3B-uncensored-heretic
- Gemma4 (official collection – many variants in different sizes): https://huggingface.co/collections/google/gemma-4
If you have an iGPU only, I recommend using one of the so called “Mixture of Experts” (MoE) releases. These are typically named like 35B-A3B or similar; the first number indicates the total number of weights (35 billion) and the second indicates how many are “active” (i.e. actually used during computation) at one time while the model is running (3 billion in the example). These models need less computation to run and stay fast on weaker GPUs. Qwen3.6-35B-A3B is very good in this space and was my go-to model for a long time.
If you have a discrete GPU and enough VRAM, I recommend using a dense model (i.e. one that activates all its weights while answering) like Qwen3.8-27B.
It’s worth noting that people don’t usually use the full quality weights (which are typically ~2 bytes per weight); they use a “quantized” version – compressed in a lossy fashion like a JPEG. Going down to 4-bits (half a byte) on average per weight is about as low as most people like to go – you will see this indicated in names like Q4_K_M. (Quantized to ~4 bits with the K quantizaation scheme, medium variant.) Usually a bigger number is better in the sense of “closer to the original quality” – at the cost of needing more RAM.
Full quality weights are often found as safetensor files on HuggingFace. Quantized weights intended for use with llama.cpp are usually in GGUF file format.
Does that help?
If you can play modern games, you can probably run a local LLM. The more VRAM the better it’ll go.
For ethics, there’s two issues. The first is power consumption. Running a model locally isn’t consuming any more power than running a game that fully pushes your system. It’s not really a concern.
The second issue is content generation. It’s all created through the theft of content others have made. There’s no getting around that. As far as I’m aware there are no ethical models that specifically have not stolen content. If you’re just using it to chat with or whatever, I think it’s fine. You probably aren’t taking work from anyone, at least to any degree more significant than piracy. If you’re generating stuff to be distributed, then you’re distributing stolen work. If that’s ethical or not is up to you to decide.
For ethics, there’s two issues.
No, there’s seven main issues.
I feel like this question is asked in bad faith if you are able to list seven “main” ethical issues here. I don’t know if I could list seven different issues that aren’t just breaking down power consumption and content theft into smaller pieces, and I’m pretty anti-AI. I guess you could include hardware consumption too. Financing maybe, but that doesn’t play a part in local models generally. Care to explain?
Then you misunderstood the question. Re-read the title?
Is it really feasible to run it 100% locally?
Yes. My Macbook Air (M2) released in 2022 can run many publicly available LLM models. The ouput is not as fast as using a large powerful datacenter, but for my local needs, I’m not in a hurry. I get about 17 to 30 tokens per second speed running 100% locally.
If yes to the previous: the software doesn’t come from nowhere
For Mac users, the interface comes from here . For the specific LLM models that is a separate question for each.
and ultimately still relies on gas-turbine-powered datacenters
DCs generating power on-site is a relatively new phenomon because existing grids are at capacity, so the only way to bring new DCs online is locally generating power at that DC, usually using gas turbines or even worse, diesel generators. Most if not all of the publicly available models for running on your own hardware were built before those gas-turnbine-generating DCs were a thing.
The public models people are running now have existed for a number of years are likely made on regular utility grid power which is whatever that nation and region uses.
and stolen IP and stolen personal data, no?
The Llama LLM is made by Meta, so probably yes for that one. Deepseek is from an AI research lab in China. QWEN is from Chinese company Alibaba. We don’t know for sure the inputs that created the Chinese models. US AI companies claim a number of the Chinese models are derived from American LLMs, but I haven’t seen (or looked for) proof of these claims.
If what I wrote above is true, what exactly are people arguing when they say it’s still possible to use LLMs ethically
Likely they mean because you don’t have to pay a large LLM owner in a rent-seeking model to run LLMs, nor does a person’s use contribute to further development by those companies.
or true to FOSS philosophy, because … ???
With the open-weight models it doesn’t rely on a commercial license to use, and the interfaces can be truly open source.
OK but what does open-weight model mean?
“Weights” in LLMs are the “final answer” numbers from the results of model training, and these are the engine of the LLM model. Lets use chocolate chip cookies as an analogy.
The chocolate chip cookies are produced with ingredients, a recipe, labor effort to produce the dough, and then cooking energy/effort to bake chocolate chip cookies. In a traditional AI company, the company gets the ingredients, they write their own recipe, do all the dough creation, and then the energy/labor for baking, and charge you money to get chocolate chip cookies.
An open-weights AI company got all the ingredients, wrote their own recipe, labor effort to produce the dough, and instead of charging you for the dough, you get as much uncooked dough as you want forever. Your only task is to take the dough and cook it yourself and you’ve got free chocolate chip cookies. You can make as many cookies as you want with your own oven. However, you are not given free raw ingredients, nor are your given the recipe to alter it in a way you might like. You can only get the dough for free.
