Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Using ChatGPT is not bad for the environment

andymasley.substack.com

Post reply on HN
Print view View original

301–310 of 337 posts

Re: Using ChatGPT is not bad for the environment

#301

Post by simonw »

alganet wrote: post #299
simonw wrote: post #293

Earlier quoted context omitted.

> LLMs are useless (for one of their main purposes) if the models get stale. I don't think that's entirely accurate. A lot of people deliberately continue to chose to use the older GPT-4 despite it not being updated since June 2023. GPT-4o has had releases in May, August and November of 2024 - so about one every 3-4 months. Anthropic's Claude 3.5 Sonnet was released in June and had a single update in October. Persona…

LLMs that can replace search (their primary and self-declared goal) cannot survive without repeated training. As I said, this is one of its main purposes. The other main purpose (military application, surveillance, autonomous psyops) is also highly dependent on continous training. Without it, properly educated healthy humans can overcome its reasoning power very quickly. All other user profiles are just cannon fodder…

If somebody told you that LLMs are a good replacement for search that person was misleading you.

Re: Using ChatGPT is not bad for the environment

#303

Post by smcin »

fastball wrote: post #210
smcin wrote: post #142

Earlier quoted context omitted.

Could you say more? Are you talking about comparing CO2 to N2O to CH4 to fluorocarbons, for example?

Yes, that. But also that these different externalities will have different responses to reduction at scale, which also impacts how effective action on a more "personal" level really is. For example the OP was talking about plastic. A 2% reduction in plastic waste has a clear benefit, because any amount of plastic reduction is a bonus. However it is not clear that a 5% reduction in CO2 emissions due to Americans drivi…

Still not understanding you, if you meant "a 2% reduction in plastic waste in country A" is of no global benefit if there's a comparable increase in country B (assuming their disposal is roughly equally bad for the environment globally, since only 9% of plastic globally is recycled). So I assume you meant "a 2% reduction in plastic waste, globally". But then the same criticism applies to country-level reduction in CO2 emissions.

Anyway, maybe the second Trump admin will exit the Paris Accord again, putting strain on China and India participating, so climate summits will be weaker till 2029 (earliest).

Likely the domestic economic impact (on the US) of major climate-related events will have more influence on US policy than international. e.g. whether one of the many causal factors to the huge 2025 LA fires was climate change; or California's water setup where Senior Drawing Rights legally predating the existence of the state of CA allow agribusiness essential unlimited water use to grow pistachios in a semidesert during a drought, or bottle and export artesian water. As well as all the other more proximate causes. Anyway, expect a long and politicized fight between state, federal, city and insurers about the LA fire root-causes, liability and federal compensation. Especially given there's a plan to weaken or abolish the US EPA at the same time.

Ditto, expect nation-level debates (in OECD countries) reappraising whether nuclear (alongside fossil-fuel and renewable) is considered safe, desirable and necessary and how that affects/limits GDP growth, and whether datacenters and AI continue to be seen as a proxy for GDP growth, or whether that assumption breaks down some year soon.

Re: Using ChatGPT is not bad for the environment

#304

Post by alganet »

simonw wrote: post #301
alganet wrote: post #299

Earlier quoted context omitted.

LLMs that can replace search (their primary and self-declared goal) cannot survive without repeated training. As I said, this is one of its main purposes. The other main purpose (military application, surveillance, autonomous psyops) is also highly dependent on continous training. Without it, properly educated healthy humans can overcome its reasoning power very quickly. All other user profiles are just cannon fodder…

If somebody told you that LLMs are a good replacement for search that person was misleading you.

LLM companies are doing that themselves with their actions. There is a clear connection between search and assistants, historically way back to "Ask Jeeves".

People literate in IT know the implementation difference. For those not literate, the difference is way less proeminent. To most people, it's the same thing and companies know it and abuse this.

They are obviously competing for the same market. That market being "the stuff you go to when you need knowledge".

Anyway, you are deviating from the point. Even if that's not the case, my argument still holds: the article is full of shit regarding the environmental sustainability of the lifecycle of model training.

Re: Using ChatGPT is not bad for the environment

#305

Post by fulafel »

wmf wrote: post #68
scottcha wrote: post #40

Yet we also see that hyperscale cloud emissions targets have been reversed due to AI investment, Datacenter growth is hitting grid capacity limits in many regions, and peaker plant and other non-renewable resources on the grid are being deployed more to handle this specific growth from AI. I think the author, by qualifying on "chatgpt" maybe can make the claims they are making but I don't believe the larger argument…

Voluntary conservation was only working by accident and guilt tripping never works. The grid needs to become clean so that we can have new industries.

