Oriiiiiii :
Listen, I am an AI developer myself and I would like to point a few things off.
Firstly, the water being used to cool off the AI datacenters isn't super fresh filtered mineral water. It's just normal water, that after it has been used to cool off tech gets evaporated back into the atmosphere as steam, which returns as water.
Secondly, while I think the idea of ecogpt is great, it has a few flaws. You see, I have asked the ecogpt model what model it is or where it comes from. Simply put it doesn't know what is is only that it is a model that is made by ecogpt or made with a mix of open source models. This means that either ecogpt made a new from scratch model and trained it from zero. For those who don't know training is the part that take the most water and resources so that isn't actually eco. if they didn't make their own model and used an open source mix, those models still had to be trained somehow so again even though it's not directly it's not eco. Then we have the hosting, hosting or running the model also takes more resources the more it's being used and also running it also take water.
Again I think that this is a great idea, and if they plant trees every 100 prompts there is no flaws with that and it's great, but if they wanted to truly be eco, they would make a way for normal people to run it locally, meaning the AI runs on our device only, no need for internet connection, and not using loads of water.
I hope I made this as accurate as I could and hopefully you understand. EcoGpt is a good idea but it's not fully working as advertised. @Carterpcs @Parker Burton
2026-08-09 07:01:57