I should write a Tridactyl script to use that as warning… it goes like this

document.querySelector(“textarea”).style.backgroundImage = “linear-gradient(to right, rgba(255,255,255, 0.7) 0 100%), url(https://programming.dev/pictrs/image/e7e7aeb4-ae1d-426b-bf45-02a4f3060bd6.jpeg?format=webp)”

It’s hard to read but a good reminder maybe it’s not worth it! :D

Moore’s law is kinda still in effect, depending on your definition of Moore’s law.

Sounds like the goal post is moving faster than the number of transistors in an integrated circuit.

LOL… you did make me chuckle.

Aren’t we 18months until developers get replaced by AI… for like few years now?

Of course “AI” even loosely defined progressed a lot and it is genuinely impressive (even though the actual use case for most hype, i.e. LLM and GenAI, is mostly lazier search, more efficient spam&scam personalized text or impersonation) but exponential is not sustainable. It’s a marketing term to keep on fueling the hype.

That’s despite so much resources, namely R&D and data centers, being poured in… and yet there is not “GPT5” or anything that most people use on a daily basis for anything “productive” except unreliable summarization or STT (which both had plenty of tools for decades).

So… yeah, it’s a slow take off, as expected. shrug

That’s been addressed few times already so I let you check the history if you are actually curious.

No one is saying training costs are negligible.

It’s literally what the person I initially asked said though, they said they don’t know and don’t care.

Yes indeed, yet my point is that we keep on training models TODAY so if keep on not caring, then we do postpone the same problem, cf https://lemmy.world/post/30563785/17400518

Basically yes, use trained model today if you want but if we don’t set a trend then despite the undeniable ecological impact, there will be no corrective measure.

It’s not enough to just say “Oh well, it used a ton of energy. We MUST use it now.”

Anyway, my overall point was that training takes a ton of energy. I’m not asking your or OP or anyone else NOT to use such models. I’m solely pointing out that doing so without understand the process that lead to such models, including but not limited to energy for training, is naive at best.

Edit: it’s also important to point out alternatives that are not models, namely there are already plenty of specialized tools that are MORE efficient AND accurate today. So even if the model took a ton of energy to train, in such case it’s still not rational to use it. It’s a sunk cost.

Indeed, the argument is mostly for future usage and future models. The overall point being that assuming training costs are negligible is either naive or showing that one does not care much for the environment.

From a business perspective, if I’m Microsoft or OpenAI, and I see a trend to prioritize models that minimize training costs, or even that users are avoiding costly to train model, I will adapt to it. On the other hand if I see nobody cares for that, or that even building more data center drives the value up, I will build bigger models regardless of usage or energy cost.

The point is that training is expensive and that pointing only to inference is like the Titanic going full speed ahead toward the iceberg saying how small it is. It is not small.

Right, my point is exactly that though, that OP by having just downloaded it might not realize the training costs. They might be low but on average they are quite high, at least relative to fine-tuning or inference. So my question was precisely to highlight that running locally while not knowing the training cost is naive, ecologically speaking. They did clarify though that they do not care so that’s coherent for them. I’m insisting on that point because maybe others would think “Oh… I can run a model locally, then it’s not <<evil>>” so I’m trying to clarify (and please let me know if I’m wrong) that it is good for privacy but the upfront training cost are not insignificant and might lead some people to prefer NOT relying on very costly to train models and prefer others, or a even a totally different solution.

Results? I have no idea what you are talking about. I thought we were discussing the training cost (my initial question) and that the truckload was an analogy to argue that the impact from that upfront costs is spread among users.

Well even a PWA still has to be developed and maintained.

Great point, so are you saying there is a certain threshold above which training is energetically useful but under which it is not, e.g. if training of a large model is used by 1 person, it is not sustainable but if 1 million people use it (assuming it’s done productively, not spam or scam) then it is fine?

I’ll assume you didn’t misread my question on purpose, I didn’t ask about inference, I asked about training.

I specifically asked about the training part, not the fine tuning but thanks to clarifying.

Edit : you might be interested in helping with https://lemmy.world/post/30563785/17397757 please

I see. Well, I checked your post history because I thought “Heck, they sound smart, maybe I’m the problem.” and my conclusion based on the floral language you often use with others is that you are clearly provoking on purpose.

Unfortunately I don’t have the luxury of time to argue this way so I’ll just block you, this way we won’t have to interact in the future.

Take care and may we never speak again.

You know what, again maybe I’m misreading you.

If you do want to help, do try with me to answer the question. I did give a path to the person initially mentioning the Model Card. Maybe you are aware of that but just in cased a Model Card is basic meta-data about a model, cf https://huggingface.co/docs/hub/model-cards

Some of them do mention CO2 equivalent, see https://huggingface.co/docs/hub/model-cards-co2 so here I don’t know which model they used but maybe finding a way have CO2 equivalent for the most popular models, e.g DeepSeek, and some equivalent (they mentioned not driving a car) would help us all grasping at least some of the impact.

What do you think?

Please, do whatever you want to protect the environment you cherish. My point though was literally asking somebody who did point a better way to do it if they were aware of all the costs of their solution. If you missed it, their answer was clear : they do not know and they do not care. I was not suggesting activism, solely genuinely wondering if they actually understood the impact of the alternative they showcased. Honestly, just do whatever you can.

Apologies for my sarcastic answer, I did actually search for that a little while ago so I do assume most people do know but that’s incorrect. The most useful tool I know of would probably be https://www.aspistrategist.org.au/uyghurs-for-sale-re-education-forced-labour-and-surveillance-beyond-xinjiang/ It wasn’t specific to children but it does show the process and I’d argue can be apply for the different criteria one would want to focus on. It dates back few years ago, when I learned about the problem so there also you might want to prefer a more up to date source.

Let me know if you are looking for something more precise. I know of few other tools which do help better understand who builds what and how, for electronics but other products too.

FWIW the person I asked did reply, they don’t care : https://lemmy.world/post/30563785/17397024

Hope it helps.

Straw-hat much or just learning about logistics and sourcing in our globalized supply chain?

Feel free to explain the down votes.

If it wasn’t clear the my point was that self hosting addresses mostly privacy for the user but that is only one dimension addressed. It does not necessarily address the ecological impact. I was honestly hoping this community to care more.