this post was submitted on 26 Aug 2024
1551 points (97.7% liked)

memes

10338 readers
1813 users here now

Community rules

1. Be civilNo trolling, bigotry or other insulting / annoying behaviour

2. No politicsThis is non-politics community. For political memes please go to [email protected]

3. No recent repostsCheck for reposts when posting a meme, you can only repost after 1 month

4. No botsNo bots without the express approval of the mods or the admins

5. No Spam/AdsNo advertisements or spam. This is an instance rule and the only way to live.

Sister communities

founded 1 year ago
MODERATORS
 
you are viewing a single comment's thread
view the rest of the comments
[–] [email protected] 6 points 2 months ago (2 children)

I'd like all ai service to publish the energy used in training the model and performing inference.

"Queries uses an average of X kWh of power. A model training run requires X MWh, and the development of this model over the years required X TWh of power."

Then we could judge companies by that metric. Off course, rich people would look for the most power-draining model for the sake of it.

[–] [email protected] 7 points 2 months ago* (last edited 2 months ago) (1 children)

That's already something that Meta is doing for their Llama models:

Source

You can extrapolate openai models consumption from these I guess

[–] [email protected] 8 points 2 months ago (2 children)

ok, but

  1. Is it still bad if they use renewables? in which case, it's not horrendous, is it?

  2. what about the rest of their servers?

  3. Fuck facebook

[–] [email protected] 2 points 2 months ago* (last edited 2 months ago) (1 children)

If we are abundant in renewable energy no but if we are still at a level where available renewable energy can be used to replace non-renewable, then AI tech needs to justify its use cases too.

and servers yes, social media related data center energy consumption should be put under heavy scrutiny too. Especially considering some energy hungry social media platforms like facebook are lately causing more harm than good (on other fronts such as political propaganda and racism). I doubt any of these are gonna happen soon though since many governments are heavily invested in using social media and LLM chatbots for propaganda and surveillance.

[–] [email protected] 1 points 2 months ago

We already have a way for society to decide what is and isn't worth spending power and effort on and it's called money.

Increase carbon taxes to incentivize clean fuel sources and ban predatory advertising and data tracking behaviour because it's problematic.

We do not need to setup a separate shadow economy to gatekeep what is and isn't worth spending eoectricity on.

[–] [email protected] 0 points 2 months ago
  1. Power consumption is still power consumption, so 2 290 000kgCo2 is a lot, even if it's way lower than what it would have been with coal plants
  2. They only talk about power consumption and not server hardware footprint, cause power consumption is the easier to offset
  3. Yes
[–] [email protected] 4 points 2 months ago* (last edited 2 months ago)

development of this model over the years required X TWh of power

This part is kind of hard to measure. When do you start counting? From the first work that informed the research direction eventually leading to this model? From the point where the concept of this final model first came about? Do you split the energy usage between multiple models that came from the same work?