So open-weight AI models (uncooked dough) are LLMs you can use on your own computers (oven) for free and have LLM output ( chocolate chip cookies), but you don’t get the training data (ingredients) nor the training parameters (recipe) that built the open-weight model.
Thanks, that was properly eli5’d!
It means the weights (the numeric data that makes up the model) are publicly available to download and use for free.
In some cases there are conditions such as, if you run the model as your business, you need to purchase a license. That’s usually the largest, most powerful models, though, not those most people would be able to run at home.
Question, what do you do with it and how is the data that underlies the model harvested?
Question, what do you do with it
I use it for both fun throwaway stuff as well as productivity tasks for learning.
Example of fun throwaway:
Do you remember that episode of Seinfeld where the Kramer starts using Facebook marketplace and starts buying the most worthless items before being robbed when trying to get a too-good-to-be-true sale? No? Because it never happened, but you can plug that premise into an crafted LLM prompt and it will pop out a whole TV script with in-character dialog for each actor as well as use of popular existing sets.
Foreign language learning:
I’m studying a foreign language and want to interact with just the level of vocabulary and grammar I have knowledge of right now at my level for practice. I can prompt the LLM to limit itself to just what I know now and adjust the conversation level so I can practice. If I ever get stuck, I can ask the LLM to explain the grammar usage or vocabulary choice.
and how is the data that underlies the model harvested?
I’m not sure what you’re asking here. Are you asking, for example, how the Deepseek model was trained? If so, I answered that above. If not, can you rephrase your question?
You run a program like llama.cpp that can use the weights to run the model, which you can then use for whatever you’d use a model for: figuring out tech stuff, coding, etc. It’s a bit involved, but there are tools that make it easy to get started. LM Studio for instance.
In some cases, the lab that created the model publishes the dataset that it was trained on, and those are usually made of publicly available data. In most cases, though, the labs don’t give details, but the answer likely involves siphoning every web page they could find.
Is it really feasible to run it 100% locally? I know there’s plenty of people with very powerful rigs indeed, but still. Or are 99% of these people really saying “it would, in theory, be possible to run that locally, therefore your concerns are invalid”?
Very useful 35B models need (more or less) 8GB of VRAM and 32GB+ CPU RAM to be usable locally. 16GB RAM might work on a lean system with an RTX Nvidia card and an exl3 quantized model.
I know we are in a RAM apocalypse, but pre-apocalypse, that’s a quite reasonable requirement, IMO.
Personally, I run Deepseek V4 07-31 Flash at 19 tokens/second on a desktop with a single RTX 3090 and 128GB CPU RAM, and that’s an extremely capable model. Again, that’s expensive these days, but pre-ram apocalypse, that is not an unreasonable workstation/homelab.
If yes to the previous: the software doesn’t come from nowhere and ultimately still relies on gas-turbine-powered datacenters…
Most open weights LLMs are Chinese. And they are:
-
Trained on pennies. They have to be, as they simply do not have a sea of GPUs like Big Tech. Training costs for their large models are in the millions or tens of millions; a single steel forge has used more energy than all of those training runs combined.
-
China relies more on renewables, and I believe the datacenters aren’t so hastily constructed with gas turbines in the middle of cities.
-
And as of now, they are transitioning away from Nvidia GPUs. Some labs already use Huawei accelerators.
…and stolen IP and stolen personal data?
Yep.
This is a huge caveat.
You can avoid this. Nvidia Nemotron models, for example, are trained on completely open datasets you can download and inspect yourself: https://huggingface.co/nvidia
They are very good, but just behind state of the art.
But in practice, the SOTA models most run use private datasets. Lord knows where the Chinese get it from, but given some common quirks between models, at least some data sources are shared (and possibly government provided?)
…However.
I would argue providing the result of the training as Apache licensed weights counts as “fair use,” in the same way non commercial fan works do.
They aren’t making a dime off releasing those weights. I’m not trying to sell anyone anything when I use them. Where is the IP theft if money isn’t changing hands?
Now, the Chinese LLM services they charge for? I have no excuse for that. Once money is on the table, it is definitely IP theft.
completely open datasets
Not completely – it is mostly open, but they use a dozen or so private datasets for things like training on global regulations, minesweeper (for some reason), etc. To their credit, they do indicate this on the model cards, but it’s not entirely clear what is in those datasets either.
(I fell for that bit of marketing myself awhile back.)
Where is the IP theft if money isn’t changing hands?
Money not changing hands is pretty much the definition of theft.
Unlike the definition of heft which is a photo of OP’s mom.
What I’m saying is it’s akin to writing a fanfic or making fanart of your favorite franchise. Or getting inspired by a painting you see, and making something similar yourself.
Do that for your personal enjoyment? That’s fair use, under the law.
But the moment you start trying to sell it is when you get in legal hot water, and when it’s indeed morally problematic.