Emissions are a collective action problem. Guilt tripping works poorly directly on behaviour but it works on awareness&public discourse -> voting -> policy.

Cf how we addressed the ozone hole, acid rain, slavery, etc.

Re: Using ChatGPT is not bad for the environment

#306

Post by fulafel »

seanmcdirmid wrote: post #60
scottcha wrote: post #40

Yet we also see that hyperscale cloud emissions targets have been reversed due to AI investment, Datacenter growth is hitting grid capacity limits in many regions, and peaker plant and other non-renewable resources on the grid are being deployed more to handle this specific growth from AI. I think the author, by qualifying on "chatgpt" maybe can make the claims they are making but I don't believe the larger argument…

Is that true though? Data centers can be placed anywhere in the USA, they could be placed near a bunch of hydro or wind farm resources in the western grid which has little coal anyways outside of one line from Utah to socal. The AI doesn’t have to be located anywhere near to where it is used since fiber is probably easier to run than a high voltage power line.

The root problem there is that fossil energy is very cheap and the state sponsors production of fossil fuels. Consequently the energy incentives are weak and other concerns take priority.

This is coupled with low public awareness, many people don't understand the moral problem in using fossils so the PR penalty from a fossil powered data center is low.

Re: Using ChatGPT is not bad for the environment

#307

Post by NBJack »

Oof. This article misses some important details.

Training is not a "one time cost". Training gets you a model that will likely need to be updated (or at least fine-tuned) on newer data. And before GPT4, there was a series of prior, less effective models (likely swept under the rug in press releases) made by the same folks that helped them step forward, but didn't achieve their end goals. And all of this to say nothing of the arms race by the major players all scrambling to outdo each other.

It also needs to compare this to the efficiency modern search engines run at. A single traditional query is far less expensive than a single LLM query.

Re: Using ChatGPT is not bad for the environment

#308

Post by smcin »

ibejoeb wrote: post #174
smcin wrote: post #131

Earlier quoted context omitted.

How effective IYO was Morgan Spurlock (RIP) the activist filmmaker ("Supersize Me", [0]) at trying to educate the general (meat-eating) US public on US meat production, the restaurant industry, nutrition, marketing? (PS:it's not a dichotomy between strict vegetarianism vs meat 4+ days/wk. Can encourage people to eat healthier and more sustainable.) [0]: https://www.imdb.com/name/nm1041597/

He didn't try to educate. He sensationalized for profit. Supersize Me was a grift. He drank himself fat and fabricated his numbers.

Can't just dismiss his 17 films as a grift. Even accepting all the valid criticisms where he misled audiences, his shockumentaries did try to educate, somewhat, but also stretched facts and went for a personalized narrative, Michael-Moore-style. He did leave a legacy of nearly single-handedly reviving documentary as a film genre relevant to nutrition, one that mainstream audiences would watch.

His sequel "Super Size Me 2: Holy Chicken!" (2017) had fewer distortions and was pretty watchable and educational. Here's a review from https://www.agdaily.com/livestock/poultry/super-size-me-2-mo... which among other things covered in depth: "Many food labels, such as organic, natural, non-GMO, gluten-free, no added hormones, free range, green, artisan, antibiotic-free, are indeed quite misleading" and how those branding terms are misused in marketing to restaurant customers. I haven't seen any industry criticism of the segment with the two branding consultants.

As to the legacy of "Super Size Me", here's a review by an ag evangelist: https://www.eater.com/24173039/morgan-spurlock-legacy-super-...

Re: Using ChatGPT is not bad for the environment

#309

Post by saaaaaam »

bdangubic wrote: post #300
saaaaaam wrote: post #297

I wish charts like this would score the carbon impact of having children. It would be off the scale.

yea lets not do that. if we all decided not to have kids we would not have to worry at all about the fucking environment eh? :)

I’m not saying people shouldn’t have children. Just that if you do, you have to stop flying, go vegan, and sort your recycling properly for ever after!

Re: Using ChatGPT is not bad for the environment

#310

Post by aziaziazi »

Thinks like heating and car heavily depends on the usage, which I guess is based on the USA average.

US houses are HUGE and even here in Europe square m2 / person double in the last decades.

- we don’t have a housing problem, we have a surface inflation problem.

- heating is directly correlated to the volume to heat. Heating 100m2/person with (coal Chinese steel, resource extracted, logistics…) solar and batteries or heat pump isn’t necessarily more carbon or water efficient that 20m2/person with gas.

Bonus point: the resident will have to think twice before filling his property with garbage consumerism.

Ps: my GF and I live in 80m2 house, the precedent family where… 2 adults and 3 children! I thing the space is wayyy enough for us but people visiting regularly remark "it’s so tiny/small! "

Post reply on HN