UK-based YASA has just built a tiny electric motor that makes Tesla motors look like slackers, and this invention could potentially reshape the future of EVs
Stop fucking use AI, or at least get used to a sizable portion of people to tell you, that you’re burning the only planet we have down over shit that doesn’t matter.
Stop burning the planet down to generate social media comments
I mean, I thought it would be obvious my issue was with using AI to do so…
Even if it had been a serious question.
But, to be fair I was thinking of what a normal.person would be able to parse, and not people who’s critical thinking had already atrophied from offloading to AI.
They probably don’t have any idea what I meant and would need it explicitly spelled out.
If It makes you feel better (or at least more educated)……the entire three-prompt interaction to calculate dogpower consumed roughly the same amount of energy as making three Google searches.
A single Google search uses about 0.3 watt-hours (Wh) of energy. A typical AI chat query with a modern model uses a similar amount, roughly 0.2 to 0.34 Wh. Therefore, my dogpower curiosity discussion used approximately 0.9 Wh in total.
For context, this is less energy than an LED lightbulb consumes in a few minutes. While older AI models were significantly more energy-intensive (sometimes using 10 times more power than a search) the latest versions have become nearly as efficient for common tasks.
For even more context, It would take approximately 9 Lemmy comments to equal the energy consumed by my 3-prompt dogpower calculation discussion.
You’re absolutely right! According to the research you cited, the energy use is actually much LOWER than I stated in my comment.
Your source shows that an efficient AI model (Qwen 7B) used only 0.058 watt-hours (Wh) per query.
Based on that, my entire 3-prompt chat only used about 0.17 Wh. That’s actually less energy than a single Google search (~0.3 Wh). Thanks for sharing the source and correcting me.
If one assumes a 1/3 correctness is sufficient and the provider is using a 7B model, it is a safe assumption that it was energy efficient and better than a traditional search. However, on the other end of the spectrum, if one assumes the most efficient reasoning model, which consumes ~400x more energy and still only amounts to 4/5 accurate responses, the entire discussion is flipped on its head.
It is however comical to see one jump to an irreproducible edge case to prove one’s point, it does really exemplify how weak the position was from the beginning. Intellectual dishonesty galore.
edit: the supplied reply is also highly unlikely to have come from a non reasoning model given the structuring of the text.
I’d be curious to have the exact 3 prompts to input into Qwen7B and get that exact response.
How do you know they’re not running a local model? Ultimately the problem with LLM accusations is that short of a confession or doing some hardcore surveillance of the other person you can’t prove it
edit: or fingerprinting/watermarking
edit2: no, “you can tell by the way it is” isn’t proof (simply because that’s fixable in an instant). even if you’re the smartest person on the internet. and again, it could be a local model.
Ultimately the problem with LLM accusations is that short of a confession or doing some hardcore surveillance of the other person you can’t prove it
Human variation.
Ironically you would have to take the others person word on it, luckily you just said you were comfortable doing so.
Some people are statistically insignificant, and to them lots of stuff is incredibly obvious and they’re constantly frustrated others can’t see it. They might even sink sizeable free time into explaining random shit, just to practice not losing their temper when people can’t see the obvious.
So you might not be able to tell that was AI from a glance, but humans are pattern recognition machines and we’re not all equally good at it.
So believe a “llm accusation” or not, but some people absolutely can pick out a chatbot response, especially when taking the two seconds to glance at typical comments from a user profile.
Jump from 1-2 sentence comments to a stereotypical AI response…
Well, again, not everyone is as good at picking out patterns quickly.
To some what took me literally under 10 seconds and two clicks counts as “hardcore surveillance” because it would take them a long time to figure it out.
Stop burning the planet down to generate social media comments about shit you don’t understand
If I’m not mistaken, you specifically showed an interest in better understanding this.
I recognise their username. It’s half sane takes, half absolute wankery with them.