The scale is different, but I’d argue a similar principle applies: if you use some model trained on public works from a protected IP, but the model and its outputs are not resold, nor profited from, it’s not theft. The point is beyond money not changing hands; there’s no profit being made from the original author’s stuff. They aren’t being taken advantage of any more than someone viewing their public stuff for free, or someone creating derivatives from private work.
But all that is off the table the moment profit and distribution is involved.
-
In my book, not paying and supporting them financially - > ethical
My issue is when you pay for a subscription, when you buy a gpu specifically for ai, when you buy a model, etc. Because then you are funding the slop machine.
To answer a lesser answered sub question of a question: what is a weight?
Imagine you have an output, its either a or b. You have an input, whatever it is. And then there is a weight in the middle. The weight determines wether the output will be a or b.
In math it would be a function input*weight=output.
ANN (Artificial Neural Networks, just afancy name) is a chain of a whole bunch of weights. There are layers and each layer generates an output from a bunch of weights and the next layer uses it as inputs. The name comes from the visualization, which looks like neurons firing to other neurons (from biology, things that make your brain work).
deleted by creator
There are ethically-sourced datasets used to train LLMs that can be powered by renewable energy and cooled by closed water systems. Maybe the question becomes “sure, but are those good for anything?” and the answer to that is “probably not right now”.
Is it really feasible to run it 100% locally?
Yes. My mid 7-year-old rig featuring Ryzen 5 1600X, 16 GB DDR4 RAM, and GTX 1060 6gb can run fairly intelligent models that are actually practical to use.
The software doesn’t come from nowhere
Correct. Training the model is the most resource-intensive and planet-churning part.
I don’t know if 5 year olds are allowed to watch 1.5h videos, but if so: this one has all the pro and con arguments explained nicely and makes a point why it’s not a good idea to shame people for using LLMs.
I have a friend who runs his AI locally because he doesn’t want to pay for tokens.
Is it really feasible to run it 100% locally? I know there’s plenty of people with very powerful rigs indeed, but still.
There are small models you can run 100% locally on a mediocre computer or even a phone, and I mean you could be 100% offline and it will still work.
The better local models require a higher end gaming GPU or an ARM Mac with a decent amount of RAM, but nothing too extreme. You could get a computer to run them for about $2000 to $3000.
the software doesn’t come from nowhere and ultimately still relies on gas-turbine-powered datacenters and stolen IP and stolen personal data, no?
The small local models typically start with a full infustrial-size model and then they “distill” it to smaller models. The concerns over training data are still valid. I’m not sure how much the concerns over power usage are about training vs running the full industrial size models commercially. Running a small local model uses a tiny fraction of the power it takes to run an industrial size model, and the training is a one-time cost (except these companies never stop training because they want to make next year’s model better and faster).
I do think it’s worth pointing out that running a model locally keeps your conversations with it private. Remember, it can be done 100% offline. So your data won’t be stolen.
Is it really feasible to run it 100% locally? I know there’s plenty of people with very powerful rigs indeed, but still.
It’s absolutely possible, and done more and more. Smaller models exist, and run on sub 8GB VRAM without any issue. If you have a gaming setup, you can run the bigger models without any issue locally. You don’t need Astra or Mythos or whatever. As a daily driver for “light” tasks (e.g. summarizing a document) small models perform without a significant difference to frontier cloud models. For bigger tasks (e.g. coding or agentic workflows) you need a bit of beef on your graphics card, but it still runs on regular consumer hardware. Anything above that is not needed for any normal use-cases.
If yes to the previous: the software doesn’t come from nowhere and ultimately still relies on gas-turbine-powered datacenters and stolen IP and stolen personal data, no?
This has multiple layers:
- for running models locally, FOSS solutions exist and are no different to any other piece of software
- models themselves are usually trained on beefy data centers and questionable data, but alternatives exist, both for training and curated data sources. Second but - those usually underperform, because LLMs need those massive data sets. Limiting the dataset limits the output quality drastically.
If what I wrote above is true, what exactly are people arguing when they say it’s still possible to use LLMs ethically or true to FOSS philosophy, because … ???
Also multiple answers:
- …because first there is a massive amount of hardliners both in the pro- and anti-LLM camp, and differentiated and objective views are rare. Ethical software and “true FOSS” are also very heated topics currently.
- …because, while possible, objectively non-ethical use of LLMs does happen more (public vibecoding of products, weaponized use in cyber and conventional warfare, use in surveillance, etc.).
- …because there are absurd amounts of money circulating in the AI bubble, it’s too big to fail, and ethical and for-profit usually do not match.
Ollama for text and multimodal (image to text). It supports structured JSON responses you can tie un to cool flows.
ComfyUI for image/audio/video/3d model generation. Use a docker image because installing and configuring it raw is a nightmare.
Huggingface is a good spot for the models. Ollama has its own library. There are several others.
You can get basic text interactions and small images on a typical office PC. You can make nice enough stuff with a typical gaming PC. More power definitely means better results though
*CivitAI also great for models, and they have a .ted domain too
