Before even getting to documentation, I see so many projects that don't have a short summary of what they do (and maybe what to not expect them to do).
As an example, Home Assistant. I can tell that it involves home automation, so can I replace Google Home with it? It seems like it doesn't do voice recognition without add-ons and it can work with Google Assistant. Do I still need accounts with the providers of smart appliances, or can it control my bulbs directly?
None of that is very clear from the website.
I've seen plenty of other projects where it's assumed there's no need to explain it's overall purpose.
A similar chart could be made for the US, proving that it does use metric: soda and wine bottles, medicine doses, eye-glasses measurements (in fact most medical things).
I think that both systems are used in schools now.
But then I see cooking instructions for a "cup of chicken strips" and a recipe having 1/4 cup of butter, and I wonder why anyone thought that volume was a good idea there.
I would make international calls frequently. I would buy calling cards. The process was: dial the 800 number on the card. Enter the id number on the card to use some of its credit. Dial the number to call. Their service would then connect me at a low rate to another country(probably making a voip call).
So I'd set up the 3 speed dial buttons with those. For each new card I'd only have to change the card's unique number.
It's true, I don't know how large the models are that are being accessed in data centers. Although if the article's estimate is correct, it's sad that such excessively-demanding models are always being used for use-cases that could often be handled with much lower power usage.
It's possible to run an LLM on a moderately-powered gaming PC (even a Steam Deck).
Those consume power in the range of a few hundred watts and they can generate replies in a seconds, or maybe a minute or so. Power use throttles down when not actually working.
That means a home pc could generate dozens of email-sized texts an hour using a few hundred watt-hours.
I think that the article is missing some factor, such as how many parallel users the racks they're discussing can support.
I like that I can interface with it in ways that I already understand (eg rclone, sync, sshfs).
Being able to run some commands on the server meant that I could use rclone to copy my AWS and OneDrive backups directly cloud-to-cloud.