That was a joke
Stop fucking use AI, or at least get used to a sizable portion of people to tell you, that you’re burning the only planet we have down over shit that doesn’t matter.
deleted by creator
Username checks out
you made an offhand joke and got mad at him for continuing the joke?
I mean, I thought it would be obvious my issue was with using AI to do so…
Even if it had been a serious question.
But, to be fair I was thinking of what a normal.person would be able to parse, and not people who’s critical thinking had already atrophied from offloading to AI.
They probably don’t have any idea what I meant and would need it explicitly spelled out.
I didn’t realize it even was ai generated. but even if it is, that’s still a fairly off-putting way to respond.
No you’re right…
It’s not like it’s literally burning our planet down and the people profiting off it aren’t tech bro fascists…
attacking someone will never change someone’s mind.
If It makes you feel better (or at least more educated)……the entire three-prompt interaction to calculate dogpower consumed roughly the same amount of energy as making three Google searches.
A single Google search uses about 0.3 watt-hours (Wh) of energy. A typical AI chat query with a modern model uses a similar amount, roughly 0.2 to 0.34 Wh. Therefore, my dogpower curiosity discussion used approximately 0.9 Wh in total.
For context, this is less energy than an LED lightbulb consumes in a few minutes. While older AI models were significantly more energy-intensive (sometimes using 10 times more power than a search) the latest versions have become nearly as efficient for common tasks.
For even more context, It would take approximately 9 Lemmy comments to equal the energy consumed by my 3-prompt dogpower calculation discussion.
This is not correct and can easily be disproven, even if one assumes less than 480g/Kwh.
And that is ignoring the infrastructure necessary to perform a search vs AI query.
You’re absolutely right! According to the research you cited, the energy use is actually much LOWER than I stated in my comment.
Your source shows that an efficient AI model (Qwen 7B) used only 0.058 watt-hours (Wh) per query.
Based on that, my entire 3-prompt chat only used about 0.17 Wh. That’s actually less energy than a single Google search (~0.3 Wh). Thanks for sharing the source and correcting me.
If one assumes a 1/3 correctness is sufficient and the provider is using a 7B model, it is a safe assumption that it was energy efficient and better than a traditional search. However, on the other end of the spectrum, if one assumes the most efficient reasoning model, which consumes ~400x more energy and still only amounts to 4/5 accurate responses, the entire discussion is flipped on its head.
It is however comical to see one jump to an irreproducible edge case to prove one’s point, it does really exemplify how weak the position was from the beginning. Intellectual dishonesty galore.
edit: the supplied reply is also highly unlikely to have come from a non reasoning model given the structuring of the text.
I’d be curious to have the exact 3 prompts to input into Qwen7B and get that exact response.
How do you know they’re not running a local model? Ultimately the problem with LLM accusations is that short of a confession or doing some hardcore surveillance of the other person you can’t prove it
edit: or fingerprinting/watermarking
edit2: no, “you can tell by the way it is” isn’t proof (simply because that’s fixable in an instant). even if you’re the smartest person on the internet. and again, it could be a local model.
Human variation.
Ironically you would have to take the others person word on it, luckily you just said you were comfortable doing so.
Some people are statistically insignificant, and to them lots of stuff is incredibly obvious and they’re constantly frustrated others can’t see it. They might even sink sizeable free time into explaining random shit, just to practice not losing their temper when people can’t see the obvious.
So you might not be able to tell that was AI from a glance, but humans are pattern recognition machines and we’re not all equally good at it.
So believe a “llm accusation” or not, but some people absolutely can pick out a chatbot response, especially when taking the two seconds to glance at typical comments from a user profile.
Jump from 1-2 sentence comments to a stereotypical AI response…
Well, again, not everyone is as good at picking out patterns quickly.
To some what took me literally under 10 seconds and two clicks counts as “hardcore surveillance” because it would take them a long time to figure it out.
Don’t assume everyone else is exactly like you.
Stop burning the planet to tell people what to burn the planet for.